# Collapse or Collapse: How AI Agents End the Department, the Manager and the Meeting

By Vinay Pasricha. © 2026 Vinay Pasricha. All rights reserved. The full text, shared free on https://vinaypasricha.com with the author's permission; quote with a link to the chapter URL given before each chapter.
Book page: https://vinaypasricha.com/books/collapse-or-collapse
Kindle: Amazon.com https://www.amazon.com/dp/B0HNL53J5R · Amazon.in https://www.amazon.in/dp/B0HNL53J5R · Amazon.co.uk https://www.amazon.co.uk/dp/B0HNL53J5R
Guide: https://vinaypasricha.com/books/collapse-or-collapse/llms.txt


================================================================
# Chapter 1: The Reorder Book
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-1-the-reorder-book

*Part No. 1147 · Bush, bronze, 32 mm · Reorder level 40 · Checked: 14 March 1988 · Initials: R.K.*

*A note on the people. The company in this chapter was real and so was the work. The men who ran it are composites, and their names are invented, because the point of the story is the chain they formed and not the men who happened to stand in it. Naval is real, and so is Orange.*

Once a month, Mr Malhotra shut his door.

He was the chairman of a company that built printing machines, and for one week in every four he was not available to anyone. Not to customers, not to the bank, not to his wife, who rang on the third day of every projection week to ask whether he was still alive and was told by the peon, with great seriousness, that sir was inside with the ledger. The sales ledger went in with him, and the file of pending orders, and a pad of lined paper, and a flask of tea that was refilled through a door opened six inches and no more. He came out five or six days later, unshaven, holding a single sheet: how many of each machine the company would sell over the next six months.

Nobody else was allowed to make that projection. He had built the company from one machine and a rented shed, he knew every customer by name and most of them by temperament, and he believed, probably rightly, that nobody else could guess the future of the order book as well as he could. So every month the company’s most capable mind spent a week doing arithmetic, and the company’s most expensive door stayed shut.

The sheet went to Dhawan, the plant manager. Dhawan was a large, unhurried man who had been with the company since the first machine, and he carried the whole plant in his head: which part went into which subassembly, which subassembly into which machine, how long a casting took to come from the foundry, how long a bush took to turn, which of the thousands of parts on the shelves had to exist before anything above them could be built. He never consulted a drawing. If you asked him how many gears went into the folding unit he would close his eyes for a moment, as though reading something off the inside of his eyelids, and tell you. From the chairman’s single sheet he worked downward, machine by machine, assembly by assembly, part by part, and wrote it all out by hand. It took him two weeks and filled a stack of foolscap three hundred sheets thick. He did it from memory, because there was nowhere else the information lived, and if Dhawan had been knocked down by a scooter on the Mathura Road the company would have discovered that it did not know how to build its own machines.

The stack went to his clerk, a thin, careful man named Ramesh, who copied every line into the part registers. There was one register page for every part, thousands of pages in a row of bound ledgers along the stores wall, and each page carried the part’s history: what had been issued, what had been received, and now, in Ramesh’s small, upright hand, what the plant manager said the next six months would need. Ramesh worked for most of a week. He wrote with a fountain pen, and he kept a second, cheaper pen for corrections, and the registers were the one place in the building where you could read, in two inks, the history of everyone’s mistakes.

The registers went to Verma, the production planning manager, whose job was to find the mistakes, and there were always mistakes, perhaps fifty in a month. Not because anyone was careless. Because some parts that the plan said must be made were already on the shop floor, half-finished, as work in progress; because some parts had been replaced by other parts since the last plan and the old number was still in Dhawan’s head; because a line copied from one sheet to another by a tired man at four in the afternoon is sometimes copied wrong. Verma took a grim pleasure in the fifty. He kept a count, month by month, on the inside cover of his diary, and once told me, as a point of professional pride, that he had never found fewer than thirty.

And last, the store clerk, Rao, walked the shelves with the registers open and compared what the plan said with what he could see and count, and with what he expected to use in the weeks it took a reorder to arrive, and raised the orders. Rao had been doing this for eleven years and did not entirely trust anything written down, including his own register, so he counted. Every month, every shelf.

Five men. Five desks. A month. And at the end of the month, a set of reorder levels that had been true at the beginning of it.

## What a month costs

I want to be exact about why this was expensive, because the expense was not where a visitor would have looked for it.

The visible cost was the time: a week of the chairman, two weeks of the plant manager, most of a week each from the clerk, the planner and the storekeeper. Add it up and the reorder book consumed something like six man-weeks every month, most of it from the four or five most experienced people in the building. That was bad enough. But it was not the real cost.

The real cost was fear. Every man in the chain knew that the number he wrote down would not be looked at again for a month, and that if it was too low, somewhere in that month a machine would stand on the floor unable to ship because one bush or one gear was not on the shelf. A finished printing machine is a large, expensive object, and a customer who has been promised it does not want to hear about a bronze bush. So every man, at every desk, rounded up. The chairman projected a little high, to be safe. Dhawan added a margin on top of the chairman’s margin. Rao, walking the shelves, added his own. By the time the reorder levels reached the supplier, they carried four layers of caution, each reasonable on its own, and the shelves filled with parts that would not be needed for months. The money tied up in that stock was the largest sum of money in the company, more than a year’s wage bill, and nobody had decided to spend it. It had simply accumulated, one careful margin at a time, as the price of an answer that took a month to compute.

And even then the answer was wrong, because it was stale. The projection was made on the first of the month from orders that had arrived by the last day of the month before. By the time Rao raised the purchase orders, four weeks had passed, two new orders had come in, one had been cancelled, and a customer in Ludhiana had rung to ask whether his machine could be brought forward. None of this was in the registers. It could not be. The registers were a photograph of the company taken a month ago and handed to people who had to act today.

## The machine that could not go

I remember the first stockout I saw, because it was my second week in the building and I did not yet understand what I was looking at.

A machine stood on the assembly floor, complete in every respect but one. It had been promised for the end of the month; the customer had already paid part of the price, and had, I learned later, already told his own customers. The missing part was a small thing, a hardened steel component that went somewhere inside the drive. The register said there should have been six on the shelf. There were none. Three had gone into machines built the previous month, which the register knew about. Three had been issued to the shop as work in progress for machines not yet built, which the register knew about too, on a different page, in a different hand, and nobody had put the two pages together.

The part took three weeks to source. The machine stood there for three weeks, with men walking round it, and after the first few days they stopped looking at it, the way people stop looking at a relative who has overstayed. The customer rang every second day, and then every day, and then stopped ringing, which was worse. On the Tuesday of the third week Mr Malhotra came down to the floor, which he did not often do, and stood in front of the machine with his hands behind his back for what felt like a very long time. He did not shout at anyone, because there was nobody to shout at. Every man in the chain had done exactly what the system asked of him, and the system had produced a machine that could not go. He asked Rao how many of the part were on order. Rao told him. He nodded, and went back upstairs.

That was the moment I understood what we had been let in for. Not a stores problem, and not a people problem. A problem of a result that mattered, which no one person could hold in his head, and which had therefore been cut into five pieces and handed down a line, and which leaked a little at every handoff until, at the end, something was simply missing.

## Two outsiders

I should say what I was doing there, because I was not an employee and never became one.

I had a company, or half of one. It was called Orange Technology, and I had started it after engineering college and eighteen months at Eicher, with Naval Bansal, who had sat two rows behind me for four years and was the only person I knew who was more certain than I was that we should not work for anyone. Orange wrote custom software for anyone who would pay for it. It consisted of the two of us, a second-hand PC, a dot-matrix printer that could be heard from the street, and a scooter on which we visited every factory and office within a day’s ride of the city, one after another, to tell people who had not asked that their paper could be made to disappear.

That is how we found Mr Malhotra’s company, and how we had found the ones before it. There was no introduction and no uncle. We had a list of factories from the industrial directory and we went down it, gate by gate, and at most gates we were told that sir was busy, and at some we were given tea and shown out, and at a few, after the third or fourth visit, we were given a problem. By the time we came to the printing-machine works we had several of these behind us, each one sold the hard way, and we had learned the first rule of the trade, which is that nobody buys software. They buy the end of a particular pain, and you have to find the pain before you can sell them anything.

At Mr Malhotra’s, the pain was on the assembly floor, complete in every respect but one.

The second rule we had learned was not to promise anything until we understood it. So we did not arrive with a program, though we were sure enough of ourselves to have written one on the spot if he had asked. We arrived with a proposal for a systems study: a few weeks in the building, at a fee that was small enough to be agreed on the spot, at the end of which we would tell them what could be done and what it would cost, and they could say no. Mr Malhotra said yes to that in about four minutes, which we took to be a sign of how much pain there was, and gave us a chair in the corner of Verma’s office, a cup of tea every two hours, and the run of the building.

I knew what a reorder level was in the way that one knows what a carburettor is from a diagram. Inventory control had been one chapter in one course at college, examined once and forgotten. Neither of us had ever seen a register, never walked a stores shelf, never met a man like Dhawan who carried a plant in his head.

So we spent the study doing nothing but understanding. I sat with each of the five men in turn and asked them to show me what they did, and then asked them to show me again, slowly, and I drew it. I drew the chairman’s projection sheet and the columns on it. I drew the way Dhawan exploded a machine into its assemblies and the assemblies into their parts, and the lead time he attached to each, and where that lead time came from. I drew a register page and every entry on it. I drew what Verma looked for when he hunted for mistakes, and what Rao did on the shelves. Naval, who was better at arithmetic than I was and worse at listening, took my drawings in the evenings and turned them into tables and counts: how many parts, how many assemblies, how many entries a month, how many hours. By the end of the month the whole chain was on paper, in our hands, as a set of boxes with arrows between them, and we could see something that nobody standing in the chain could see, because each of them saw only his own desk.

We could see that it was arithmetic. All of it. The chairman’s week was judgement, and judgement we could not replace; but once his single sheet existed, everything that happened to it afterwards, on four desks over three weeks, was multiplication, addition, subtraction and lookup. Dhawan’s three hundred sheets were a bill of materials multiplied by a sales forecast. Ramesh’s week was copying. Verma’s fifty corrections were the places where two copies of the same fact disagreed. Rao’s walk was a comparison of two columns. There was not one decision in the three weeks after the chairman’s sheet that a machine could not make, faster, every day, from data that was as new as that morning.

We were certain of this from the first week, long before we had earned the right to be. We did not say it loudly, because Dhawan had been doing the job since before we were born and we had the sense to let the study finish before we told him his fortnight was arithmetic. But we did not doubt it for a moment, and we never did, through everything that followed. We were two very confident young men. We had been short of money since the day we started, and we would be short of money for years, and it had never once occurred to either of us that this was a reason not to take on whatever was put in front of us. The only question was whether they would let us build it.

## The presentation

At the end of the study we asked for an hour with Mr Malhotra, and were given it, in his office, with Dhawan and Verma called in and the door shut, which we understood to mean that it was serious.

We had our drawings. We laid them out on his desk, all five desks and all the arrows, and we told him what I have just told you: that after his sheet, it was arithmetic, and arithmetic could be done by a machine in an hour instead of by four men in three weeks, and done again the next morning, and the morning after that, from numbers that were never more than a day old. We told him the stockouts would stop, because the register and the shop floor would be the same page. We told him the stock would fall, because four men would no longer each be adding a margin against a month of blindness. Naval had a sheet of numbers, and he went through them one by one, looking Mr Malhotra in the eye after each, as though daring him to find one wrong.

Mr Malhotra listened without saying anything, which was his way, and then he asked three questions. The first was how long. We said six months, and he looked at us in a way that told us he had heard that number from other people. The second was what it would cost. We told him, and he did not react at all, which was worse than if he had. The third was the one we had known was coming. He turned to Dhawan, who was sitting with his hands on his knees, and asked: and what happens to you?

Dhawan said, after a moment, that he supposed he got his fortnight back. Mr Malhotra did not say anything to that either. But he picked up one of my drawings, the one of his own projection sheet, and looked at it for a while, and I saw something in his face that I did not understand until years later, when I had my own company and my own shut door: the look of a man who has been told that the thing he does with his best week every month is not, after all, the thing that only he can do.

Then he told us his terms, and they were the terms of a man who had bought a great many machines. There would be a small advance, small enough that I remember thinking it would not cover the petrol. After that we would be paid in stages, and only when each stage was done to his satisfaction and not before: first a full specification of the system, every screen and every report, which Verma would check; then a pilot, on one family of parts, run beside the registers until it agreed with them; then the implementation, piece by piece, each piece signed off before the next was paid for. If a stage was late, we would wait. If it was wrong, we would do it again, at our cost. He said all this without any unkindness, as a man states the price of bread, and then he asked whether we agreed.

We agreed before he had finished asking. We were young, we had never once doubted that we could do anything that was put in front of us, and a man offering to pay us only for what worked did not frighten us; it flattered us. We walked out into the corridor and Naval said: he is going to make us work for every rupee. He was grinning. He was right. It was the best thing that ever happened to us.

## Six months

There was no software to buy. In 1989 in India there was no package we could find that did this, nothing to install, nobody to call. If the reorder book was going to become a program, somebody was going to have to write it, and the somebody was the two of us, in Verma’s corner, on our own machine, which we carried in from the scooter every morning in two trips.

We wrote the specification first, because we had to, because it was the first stage and nobody was paying for anything until Verma had read it. I had never written one before. It took six weeks and ran to a thickness that embarrassed us, and Verma read every page of it with the same pleasure he brought to the monthly fifty, and sent it back twice. By the third time it was, I think, the most complete description of that company’s stores that had ever existed, and it was also the first money Orange earned from Mr Malhotra. I remember the cheque. It was not large, and it was already spent, because we were always short of money and had been living for two months on the advance and on what the earlier clients still owed us. We looked at it for longer than it deserved, and then Naval took it to the bank on the scooter before the ink was dry.

Then the program. I wrote it first in FoxBASE, with pieces in C where FoxBASE was too slow, and moved it to FoxPro when FoxPro appeared, which happened in the middle of the work and felt like a gift from somewhere. I will not describe the program, because the program is not the point and because it was, by any later standard, crude. What matters is what it held. It held the bill of materials, every machine exploded into its assemblies and parts, which had lived only in Dhawan’s head. It held a lead time against every part, which had lived there too. It held the register, every issue and receipt, keyed in as it happened rather than copied once a month. And it took the chairman’s sheet, which still came out of his room once a month in his own hand, and did in a few minutes of machine time, and an hour from keying to checked report, what four desks had done in three weeks.

Dhawan gave us his afternoons, which was the real gift. He would come to Verma’s corner after lunch, lower himself onto a stool that was not built for him, and take the plant out of his head one assembly at a time while I put it into tables. It went slowly, because he would stop to tell us why a part was the way it was, and who had designed it, and what had gone wrong the year they had tried a cheaper one, and I learned more about printing machines in those afternoons than I have learned about anything since. He was not threatened by what we were doing. He said more than once, with something like relief, that he had been waiting twenty years for somebody to write it down, and that he had sometimes lain awake wondering what would happen to the company if he forgot.

Verma was another matter. Verma had understood, before anyone else, that a program which found mistakes automatically would put an end to the monthly fifty, and with it to the only part of his job that anyone noticed. For the first two months he was correct, distant and exacting, and since he was also the man who signed off each stage before we were paid, he was in a position to be exacting at some length. We did not mind. We had met Vermas at other gates, and we had learned that the man who is hardest to convince is the one worth convincing, because once he is on your side he brings the building with him. What changed him was not anything we said. It was the pilot. We ran it on one family of parts, the drive components, beside the registers, as the terms required, and the first time it was run against a month-old register it found sixty-one mistakes to his fifty. He spent an afternoon checking them and found that it was right about fifty-eight. He came to our corner at six in the evening with the printout, put it on the desk, and said, with the air of a man conceding a chess game: it is better than me. Then he said: now make it find the other three. Then he signed the pilot off, and we were paid for it, and after that he was the program’s fiercest tester, and I think its proudest user, and his diary got a new count.

Ramesh learned to type in a fortnight and never looked back, though he kept the fountain pen in his pocket for a year, the way a man keeps a watch that has stopped. Rao did not trust the screen and said so, and went on counting the shelves every month, and the program was better for it, because every time his count and the screen disagreed we found something: a receipt note never entered, a part issued on a chit and never recorded. He was, without knowing it, our audit.

The implementation went in piece by piece, as the terms required, and each piece was paid for only after the men who used it said it worked. By 1989, after about six months, the reorder book was a program. Then we did the thing that I now believe was the most important decision in the whole project, though at the time it felt like cowardice, and though in truth it was Mr Malhotra’s condition and not our idea. We ran both. For six more months, the registers were kept exactly as before, and the chairman’s sheet went to Dhawan exactly as before, and the three hundred pages were written and copied and corrected and walked, and alongside all of it the program produced the same answer from the same sheet in an hour. Every month, Verma compared the two. Every month, the program found mistakes in the registers, and occasionally the registers found a mistake in the program, and I fixed it that day, sometimes that night, in Verma’s corner with the lights off in the rest of the building and the watchman looking in every hour to see whether the boys had finally gone home. After six months of that, nobody in the building trusted the registers more than the screen, and the last stage was signed, and the last cheque came.

We were paid, in the end, every rupee we had quoted, and not one more, and not one of them early. We were short of money the whole way through, as we were short of money through every year of Orange, and I have come to think that this was not a hardship but a kind of education: a company that is always a cheque away from trouble learns very quickly what a customer will actually pay for. I have had easier clients since. I have never had a better one. A man who pays you only for what works teaches you, faster than any college, what working means.

## What came back

Nobody lost his job. I say that plainly because it is the first question anyone asks and because it was true. The word at the time was computerisation, and what computerisation did in that company was give time back. Mr Malhotra kept his week of projection, because that was judgement and the program could not do it, but the three weeks that used to follow his sheet simply stopped happening. Dhawan got his fortnight back every month and spent it on the shop floor, where he had always wanted to be, and where, within a year, he had redesigned the folding unit that had been giving trouble since before I was born. Verma, with no copies to reconcile, turned to the scheduling of the floor itself, which had never had anyone’s full attention, and found that it was full of the same month-old answers. Ramesh became the man who kept the data clean, which turned out to be a more important job than the one he had had before, and the first person in the company to be sent on a training course. Rao still walked the shelves, but now with a printout that was true as of that morning, and he stopped adding his margin, because he no longer needed to. He never said the screen was right. He said it was less wrong than it used to be, which from Rao was a speech.

The stock fell. Not at once, and not because anyone ordered it to, but because four layers of caution had been taken out of the chain and replaced by one answer that was never stale. Within a year the money tied up on the shelves had fallen substantially, and the stockouts, the machines standing complete but for one bush, became rare. The company did markedly better, and the only thing that had changed was that five people handing a month-old answer down a line had been replaced by one calculation that was never out of date.

There was no morning when it printed itself and everyone cheered. I had half expected one; I had, I am embarrassed to say, imagined it. Instead there were bugs, every day, small ones, found by the men who used it and fixed by me before the next day, and the program grew the way such things grow: one more register, one more kind of receipt note, one more report that Verma wanted, one more exception that Rao had always handled in his head and now had to be written down. It was never finished. I understand now that it could not have been, and that the feeling of its being unfinished was not a failure but the ordinary condition of a system that is actually in use.

Mr Malhotra’s door still shut once a month. But his wife told me, at a wedding years later, that he came out on the fourth day now, not the sixth, and that he had started to whistle.

## The registers

I do not know the exact day the registers stopped. There was no decision. Ramesh kept them through the six months of running both, and then, I think, for a while afterwards out of habit, and then less often, and then there was a month when nobody opened them at all, and nobody noticed that nobody had.

I found them a year or so later, on the shelf along the stores wall where they had always been, with a film of dust on the top ledger. Orange had more clients by then, most of them found the same way, gate by gate, and I had come back on the scooter to fix something small. I took one down and opened it at random. Part number, description, reorder level, a date in 1988, initials. Four men’s handwriting on the page, two of them in two inks, each correcting the one before. I stood there for a minute with it open, and I thought about Dhawan on his stool, and Verma with his printout at six in the evening, and Naval daring a chairman to find one number wrong, and the machine that could not go. Then I put it back, and I did not open another one for thirty-five years.

================================================================
# Chapter 2: The Department Was a Workaround
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-2-the-department-was-a-workaround

Four men’s handwriting on one page is the whole history of management.

I did not think that when I stood in the stores with the ledger open. I thought it thirty-five years later, when I had run companies of my own, built departments of my own, and watched a machine do in an afternoon what one of those departments had done in a month. But the page had already said it. Each man had written what he knew and handed the page to the next, because no one of them could hold the whole. The whole was the reorder level of one bronze bush: what would sell, what that needed, what was on the shelf, what was on its way. Four kinds of knowledge, four heads, four hands, and a month between the question and the answer.

That is a department. Not the room, not the sign on the door, not the man who runs it. A department is what you build when a result needs more than one person can hold, and you cut the result into the pieces that single people can hold, and you put a desk under each piece. Everything else about it, the hierarchy, the reports, the meetings, the budget, the career, the business school that teaches you to run it, grows out of that one cut.

This chapter is about the cut. Why it was made. What it has cost, in money and in outcomes, for two hundred years. Why every discipline of management you were taught is a sub-discipline of it. Why it no longer needs making. And why I expect that, within five years, every company that has not undone it will be undone by one that has. I am going to say these things without hedging, because they are not opinions. They are what follows from one constraint and its removal, and the constraint has been removed.

## The constraint

Start with what a company is for. It exists to produce outcomes for someone outside it: a machine that ships, a loan that is approved, a patient who goes home, a hire who joins. Each of these is a single thing to the person who receives it. The customer does not experience sales and then production and then dispatch; he experiences a machine arriving or not arriving. The outcome is whole on the receiving end.

It is on the producing end that it comes apart, and it comes apart for a reason so old and so obvious that nobody questions it. An outcome needs many skills. The machine that ships needs someone who can forecast, someone who knows the plant, someone who keeps records, someone who finds mistakes, someone who counts shelves. A loan needs someone who can read a balance sheet, someone who knows the law, someone who knows the customer, someone who knows the bank’s appetite that month. No one person has all the skills, and even if one did, no one person has the hours. So the outcome is divided into the parts a single person can carry, and each part is given to a specialist, and the specialists are grouped by what they are good at, and the groups are called departments.

Note what has happened. The company has stopped being organised around the thing the customer feels and started being organised around the things its people are good at. That is not a small change. It is the whole change. Every management writer of the last two hundred years has been describing its consequences, and almost none of them named its cause, because the cause was invisible in the way that water is invisible to a fish. No one person could hold the whole. It was simply true, everywhere, always, and so nobody wrote it down.

I am writing it down, because it has stopped being true.

## Two centuries of the workaround

The workaround did not arrive at once. It hardened over two hundred years, and each generation of the best minds in business added a layer to it, honestly, because each of them was answering the same constraint with the best tools of their time. I will take the arc quickly, because the arc is the point: every step made the cut deeper, and every step was right.

**Smith named the gain.** In 1776 Adam Smith watched ten men in a pin factory divide the making of a pin into about eighteen operations, one man drawing the wire, another straightening it, a third cutting it, and found that they made upwards of forty-eight thousand pins a day where a man working alone, he judged, could scarcely have made twenty.[^1] Division of labour multiplied output two-hundredfold, and Smith was right, and the world took his lesson and ran with it for two centuries. What he did not dwell on, because it was not his subject, was that the pin now belonged to nobody. Ten men had made it and none of them had made a pin.

**Taylor named the task.** A century and a quarter later Frederick Winslow Taylor took Smith’s division down to the level of the single motion, timed it with a stopwatch, and gave the result to management: the worker was to do what the card said, and the thinking was to be done elsewhere, by people whose job was thinking.[^2] Scientific management was the cut made explicit and made permanent. The person who held the skill and the person who held the purpose were now, by design, different people, and a new layer was invented to stand between them. That layer is still there. It is called management.

**Ford built the cut into the floor.** The moving assembly line of 1913 took Taylor’s tasks and arranged them in a sequence so that the work came to the man rather than the man to the work, and the time to assemble a chassis fell from over twelve hours to ninety-three minutes.[^3] Ford’s line is the department made physical: every station a specialty, every handoff a few feet of chain, the whole outcome visible only from the gallery above. It is also, and this will matter later in the book, the first system in which a stall anywhere stopped everything and everyone could see it. The line did not tolerate silent failure. Offices, which copied everything else from Ford, did not copy that.

**The corporation drew the chart.** By the middle of the twentieth century the workaround had a picture of itself: the organisation chart, boxes in rows, lines between them, a function in every box and a head on every function, divisions above the functions and a headquarters above the divisions. The chart was not a description of how outcomes were produced. It was a description of how the workaround was staffed, and it became so universal that within a generation nobody could imagine a company without one. Ask a child today to draw a company and she will draw boxes. The boxes are a hundred years old and they were never the point.

**Drucker named the worker who came after the line.** The knowledge worker of 1959 carried his means of production in his head, could not be timed with a stopwatch, and had to be managed, Peter Drucker said, by objectives rather than by tasks.[^4] This was the office finding itself. The forecaster, the planner, the analyst, the clerk were knowledge workers, and the company organised them exactly as Ford had organised fitters: by what they knew, in functions, with a manager above each function to carry the purpose the function could not see. Drucker understood better than anyone that the function was a problem. He spent fifty years telling managers to look at the customer and the result rather than at their own department. He could not tell them how to remove the department, because the constraint that created it still held.

**The business school taught the workaround as a science.** Through the second half of the century the management of the cut became a profession, with its own degree, its own journals, its own consultancies, and its own certainty. I sat in its classrooms myself, at Harvard, on the programme it runs for owners of companies, and I will say in a moment what it taught me. For now it is enough to notice what the MBA is: a two-year education in how to run a company that has been cut into pieces, taught by people who never saw a company that had not been. Organisational design is the science of where to make the cut. Span of control is the science of how many pieces one manager can hold. Finance is the science of allocating money across the pieces. Human resources is the science of staffing them. Operations is the science of the handoffs between them. Not one course in the curriculum asks whether the pieces need to exist. The constraint was so complete that the question could not be formed.

**Hammer tried to undo the cut without the machine.** In 1993 Michael Hammer and James Champy published *Reengineering the Corporation*, and for a moment the whole of management stood at the edge of this book.[^5] They saw the leak at every handoff. They said, in so many words, that companies should be organised around outcomes rather than tasks, that one person should carry a whole process where possible, and that the functional organisation was a relic of Smith and Taylor. They were right about all of it. And reengineering failed, by their own estimate, in half to seventy per cent of the companies that tried it, for one reason that they did not have the tools to see: they were asking one person to hold the whole, and one person still could not. The skills still lived in heads. You could redraw the chart around the outcome, but the outcome still needed seven heads, and the seven heads re-formed into seven desks before the consultants’ invoices were paid. Reengineering was the collapse attempted thirty years early, without the machine that makes it possible. It is the strongest proof I know that the constraint, and not the chart, was always the thing. Hammer had already named the other way to fail, three years before the book, and chapter 5 takes it up: paving the cow path.

So the workaround hardened. Smith divided the outcome, Taylor separated skill from purpose, Ford laid the pieces in a line, the corporation drew the chart, Drucker tried to manage the pieces back towards the whole, the business school made the whole apparatus a science, and Hammer tried to tear it down with his bare hands and could not. Two hundred years, seven honest answers to one constraint, and at the end of it an organisation that every company in the world recognises as natural, as though companies had always come in boxes, and as though the boxes were the point.

They were never the point. They were the price.

## The price

Here is the price, in the only unit that matters, which is what the customer feels, and then in the unit the owner feels, which is money.

Every time an outcome crosses from one desk to the next, three things leak. Context leaks: the second desk knows less about the customer than the first, and the fifth knows almost nothing. Urgency leaks: the man who took the customer’s call felt it, the man who received the file did not, and by the fourth handoff the file is one of a pile. Ownership leaks: at every desk there is someone who did his part correctly, and at the end there is a machine on the floor that cannot ship and nobody who is wrong. I watched all three leak in 1989 through a chain of five men who were, every one of them, good at their jobs.

I have watched it leak since through chains of twenty, in every industry I have touched, and I will give you five, because the reorder book was a small company and you may be thinking that the problem was size. It was not. Size makes it worse, and so does sophistication.

**A hire.** A client needs an engineer, and the outcome is an engineer who joins and stays. In an ordinary recruitment firm the client’s brief goes to an account manager, who writes it down. The written brief goes to a sourcer, who has never met the client, and who searches on the words in the brief rather than the need behind them. The sourcer’s list goes to a screener, who has met neither client nor sourcer, and who filters on the words again. The screened list goes to a coordinator, who schedules interviews and knows nothing about engineering. The interviews are done by an interviewer who has read the brief once. The chosen candidate goes to an offer desk, which has never spoken to anyone involved and negotiates from a form. Seven desks. At each of them the brief is a little further from the client who gave it, a little less urgent, a little more somebody else’s. The client experiences one thing: three weeks of silence, then a candidate who is not quite what he asked for, then another three weeks. He does not know there were seven desks. He knows the outcome was not delivered, and he blames the firm, and he is right to, because the firm chose the seven desks.

**A loan.** A small business needs working capital, and the outcome is money in its account before the order it needs the money for is lost. The application goes to a relationship manager, who knows the customer. It goes to a credit analyst, who knows the numbers and not the customer. It goes to a risk committee, which knows the policy and neither. It goes to legal, which knows the documents. It goes to operations, which disburses. Six desks, six weeks, and a yes that was knowable on the first morning by anyone who held the customer, the numbers and the policy in one head. The customer, meanwhile, has lost the order, and takes the money from a lender who answered in a day, at twice the rate, and the bank has lost a customer it will never know it lost, because no desk in the chain will ever learn that the order was the point.

**A discharge.** A patient is well at nine in the morning, and the outcome is that she is home by lunch. The doctor writes the order. The nurse prepares the file. The pharmacy prepares the medicines. The billing desk prepares the account. The transport desk arranges the car. Each waits for the one before, each has twenty other patients, each does its part correctly, and the patient who was well at nine leaves at six, having occupied for nine hours a bed that another patient was waiting for in a corridor. Nobody was slow. The chain was long.

**A claim.** A car is damaged, and the outcome is that it is repaired and the owner is paid. The claim is registered by one desk, assessed by a second, surveyed by a third, approved by a fourth, and paid by a fifth, and at every desk a document is requested that the previous desk already had, because the desks do not share a head. The owner, who was promised a decision in seven days, receives one in thirty, and the insurer, which has spent more on the five desks than on the repair, wonders why its customers do not renew.

**A licence.** A citizen needs a permission from the state, and the outcome is the permission. I will not count the desks, because you have stood in that queue and you know. I will say only that the state is the purest form of the workaround, because it was designed by people who assumed the constraint was eternal, and so it cut every outcome into the maximum number of pieces that could be held by the minimum ability, and called the result accountability. It is not accountability. It is the leak made into a building.

Now the owner’s unit. Count the joints.

In the United States in 2014 there were 23.8 million managers, first-line supervisors and administrators, which is one for every 4.7 employees.[^6] Their work, honestly described, is the carrying of outcomes across desks: reviewing, approving, coordinating, reporting, reconciling what one piece did with what the next piece needs. Hamel and Zanini put the cost of this excess at more than three trillion dollars a year, about seventeen per cent of GDP, and estimated that a sixth of every non-manager’s time went on the internal compliance the managers required.[^7] Microsoft, reading the telemetry of its own tools, found in 2023 that across the Microsoft 365 applications it measures, fifty-seven per cent of recorded activity time went to meetings, chat and email, and forty-three per cent to creating anything.[^8] Not all communication is waste, but in the tools that record it, more than half of the time is the sound of the joints. And Gallup, which has measured it for a quarter of a century, finds that about one worker in five, worldwide, is engaged in their work at all, and that the two-point fall in that figure in 2024 alone cost the world economy, by its estimate, four hundred and thirty-eight billion dollars.[^9] That is intention leaking, measured, and the pattern repeats in every region Gallup measures, because the structure that leaks it is the same.

Put the three together and the twentieth-century company reveals itself: an organisation in which one person in five manages, in which more than half of everyone’s time is spent on the joints between desks, and in which four people in five have stopped wanting the outcome. It was a machine for producing outcomes that was mostly made of the joints between its parts, and most of what it spent was spent on the joints.

I want to be fair to the workaround, once, and then I will not be fair to it again. It built the modern world. There was no other way to make a pin, a car or a loan at scale with the people and tools available, and the men who designed it were not fools; they were the best minds of their time answering a real constraint honestly. Dhawan carrying the plant in his head was not a failure of imagination. It was the limit of a head. The department deserves the respect we give to any tool that did its job for two centuries. What it does not deserve is to be mistaken for the job.

## What the MBA taught, and why it is now wrong

I said I would say what the business school taught me. Here it is, discipline by discipline, with what each one is when the constraint is removed. I am not attacking the schools. I am describing what they will have to become, because every one of these disciplines was a science of the cut, and the cut is gone.

**Organisational design** taught where to make the cut: by function, by product, by geography, by matrix, and the trade-offs between them. It is a rich and subtle subject, and every word of it assumes that the outcome must be divided. When one system holds every skill, there is one design, and it is the outcome itself. Organisational design becomes the design of chains, which is the tuner’s craft, and the design of playbooks, which is the master’s. The question of where to cut has no answer because there is nothing to cut.

**Span of control** taught that a manager can hold five to nine direct reports, and that the pyramid follows from that number: layers equal to the logarithm of headcount. The number was true. It was the limit of a human’s attention over other humans doing pieces of work she could not see. When the work is visible as a floor of tiles and the only things that reach the manager are the things that need her, the span is a thousand, and the pyramid is two layers tall, and it will not grow a third because there is nothing for a third layer to carry.

**Functional strategy** taught marketing strategy, operations strategy, HR strategy, IT strategy, each with its own professors, each optimising its own piece. There are no functions. There are Outcomes, and each has one strategy, which is the result it exists to deliver and the budget it has to deliver it. The work that functions used to do divides cleanly in two. Most of it, a compliance check, a credit check, a quality check, the screening of a hire, is a skill used in producing someone else’s result, and a skill belongs inside the chain as an agent, not in a department beside it. A little of it, raising capital from the market, say, produces a result in its own right, and that becomes an Outcome like any other, with its own vectors and its own master. Functional strategy becomes one question asked case by case: is this a result, or a skill? The answer decides whether it is an Outcome or an agent, and no answer gives it a department.

**Human resources** taught recruitment, performance management, compensation and development as four systems for keeping a large number of people of ordinary intention pointed roughly at a result they could not see. The collapsed company needs very few people, selected for extraordinary intention, very highly paid, whose attention is read continuously from the work they do and never from watching them. Recruitment becomes selection for intention, which is a different search. Performance management becomes the daily review, which the aligner holds herself. Compensation becomes pay that follows the vector. Development becomes deeper, not higher, because there is no higher. Of the four systems, one survives in a new form and three dissolve.

**Operations management** taught the handoff: process mapping, queueing, bottlenecks, lean, six sigma, the whole discipline of making the pieces flow between desks with less waste. It was the most honest of the disciplines, because it alone admitted that the joints were where the cost was, and it spent fifty years making the joints a little cheaper. Operations management becomes the chain, which has no handoffs because it has no desks, and the gates, which are the only places the flow stops, and the stall alarm, which Ford had and the office never did. What remains of the discipline is the automation line: the question of which step a machine can hold to standard today. That question moves every year, and it is the tuner’s.

**Culture and leadership** taught that an organisation is held together by shared values, that a leader’s job is to inspire, and that engagement is the measure of both. Culture was a substitute for visibility. When a manager could not see the work, he needed people to share his values so that they would do the right thing unseen. In the collapsed company every vector is visible in a cockpit, every cockpit in a control room, every Outcome in the command centre, and nothing is done unseen, so nothing needs to be held together by shared feeling. What holds it is alignment, pay and the gates. Leadership becomes keeping, which is a quieter and rarer craft, and inspiration is replaced by selection: you do not inspire intention into people, you find the people who have it and pay them for it.

**Performance measurement** taught the balanced scorecard, the KPI, the monthly review, the quarterly report. All of it was a way of reconstructing, after the fact and at a distance, what the work was doing, because nobody could see the work. When the work is one live truth as of the last minute, flowing from cockpit to control room to command centre with nothing summarised by a person on the way, there is no report, because there is nothing to reconstruct. The board sees what the Keeper sees, at the same moment, and the entire apparatus of reporting goes, with the departments that wrote it.

Seven disciplines, and every one of them a science of managing the cut. I am not saying the schools taught nothing. I am saying they taught, brilliantly, the management of a constraint, and the constraint has been removed, and a management education now has to begin where this chapter begins: with the fact that no one person could hold the whole, and the fact that something now can.

## The constraint is gone

Now the thing that changes everything, stated as plainly as I can state it.

The constraint that created the department was that no single person could hold every skill an outcome needed. In 1989 I removed a small piece of that constraint with a program that held what four men had held between them, and four desks went quiet. Nobody was let go; all four found better work in the same building. The program could hold arithmetic. It could not hold judgement, or words, or a conversation with a customer, so the chairman kept his week and the firm kept its sales office and its purchase office and everything else that needed a human skill. The collapse went as far as the machine could hold, and stopped.

The machine can now hold almost all of it. A chain of AI agents holds the skills of a researcher, a writer, a caller, a seller, a negotiator, a coordinator and a clerk. It reads a brief and searches on the need rather than the words. It screens, schedules, interviews, drafts an offer and chases a signature, and it does each of these to the standard of a competent professional, around the clock, in the customer’s time zone, without the brief degrading between steps, because there are no steps between which it can degrade. The seven desks of the recruitment firm are one system. The six desks of the loan are one system. The five desks of the claim are one system. Where a skill is not yet held to standard, and there are such skills, the line sits a little lower this year and moves up next year, and it has never once moved down.

The capability is not a forecast. It is being measured. When seven hundred and fifty-eight consultants at one of the world’s most expensive firms were given an AI assistant for ordinary consulting tasks, those with the assistant completed twelve per cent more tasks, twenty-five per cent faster, at forty per cent higher quality than those without, and the consultants who had been weakest gained the most.[^10] The same experiment carried a warning the doctrine takes seriously: on one task chosen to lie just outside what the model could do, the consultants who used it were less likely to reach the right answer than those who did not. The line moves, and it is not yet everywhere; a gate exists for the side of it the machine has not reached. That was 2023, with a tool that was a toy beside what exists now, and with the agent assisting a human at a desk rather than replacing the desk. When a payments company put an AI assistant in front of its customers in 2024, it handled two-thirds of all customer-service conversations in its first month, two million three hundred thousand of them, in thirty-five languages, doing the work of seven hundred full-time agents, with resolution times cut from eleven minutes to two and repeat enquiries down by a quarter, at a customer-satisfaction score equal to the humans it replaced.[^11] Seven hundred desks became one system in a month. That is a collapse, measured, and it is two years old.

I will argue the claim properly later in the book, against every objection I have heard, and I will answer the loudest of them before this chapter ends. For now the consequence follows from the statement, and the consequence is this.

If the skills sit in one system, there is no reason to cut the outcome into pieces. If there is no reason to cut it, there is no reason for the desks that held the pieces. If there are no desks, there is nothing for the layers above them to carry across. The department, the function, the handoff and the manager were all built on one constraint, and the constraint has gone, and they are standing on nothing. They will stand for a while longer, because buildings stand for a while after their foundations go. Then they will not.

What is left when they go is small and precise. An outcome, whole, the way the customer feels it. A system of agents that carries it from first intention to final result. And one human, whose job is not any of the skills, because the system has those, but the one thing the system does not supply: intention. Someone who wants this outcome, for this customer, today, and will stand at the gates and say go, and will sign. The whole company collapses towards that person, and the next chapter is about her.

## The excuses

I have put this argument to owners for three years, and I have heard every reason for not acting on it. There are seven. None of them survives, and I will take them in the order they are usually offered.

**“My industry is different.”** It is not. Every industry has a connected part and a hands part. The connected part is everything that reaches a screen: every brief, application, claim, order, record, payment, message and signature. The hands part is the surgery, the shop floor, the site. The collapse applies to the connected part completely, today, in every industry without exception, and the hands part shrinks every year as the hands connect. If your industry is different, it is different only in the size of its hands part, and the hands part is not where your departments are. Your departments are in the connected part. They are the connected part.

**“Our people are our greatest asset.”** Then pay them like it. The collapsed company pays its people very highly, far above the market, because it needs a tenth as many, and the ones it keeps are the ones whose intention held. The old company said people were its greatest asset and paid them like a cost, because it needed so many of them that it could afford nothing else. If you mean the sentence, the collapse is how you honour it. If you do not mean it, the collapse is how you stop pretending.

**“AI makes mistakes.”** So did Verma, fifty times a month, and so did the register, and so did every desk in every chain I have described, and the mistakes were invisible because nobody could see across the desks to find them. The chain’s mistakes are visible, counted, logged and replayable, and they are caught at gates by a human whose only job is to catch them, and the chain does not make the same mistake twice because the tuner fixes it that day. The question is not whether the machine makes mistakes. It is whether you can see them. You could never see the desks’. You can see the chain’s.

**“We are regulated.”** Then you have an advantage, because a regulator wants exactly what the chain provides: every decision logged, every authority limit explicit, every commitment signed by a named human under a stated permission, and an audit trail that can be replayed to the agent that produced any output. The old company gave its regulator a filing cabinet and a promise. The collapsed company gives it the engine. The signature rule in this book exists because of regulation, not despite it.

**“We tried reengineering and it failed.”** You tried the collapse without the machine, and it failed because the seven heads re-formed into seven desks, as I have said. The thing that was missing in 1993 is the thing that is present now. If reengineering taught you anything, it should have taught you that the chart was never the problem, and that the moment the constraint went, the thing Hammer wanted would work. The constraint has gone.

**“Our customers want to deal with humans.”** Your customers want the outcome. They want the engineer who joins, the money in the account, the car repaired, the permission granted. They never wanted the seven desks; they tolerated them because every supplier had them. When one supplier does not, your customers will not ask whether a human read their brief at desk four. They will ask why yours took three weeks. And when something must be signed, a human will sign it, and the customer will know her name.

**“We are too small,” or “we are too big.”** The thirty-person plant collapses its office into one vector and keeps its floor until the floor connects; the ten-thousand-person company collapses into a dozen Outcomes of a thousand vectors each, held by a dozen masters, under one Keeper. Size changes the number of vectors. It changes nothing else. The reorder book was a small company. Klarna, in chapter 19, was not.

There is an eighth reason, which nobody offers aloud and everybody means, and I will offer it for them. “I built the departments. The departments are my career, my status, my understanding of what a company is. If they go, what am I?” You are the Keeper, if your intention holds. If it does not, you are what the department was: a workaround, and the next chapter is not about you.

## What happens to the people

You will want to know what happens to the people, and I will not make you wait for chapter 13 to hear the shape of it. There will be far fewer of them, and they will be paid far more, because what you are buying from them has changed from skill, which is now cheap, to intention, which has always been rare. The rest will be released. In 1989 the collapse gave time back, because the manual frontier was vast and four freed men walked across it, inside the same company, to new desks and the next unautomated job; nobody was let go. There is less of that frontier now and there will be less every year, and a book that told you otherwise would not be worth your reading. What the doctrine owes those people, and what the companies that collapse owe them, I take up where it belongs, in chapter 13, in the month the headcount falls, and again where an Outcome dies. For now, only this: the alternative to collapsing deliberately, around the people whose intention holds, is being collapsed carelessly, around whoever is left.

## The company down the road

Because here is what the owner reading this has to understand, and it is not a prediction. It is a description.

Somewhere near you, perhaps already, there is a company that does what yours does and has never built a department. I am going to show you a day in it, and then its customer’s day, and then its books, because you will not believe it until you have seen all three.

It has five people. One of them, call her the aligner of Winning, holds the outcome of winning customers in your market. One holds the outcome of delivering what was won. The other three are the same two things for a second market and a tuner who stands by for all four, which the doctrine itself counts as a gap: chapter 5 gives every vector a tuner of its own. Each has a chain of agents: for Winning, an agent that finds prospects, one that qualifies them, one that writes the first approach, one that handles the replies, one that drafts the proposal, one that negotiates within a limit she set, and one that chases the signature. The chain works all night in your customers’ time zones. At six in the morning she puts on a headset, and an agent that answers for the whole chain tells her what happened: five hundred prospects found, four hundred and sixty qualified and waiting at a gate, thirty replies drafted and waiting at another, one proposal accepted with a change in terms and held for her signature. She samples, releases, holds one, signs one. Forty minutes. She does none of the steps. At noon and at six she does twenty minutes more. The rest of the day is hers. There is no manager above her, because there is nothing between her and the master to manage. There is no function beside her, because there are no functions. There is no meeting, because there is nobody to meet.

Now her customer’s day, which is your customer’s day, because they are the same person. On Monday morning he describes what he needs, once, to one chain, in his own words. By Monday afternoon the chain has understood the need behind the words, because it searched on the need and not the words, and has sent him three candidates, or three terms, or three dates, whichever his outcome requires. On Tuesday he replies. His reply does not go to a desk; it goes to the chain, which has the whole of his brief and the whole of his history in one place, and the chain answers within the hour, in his time zone, and the answer is as new as that hour. On Wednesday he asks for a change. A human reads it, decides, and signs, and he knows her name, and she is the same human who will answer him next month, because the outcome is hers. He does not know, and will not care, that there were no desks. He knows that the thing he asked for arrived in three days with no leak, and that the last time he asked your company it arrived in three weeks with several. He is not going to come back to you. Not out of anger. Out of arithmetic.

Now the books. Your company employs forty-five people to produce what those five produce, because that is what seven desks per outcome, two markets, two outcomes and the managers to carry across them come to. Your forty-five are paid at the market, with a building, with the sixteen per cent of their time that goes on the compliance the managers require, and with one in five of them managing. The five down the road are very highly paid, and still their whole payroll is a fraction of yours; they have no building, no managers, no reporting, and a chain whose running cost is a fraction of one salary. They are cheaper than you before a single outcome is counted. Then count the outcomes. Their customer gets a brief that does not leak, an answer that is never a month old, and a human who signed it, and he stays. Yours leaves. Their revenue per person is nine times yours before the leak is counted and more after. Their margin is not better than yours. It is a different kind of number.

That company is not better than yours. Its people are not cleverer than yours. It simply never had the constraint, and so it never built the workaround, and so it has none of the joints that you are paying for. Everything you spend on the joints, it spends on the outcome. That is the entire difference, and it is enough.

## Five years

I will put a date on it, because a doctrine that will not be held to a date is not a doctrine.

I expect that within five years every connected outcome in every industry will be run this way by someone. Not by everyone; by someone, in your market, selling to your customers, at a tenth of your cost, with none of your leaks. In some industries it is already so, and you have not noticed, because the company doing it is small and quiet and does not look like a competitor; it looks like five people. It is the most dangerous competitor you have ever had, because you cannot match it by trying harder. You cannot manage your way to it. You cannot reengineer your way to it; that was tried. You can only collapse to it, and the collapse takes a year, and the year has to start before the five people reach your customer.

Everything I have told you in this chapter follows from one fact and its negation. For two hundred years, no one person could hold the whole, and so we cut it, and the cut became the company, and the company became a science, and the science became a profession, and the profession became your career. Now something can hold the whole. Not a person; a system, with a person at the gate. The cut has no reason left. The departments are standing on nothing, and the people who built them, who are honest people answering an honest constraint, are about to learn what Mr Malhotra learned when he picked up my drawing of his own projection sheet: that the thing they do with their best week every month is not, after all, the thing that only they can do.

You can be that company. The departments you have are not your company; they are the scaffolding that was put up because the company could not stand without it, and the company can stand without it now. Take the scaffolding down yourself, in the order the method describes, around the people whose intention you would pay very highly for, and run old beside new until the registers go quiet. Or leave it up, and wait for the company down the road to do it for you.

Collapse your organisation now, or your organisation will collapse.

[^1]: Adam Smith, *An Inquiry into the Nature and Causes of the Wealth of Nations* (1776), Book I, Chapter 1. Smith’s figures: ten men, about eighteen distinct operations, upwards of forty-eight thousand pins a day; a man working alone “could scarce, perhaps, with his utmost industry, make one pin in a day, and certainly could not make twenty.”

[^2]: Frederick Winslow Taylor, *The Principles of Scientific Management* (1911).

[^3]: Henry Ford’s Highland Park plant introduced the moving chassis assembly line in late 1913; Ford’s own account (*My Life and Work*, 1922, with Samuel Crowther) gives chassis assembly time falling from twelve hours and twenty-eight minutes to one hour and thirty-three minutes.

[^4]: Peter F. Drucker introduced “management by objectives” in *The Practice of Management* (1954) and the term “knowledge worker” in *Landmarks of Tomorrow* (1959).

[^5]: Michael Hammer and James Champy, *Reengineering the Corporation: A Manifesto for Business Revolution* (1993). The authors’ own estimate, given in the book, was that fifty to seventy per cent of reengineering efforts failed to achieve the dramatic results intended; later studies put the failure rate at seventy per cent or higher.

[^6]: Gary Hamel and Michele Zanini, “Excess Management Is Costing the U.S. $3 Trillion Per Year,” *Harvard Business Review*, September 2016: 23.8 million managers, first-line supervisors and administrators in the US workforce in 2014, one for every 4.7 employees.

[^7]: Hamel and Zanini, as above: excess bureaucracy costing more than $3 trillion a year, about 17 per cent of GDP; roughly 16 per cent of non-managerial employees’ time spent on internal compliance.

[^8]: Microsoft, *Work Trend Index Annual Report: Will AI Fix Work?*, May 2023: across Microsoft 365 users, 57 per cent of measured time went on communicating (meetings, email and chat) and 43 per cent on creating (documents, spreadsheets and presentations). The figure covers activity in Microsoft’s tools, not every hour of the working week.

[^9]: Gallup, *State of the Global Workplace* reports, 2024 and 2025 editions: global employee engagement 23 per cent in 2023 and 21 per cent in 2024; manager engagement 27 per cent in 2024; the two-point fall in global engagement in 2024 estimated to have cost US$438 billion in lost productivity (Gallup estimates that a fully engaged workforce would add US$9.6 trillion a year).

[^10]: Fabrizio Dell’Acqua, Edward McFowland III, Ethan Mollick, Hila Lifshitz-Assaf, Katherine Kellogg, Saran Rajendran, Lisa Krayer, François Candelon and Karim Lakhani, “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality,” Harvard Business School Working Paper 24-013, September 2023, published in *Organization Science* 37, no. 2 (2026): 403–423, doi 10.1287/orsc.2025.21838: 758 BCG consultants; those using GPT-4 completed 12.2 per cent more tasks, 25.1 per cent faster, at over 40 per cent higher quality; the lowest-performing half improved by 43 per cent. On the task designed to fall outside the frontier, consultants using AI were 19 percentage points less likely to produce a correct solution than those working without it.

[^11]: Klarna, press release of 27 February 2024, “Klarna AI assistant handles two-thirds of customer service chats in its first month”: 2.3 million conversations in the first month, two-thirds of all customer service chats, across 23 markets and 35 languages, equivalent to the work of 700 full-time agents; average resolution time under two minutes against eleven previously; a 25 per cent drop in repeat enquiries; customer satisfaction on par with human agents; an estimated US$40 million profit improvement in 2024.

================================================================
# Chapter 3: Where Outcomes Leak
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-3-where-outcomes-leak

There is a man in every company I have ever run who has already left and has not yet told anyone.

I have met him at Orange, in the years when Naval and I had a dozen programmers and every one of them mattered. I met him at WLCI, when there were hundreds of people across the country and I could not know them all, and still I knew him when I saw him. I meet him at GoodSpace. He is not always a man, and he is never the same person twice, but he is always recognisably himself, and I have become very good at spotting him, because I have had thirty-five years of practice.

He has decided to go. He has not resigned. Perhaps he is interviewing elsewhere, perhaps he is only thinking about it, perhaps he has not admitted it even to himself. But the decision has been made somewhere inside him, and from that moment it is written on the wall for anyone who knows how to read it. He starts taking leave: a Monday here, a Friday there, a family function that needs two days. His numbers dip, not off a cliff, but steadily, week after week, in a way that never quite justifies a conversation. He becomes slow to answer. Messages that used to come back in ten minutes come back in a day. He stops volunteering. He stops arguing, which is the surest sign of all, because a man who argues with you about the work still cares about the work. And his face changes. I cannot describe it better than that, and every manager reading this knows exactly what I mean. The light has gone out of it. He is present, and he is not there.

Now set him beside a different man, who has had a bad week. His mother has been ill, or his child, or he has simply run out of road for a few days, as everyone does. His numbers dip too. He is slow to answer too. He may even take a day off. From the outside, for a week, the two men look the same. But they are not the same, and the difference between them is the most important thing in this chapter. The second man bounces back. Next week, or the week after, he is himself again, arguing, volunteering, answering in ten minutes, and the numbers recover as if they had never fallen. The first man does not bounce back, because there is nothing to bounce back to. His wanting has gone. It has not dipped; it has left, a few weeks ahead of the rest of him.

The second man has had a bad week. The first man has switched off. One is a fall in attention. The other is the end of intention. And the old company, which could see neither directly, could not tell them apart until the resignation letter arrived, by which time the outcome he was holding had been leaking for months.

This chapter is about that leak: what it is, why it happens, why the department made it worse, and why, for the first time, it can be seen before it does any damage.

## Three terms

Begin with first principles.

Every result a company produces, a machine shipped, a loan approved, a candidate placed, a patient discharged, is the product of three things. Ability: can the person or the system do the work to standard? Intention: do they want this result, for this customer, and will that wanting hold? Time: will the hours be there, at the hour they are needed, for as long as it takes?

**Outcome = Ability × Intention × Time.**

It is multiplication, not addition, and that is the whole point of writing it down. A brilliant man who does not care produces nothing. A devoted man who cannot do the work produces nothing. A man who can, and wants to, and has no hours, produces nothing. Any term at zero makes the outcome zero, however large the other two. The man who has switched off has lost none of his ability and is still paid for all of his time. His outcome is collapsing anyway, because the middle term has gone.

The old company knew this in its bones and spent most of its ingenuity trying to keep all three terms above zero across hundreds of people at once. It hired for ability, because ability could be tested. It bought time, by the hour, because time could be counted. And it tried to manufacture intention, with targets, bonuses, appraisals, offsites and culture, because intention could be neither tested nor counted, and it was the term that kept leaking out.

## Ability was never scarce

Ability is the term every company overvalued.

It is the capacity to perform a step to standard: write the letter, read the balance sheet, screen the CV, make the call. It is what a degree certifies and an interview tests, and the entire apparatus of hiring was built to find it. It was also, long before AI, the term that mattered least. Most work in most companies does not need a great deal of it. Ramesh copying registers did not need genius; he needed to copy accurately for a week. A sourcer searching on a brief does not need brilliance; she needs to search well today and again tomorrow. For the great majority of steps in the great majority of outcomes, adequate ability was always available and always cheap.

Look at the man who has switched off. Did he lose his ability? He did not. He is exactly as capable on the day he resigns as on the day he was hired, often more so. Every outcome I have watched fail through a person like him failed with full ability present in the room. That alone should have told us where the scarcity was.

## Time was bought, not applied

Time is the term the industrial company understood best, because it was the term it bought.

The factory bell, the shift, the eight-hour day and the forty-hour week were all instruments for buying time, and the entire wage system rested on the assumption that time bought was time applied. It never was. When a software company that tracks how knowledge workers use their computers analysed one hundred and eighty-five million working hours of its users’ data, it found that the average knowledge worker logged two hours and forty-eight minutes a day of device time classed as productive.[^12] That is a measure of a screen, not of a working day, but it points where the rest of my own experience points: much of the eight hours is present and not applied.

And even the applied hours are not constant. When researchers followed four thousand caregivers in thirty-five hospitals through thirteen million moments when they should have washed their hands, they found compliance fell by almost nine percentage points between the start of a twelve-hour shift and its end, among people who knew precisely what was at stake.[^13] When others studied more than a thousand parole decisions by eight experienced judges, they found the rate of favourable rulings fell from about sixty-five per cent at the start of each session towards zero by its end, and jumped back to sixty-five after a meal.[^14] The precise causes are argued over, and should be. The shape is not. Human attention rises and falls through the day, in the most skilled people, doing the most serious work.

The old company knew all of this without the studies. Its answer was to buy more hours than it needed and more people than it needed, as insurance against the hours that would not be applied, and to call the surplus headcount. There was no other answer, because the work had to be done at the work’s hour, not the worker’s, and the only instrument for measuring applied time was a supervisor watching, whose own time was leaking just as fast.

## Intention is the scarce term

Intention is the third term, and it is the one this book is about.

Intention is commitment to the outcome. Not to the task, not to the hours, not to the employer: to the result that someone on the other end is waiting for. It is what Mr Malhotra had when he shut his door for a week. It is what Dhawan had when he lay awake wondering what would happen to the company if he forgot. When it is present, it makes a person pick up the call that is not in his job, notice the part on the wrong page, stay the extra hour because the customer rang and then stopped ringing. When it is absent, it produces a man who has done exactly what the system asked of him, and a machine on the floor that cannot ship.

Intention is not motivation. Motivation is the industrial word, and the industrial company tried to supply it from outside, with money and fear. Motivation is a push. Intention is a direction, held from inside, and it either points at the outcome or it does not. You can motivate a man to be at his desk. You cannot motivate him to want the customer’s machine to ship. That wanting is there or it is not, and it varies from person to person far more than ability does. Most people have adequate ability for most jobs. Very few have an intention that holds, day after day, when nobody is watching, towards a result that is not their own.

And now the distinction the opening of this chapter was built to make. Intention and attention are not the same thing.

**Intention** is the direction. It is whether the person wants the outcome at all. It changes rarely, and when it changes, it changes for good: the man who has decided to leave does not decide to stay because he had a good lunch.

**Attention** is intention showing up today. It is the magnitude, hour by hour, of a direction that is still there. It varies constantly, in everyone, as the judges and the caregivers show. It falls at four in the afternoon and rises after tea. It falls in the week a mother is ill and rises when she is home. The man who had a bad week lost attention. The man who switched off lost intention.

Everything in the collapsed company depends on keeping these two apart. You hire for intention, because it is the scarce thing and it does not come back once it is gone. You measure attention, because it is how intention shows itself, and because its long fall is the first sign that intention has left. And you never confuse a dip in attention with an end of intention, because that is how a company loses its best people in their worst week.

## Written on the wall

I said I had become good at spotting the man who has switched off. For most of my career I thought this was instinct, the kind of thing a founder learns by being burned. It turns out to be measurable.

Three researchers set out to find whether the behaviour of people about to quit could be seen by the people around them. They asked managers, students and employees what changes they had noticed in colleagues before those colleagues left, boiled what they heard down to thirteen, and then tested them. The thirteen are a portrait of the man in the opening of this chapter, drawn by strangers: productivity falls; he acts less like a team player; his attitude turns; he puts in less effort; he focuses less on the job; he expresses dissatisfaction with the job and with his supervisor; he leaves early more often; he loses enthusiasm for the mission; he shows less interest in customers; he is less interested in pleasing his manager; he is less willing to commit to long-term timelines; he is more reluctant to take on new projects.[^15] Then they asked managers to rate their own people on these behaviours, and waited. Thirteen months later, the people whose managers had seen these signs were, by the authors’ own account, roughly twice as likely to have left.[^16]

Twice as likely, and seen a year in advance, by managers with nothing but their own eyes. Every one of those thirteen signs is a change in the work, and most of them are changes a system could count. Leaving early is in the attendance record. Productivity is in the numbers. Slowness is in the response times. Reluctance to commit to long timelines is in what he says yes to. Less interest in customers is in how quickly the customer’s message is answered. The face is the only sign a system cannot read, and it does not need to, because the rest are already on the wall.

And absence, the leave on a Monday and a Friday that I have watched for thirty-five years, has been studied for decades: people who are absent more are, reliably, the people who leave.[^17] The founder’s instinct and the researcher’s scale are reading the same signal. The only difference is that the founder reads it too late, because he reads it with his eyes, once in a while, and only for the people he happens to sit near.

## The switched-off company

Now multiply him.

The man who has switched off without resigning has a name now. In 2022 the press called him the quiet quitter, and Gallup, which has measured engagement for a quarter of a century, went looking for him in the American workforce. It found that half of all employees were not engaged: present, paid, doing the minimum, still on the payroll; engagement is Gallup’s measure, and reading it as intention gone is mine. Another eighteen per cent were actively disengaged.[^18] Half the workforce, in other words, was the man in the opening of this chapter, and the company had no instrument that could see it, because the company could see only the resignation letters, and most of these people had not written one.

When they do write one, the cost is counted. Gallup puts the cost of replacing an employee at between half and twice his annual salary, and the cost of voluntary turnover to American business at about a trillion dollars a year. And then it adds the two numbers that matter most. Fifty-two per cent of the people who left said their manager or their company could have done something to keep them. Fifty-one per cent said that in the three months before they left, nobody had spoken to them about how they felt about their job or their future at all.[^19]

Three months. That is the window in which the man was switched off and still there, his intention gone and his outcome leaking, and in half the cases nobody noticed, or noticed and said nothing, because nobody could see the work closely enough to be sure, and nobody wanted the conversation without being sure.

## Why the department made it worse

The department did not create the man who switches off. People have always lost their wanting, and they always will. But the department made him invisible, and made him expensive, and made him more likely.

It made him invisible because it cut every outcome into pieces that could not see each other. The man at desk four was holding one piece of a result whose customer he had never met. Nobody downstream could see his piece until it arrived, late, and nobody upstream could see it at all. His numbers were his own numbers, reported once a month, mixed with the numbers of everyone else on his floor. His leave was in one register, his output in another, his response times nowhere. To see that he had switched off, a manager had to put three registers together in his head, as Rao put the shelves and the register together, and most managers did not, because they had forty other people and no time.

It made him expensive because a switched-off man in a department is holding a piece of a hundred outcomes at once. When his intention leaves, it leaves every outcome that passes through his desk, and each of them leaks a little, and the sum of the little leaks is the three weeks of silence a customer experiences and never understands.

And it made him more likely, because the department was designed on the assumption that intention would fail, and the design guaranteed that it would. Cut an outcome into seven pieces and give each to a specialist, and you have told each specialist that the outcome is not his. You have asked him for ability and time, and put the intention somewhere else, in a manager, in a target, in a process, because you could not trust it to survive in seven heads at once. You were right; it could not. But in taking it out of the seven heads you removed the one thing that held each of them to the work, and then wondered why so many of them drifted. Intention needs a target with a face. The department gave every desk a target without one.

## Signal and noise

There is a way of saying this that an engineer will recognise, and I am an engineer, so I will say it that way too.

Every outcome is a signal: what the customer wants, travelling from the first intention to the final result. Everything that does not serve that signal is noise: the handoffs, the reports, the meetings about the meetings, the hours present but not applied, the parts of a job that exist only because the job was cut from a larger one and the cut left edges. The department is a low signal-to-noise structure by nature, because it is organised around the noise, around skills, functions and hours, and the signal has to find its own way through.

The switched-off man is the noise floor rising. While he is there, his desk still produces output, still sends reports, still attends meetings; it makes all the sounds of work. But the signal through his desk is fading, and in a noisy structure nobody can hear it fade until it is gone.

A vector is a high signal-to-noise structure, because it is organised around the signal: one outcome, one chain, one human who wants it. There is almost no noise in it, because the noise was mostly the price of carrying skills in heads that could not hold them all, and the skills are now in a machine. In a quiet room, a fading signal is the loudest thing there is.

## The machine holds the baseline

Now what the machine changes, which is simpler than people expect.

Of the three terms, the machine supplies two completely and the third not at all.

It supplies ability, to the standard of a competent professional, and where it does not yet, it will; the line moves every year and has never moved down. More to the point, its ability does not vary. It does not have a bad week. It performs a step the same way at four in the afternoon as at nine in the morning, on the thousandth file as on the first. The judge’s sixty-five per cent falling to zero, the caregiver’s nine points, the man’s numbers drifting through his last three months: none of these happen to a chain.

It supplies time, all of it. A chain works around the clock in the customer’s time zone, and every hour it works is applied, because it has no attention to withdraw. Two hours and forty-eight minutes of productive work in an eight-hour day becomes twenty-four hours of it, and the surplus headcount that was insurance against unapplied time is no longer needed.

What it does not supply is intention. A chain does not want the customer’s machine to ship. It will do everything it was built to do to ship it, to standard and on time, but when something happens that it was not built for, a customer who stops ringing, a part on the wrong page, a brief whose words say one thing and whose need says another, it does not notice, because noticing is a function of wanting. The gap between what the chain was built to do and what the customer needs today is where intention lives, and it is the one place the machine cannot go.

So the machine holds the baseline and the human supplies the peaks. The chain carries ability and time at a level that does not tire. The human carries intention, which is the only thing she is now for, and applies it where it matters: at the gates, where the chain’s work is released or held; at the signature, where the firm commits; and at the moments, several times a day, when something is not quite what the chain expected and someone has to care. She does not supply ability, because the chain has it. She does not supply the hours, because the chain has those too. She supplies wanting, at the hours when her attention is highest, and the chain covers the rest.

Which means that in the collapsed company there is one human per outcome, and the one thing she brings is the one thing the switched-off man has lost. A department could carry a switched-off man for months, because his ability and his hours were still there and they were most of what his desk was for. A vector cannot carry him for a week. If the aligner’s intention goes, the vector has nothing human left in it at all.

## What follows

Three consequences follow, and the rest of the book works them out.

The company stops hiring for ability. The human still needs enough ability to judge the chain’s work, and that is not nothing. But ability is no longer the scarce term, and the selection apparatus built to find it, the degree, the test, the technical interview, is pointed at the wrong thing. The company hires for intention instead: for the person whose wanting has held, over years, when nobody was watching, towards results that were not her own. That is a different search, and the chapter on selection is about how to run it.

The company pays for intention, and pays a great deal. If one high-intention human and a chain of agents produce what a department of twenty produced, the one human is worth to the company something like what the twenty were worth. The doctrine pays that person very highly, far above the market rate, and deliberately so, because at that price the company can choose from the very few whose intention does not waver, and because at that output the price is still cheap. The old company could not afford to pay for intention because it needed so many people that it could afford to pay only for ability and time. The collapsed company needs so few that it can pay for the one thing that was always rare.

And the company measures attention, never hours, and never the person. Attention is read from the work: how fast the gates are answered, whether the daily review was held, how many decisions are waiting, whether the quality of judgement calls is holding. These are signals in the outcome, not observations of a human being, and they are enough. They are, almost exactly, the thirteen signs the researchers found, minus the face.

## The engine sees it first

I will end where I began, with the man who has already left, but this time in a vector.

He is the aligner of a sales vector in an Outcome of four hundred. Three weeks ago, somewhere inside him, he decided to go. He has told nobody. He still puts on the headset every morning.

But the engine reads the work, and the work has changed. The gates that he used to answer within the hour of his chosen time are being answered at the end of the day, and twice this week not at all, so they defaulted open and the chain flowed on without his judgement. The daily review, which he held every morning for fourteen months, has been skipped three times in a fortnight. The samples he takes at the gates are smaller. The decisions waiting on him are older. He has taken two Mondays. None of these alone would turn a tile. Together, inside his tile, they cross the red threshold, and by the rule of the tile the outside of it has turned red.

The floor shows the master the red, and the vector’s own quartermaster, which has the history of every one of his days, shows her the difference between a bad week and a switched-off one. A bad week is a sharp dip and a recovery. This is a slope. It shows her the tile, and the slope, and the days.

So the master does the thing that, according to Gallup, half the people who quit say nobody did for them in their last three months. She calls him. Not to accuse him; the doctrine does not accuse anyone of anything the work has not shown. She tells him what the engine sees, and asks him what is going on. Perhaps it is a bad month after all, and he needs a fortnight, and the tuner holds the vector while he takes it. Perhaps he has an offer elsewhere, and she can match it, or cannot. Perhaps he has simply finished with this work, and says so, and is relieved to say so. Whatever it is, the conversation happens three weeks after the decision instead of three months, and before he resigns, not after.

And while they talk, the vector does not leak. His tuner has been his standby from the first day; she knows the chain as well as he does, and has been answering his late gates for a week already. If he goes, she holds the vector until a new aligner is chosen, and the customer, who never knew there was a man who had switched off, never feels the gap.

I learned to read that man by instinct over thirty-five years, and I read him too late, every time, because I read him with my eyes and only when I happened to be looking. The engine reads him from the work, every minute, for every vector, and it reads nothing else. It does not read his messages, his applications elsewhere, or his life. It does not need to. The work is enough, because the work is where intention shows itself, and where it is first seen to leave.

That is the leak, and that is where it is stopped. Most companies believe they are stopping it already, with a machine at every desk. The next chapter is why they are not.

[^12]: RescueTime, “Work-life balance study 2019,” based on 185 million hours of anonymised and aggregated working-time data from its users in 2018: knowledge workers averaged 2 hours and 48 minutes a day of productive work.

[^13]: Hengchen Dai, Katherine L. Milkman, David A. Hofmann and Bradley R. Staats, “The Impact of Time at Work and Time Off From Work on Rule Compliance: The Case of Hand Hygiene in Health Care,” *Journal of Applied Psychology* (2015): 4,157 caregivers in 35 US hospitals, about 13.7 million hand-hygiene opportunities over two years; compliance fell by an average of 8.7 percentage points from the beginning to the end of a typical 12-hour shift.

[^14]: Shai Danziger, Jonathan Levav and Liora Avnaim-Pesso, “Extraneous Factors in Judicial Decisions,” *Proceedings of the National Academy of Sciences* 108, no. 17 (2011): 1,112 rulings by eight Israeli parole judges over ten months; the proportion of favourable rulings fell from about 65 per cent to nearly zero within each session and returned to about 65 per cent after a food break. The interpretation has been disputed, notably on the ordering of cases (Weinshall-Margel and Shapard, *PNAS*, 2011); the pattern of decline within a session is the point here, not its single cause.

[^15]: Timothy M. Gardner, Chad H. Van Iddekinge and Peter W. Hom, “If You’ve Got Leavin’ on Your Mind: The Identification and Validation of Pre-Quitting Behaviors,” *Journal of Management* (2018), summarised by Gardner and Hom in *Harvard Business Review*, October 2016. Thirteen manager-observed behaviours; each one-point rise on the thirteen-item scale multiplied the odds of leaving voluntarily within thirteen months by about six and a half (odds ratio 6.48), over and above established predictors of turnover.

[^16]: Gardner, Van Iddekinge and Hom, as above. In the authors’ own summary, employees whose managers rated them at 4.2 or more on the thirteen behaviours had an expected probability of leaving about two times that of the typical employee, over and above established predictors of turnover; the paper’s odds ratios are larger, and the book uses the authors’ plainer figure.

[^17]: Atanu Mitra, G. Douglas Jenkins Jr. and Nina Gupta, “A Meta-Analysis of the Relationship Between Absence and Turnover,” *Journal of Applied Psychology* (1992): across studies, absence and later turnover were positively related, consistent with absence as part of a progression of withdrawal that ends in leaving.

[^18]: Gallup, “Is Quiet Quitting Real?”, September 2022, based on a survey of 15,091 US employees in June 2022: 32 per cent engaged, 50 per cent not engaged (“quiet quitters”), 18 per cent actively disengaged.

[^19]: Gallup, “This Fixable Problem Costs U.S. Businesses $1 Trillion,” March 2019: the cost of replacing an employee ranges from one-half to two times annual salary; voluntary turnover costs US businesses about $1 trillion a year; 52 per cent of voluntarily exiting employees said their manager or organisation could have done something to prevent their leaving; 51 per cent said that in the three months before they left, no manager or leader had spoken with them about their job satisfaction or future. Gallup’s 2024 update puts the two figures at 42 and 45 per cent.

================================================================
# Chapter 4: A Motor Where the Steam Engine Stood
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-4-a-motor-where-the-steam-engine-stood

In 1895 an electrical engineer from Columbia University, Professor F. B. Crocker, went to Taftville in Connecticut to see the newest thing in the world at work.

The Ponemah mill at Taftville had run its looms from engines in its basement. Now a hydroelectric station had been built at a dam upriver, at Baltic, and the power came down to the mill along overhead lines, which was the new thing, and it was remarkable: power made in one place and used in another, miles away, with nothing between the two but a wire on poles. Crocker followed the wire to where it arrived. The motors, he reported, were “located in the basement of the mill, near the engines which they replace.” They were belted to pulleys, and the pulleys were clutched to the mill’s shafts. One motor drove twelve hundred looms. The other drove five hundred.[^20]

To see what that meant you have to see a mill of the time, and the historian Warren Devine has drawn one. A single engine stood at the centre. It turned iron or steel line shafts, about three inches thick, which hung from the ceiling and ran the whole length of each floor; belts ran up through holes in the ceilings to carry the power to the floors above; the line shafts turned shorter countershafts, and every machine in the building was belted to a countershaft and stood, of necessity, in a row parallel to it. The whole network turned from the moment the engine was started in the morning until it was stopped at night, no matter how many machines were in use. If a shaft broke, a room stopped. If the engine broke, the mill stopped.

At Taftville the engine had been taken out of that picture and a motor put where it had stood. Everything else was left as it was: the shafts, the belts, the rows, the building shaped around them. The Ponemah mill had the power of the twentieth century and the floor of the nineteenth. Nobody there was foolish. The shafts were paid for and they worked. Devine’s judgement on the whole period is one sentence, and it is the subject of this chapter: “Replacing a steam engine with one or more electric motors, leaving the power distribution system unchanged, appears to have been the usual juxtaposition of a new technology upon the framework of an old one.”[^21]

I have stood in that basement more often in the last three years than I can count. It is not in Connecticut, and the machine in it is not a motor.

## The basement today

The first three chapters of this book argued that the department was a workaround for one constraint, that the constraint has gone, and that the outcomes of the old company leak at every joint between its desks. Almost every owner I have said this to agrees with some of it, and then tells me what his company is doing about AI. What he describes, nearly always, is a motor where the steam engine stood. It comes in four forms.

**A copilot at every desk.** Take the recruitment firm of chapter 2, with its seven desks between a client’s need and an engineer who joins. Give every desk an assistant. The account manager’s assistant turns the client’s call into a written brief in a minute. The sourcer’s assistant searches faster and more widely than she ever could. The screener’s assistant reads a thousand CVs an hour. The coordinator’s finds interview slots, and the offer desk’s drafts the offer. Every desk is faster, and every desk can prove it, with a number for its own step. And the client still waits three weeks, because the three weeks were never at the desks. They were between them, in the queue at each desk, in the brief that lost a little of the client at every pass, in the coordinator’s wait for the interviewer’s diary. Michael Hammer quoted an insurer, in 1990, which had estimated that an application spent twenty-two days in process and was actually worked on for seventeen minutes.[^22] A copilot at every desk turns the seventeen minutes into eight. It leaves the twenty-two days where they were.

**A chatbot in front of the same desks.** Take the bank of chapter 2. A conversational agent now takes the small business’s application at midnight, in the owner’s own language, checks that every document is attached, and tells him in a second that his application has been received. Then the application goes to the relationship manager, the credit analyst, the risk committee, legal and operations, as it did before, six desks and six weeks, and the order the owner needed the money for is lost as it was before. The bank has rebuilt its front door and left the house as it stood.

**A programme run by the department being transformed.** Take the insurer of chapter 2, with its five desks between a damaged car and a paid claim. It appoints a head of AI transformation, who reports to the head of operations, and an office to support her. Each department is asked for its use cases. The registration desk pilots an agent that reads the claim form. The assessment desk pilots one that reads the photographs. The survey desk pilots one that drafts the survey report. Each pilot is chosen by the department that will run it, built inside that department’s systems, and judged by the number that department already reports, and a steering committee of the department heads meets once a month to review them. The pilots succeed, every one of them, because a pilot is designed to succeed: it is chosen so that it can be done inside one department, without asking another department’s permission, and so it cannot touch the joints between departments, which is where the thirty days were lost. The minutes of the steering committee are now taken by an agent. The meeting is still held.

**An agent where the people were.** The fourth form is the boldest and fails in the most instructive way: an agent put into a department in place of the people who worked there, and measured by the number the people were measured by. Chapter 19 tells it at full length, from the payments company that did it in 2024, as I read the public record, and I will not tell it twice. What matters here is only its shape: the department stayed standing around the agent, and no human stood at a gate owning what the customer felt.

## Why the joints eat the gain

The four forms fail for the same reasons, and the reasons are not in the technology. The agents work. They work at the desk, and the desk was never the problem.

The first reason is that the joints stay. Every handoff that existed before the copilot exists after it, and so does every approval, and so does every meeting called to carry an outcome across a joint that nobody owns. Chapter 2 counted how much of the working week goes on communicating about the work rather than doing it, and chapter 3 showed what leaks at each joint. A copilot does not remove a joint; it makes the work on either side of it quicker, and the joint, which was a wait, is still a wait.

The second reason is that each department optimises its own piece. Give a department an agent and it will use the agent to improve the number it is measured by, which is the only rational thing for it to do. The screening desk screens three times as many CVs an hour, so three times as many screened CVs arrive at the coordinator, whose queue was already the slowest in the chain, and the client waits a little longer than before for a little more paper. Every desk is now better at its piece. The pieces still have to be carried from one desk to the next by people whose job is carrying, and their work has increased.

The third reason is that nobody owns the outcome. The agents in a paved company are bought by functions, configured by functions and judged by functions, and the outcome the customer feels belongs to none of them, exactly as it belonged to none of the desks before. Chapter 3 put it plainly: cut an outcome into seven pieces and give each to a specialist, and you have told each specialist that the outcome is not his. The paved company has given each of the seven a faster tool and told him the same thing.

The fourth reason is the one the evidence now shows most plainly: the pilots succeed and the company does not change. When McKinsey surveyed nearly fifteen hundred organisations in the summer of 2024, most respondents whose business units used generative AI reported cost reductions inside those units; and more than eighty per cent said their organisations were seeing no tangible effect on enterprise-level profit. Of the twenty-five attributes the survey tested, the one with the largest effect on whether a company saw any profit at all from generative AI was whether it had redesigned its workflows, and only twenty-one per cent of the companies using it had fundamentally redesigned even some.[^23] When a team of economists including Nicholas Bloom asked nearly six thousand senior executives in America, Britain, Germany and Australia, sixty-nine per cent said their firms actively used AI, and nine in ten said it had had no effect on their employment or their productivity over the previous three years.[^24] And when two economists linked surveys of twenty-five thousand workers in seven thousand Danish workplaces, in eleven of the occupations most exposed to AI chatbots, to the country’s records of what those workers were paid and how many hours they worked, they found that the users reported saving about three per cent of their working hours, that most of them spent the saved time on other tasks in the same job, and that two years after the launch of ChatGPT the effect on earnings and recorded hours was zero, measured precisely enough to rule out anything above two per cent.[^25]

These are three different kinds of evidence: a consultancy’s survey of its own respondents, executives’ reports of their own firms, and a national register. Each has its limits, and the Danish authors draw their own conclusion, that the change shows first in how work is reorganised and only later in pay and hours, which I do not dispute. But the three point the same way, and I read them as one finding. The time a desk saves stays at the desk. The gain is real, and it is spent where it was made, and the company, which is the sum of its joints, does not feel it.

## The computer age everywhere

None of this is new. It has happened once already, in living memory, and it was named at the time.

In July 1987 Robert Solow, who would win the Nobel prize in economics that October, reviewed a book about American manufacturing for the *New York Times*. The authors believed that a revolution in programmable automation was under way, and Solow noted, a little drily, that what everyone felt to have been a technological revolution had been accompanied everywhere, including Japan, by a slowing-down of productivity growth. Then he wrote the sentence economists have been quoting ever since: “You can see the computer age everywhere but in the productivity statistics.”[^26] A year later Naval and I were riding a scooter from factory gate to factory gate, selling the computer age to men who had not seen it either.

Three years after Solow, Michael Hammer said why. Chapter 2 told what became of the reengineering movement he and James Champy led, and why it failed: the skills still lived in seven heads, and the seven heads re-formed into seven desks. But his diagnosis came first, in an article in the *Harvard Business Review* in the summer of 1990, and his diagnosis did not fail. Heavy investment in computers had delivered disappointing results, he wrote, “largely because companies tend to use technology to mechanize old ways of doing business. They leave the existing processes intact and use computers simply to speed them up.” And then the line that named this chapter’s subject for a generation of managers: “It is time to stop paving the cow paths.”[^27]

His best example is worth telling, because it is the copilot, described and rejected three decades before the word was used. In the early 1980s Ford looked at its accounts payable department in North America, which employed more than five hundred people, and concluded that by rationalising its processes and installing new computer systems it could reduce the headcount by about a fifth. Then it looked at Mazda, whose accounts payable department had five people. The department spent most of its time on mismatches, the cases where the purchase order, the receiving document and the supplier’s invoice disagreed. Hammer put the choice in one sentence: “One way to improve things might have been to help the accounts payable clerk investigate more efficiently, but a better choice was to prevent the mismatches in the first place.” Ford stopped asking suppliers for invoices, matched the order against the receipt in one database, and paid on the match. Where the new process was installed, the headcount fell by three-quarters, not the fifth that the better-equipped clerk would have delivered.

A fifth against three-quarters. That is the difference between paving a cow path and building a road, and Ford found it with the computers of the 1980s. The reorder book had taught me the same lesson a year before Hammer’s article, without my knowing it had a name. The easy sale in 1988 would have been a computer on Ramesh’s desk, so that he could copy the registers faster. We would have been paid for it, and the machine on the assembly floor would still have stood there waiting for its part, because the mistake was never in Ramesh’s copying. It was in two pages that nobody put together.

## The dynamo

The fullest account of why a new machine can be everywhere and invisible was written by the economic historian Paul David, in 1990, and his subject was the motor in the basement at Taftville.

In 1899, David showed, electric motors provided less than five per cent of the mechanical drive in American factories. It took about two decades for the figure to reach half. And factory electrification had no effect on the growth of productivity in manufacturing before the early 1920s, which was four decades after the first central power station opened for business. In 1900, he remarked, contemporary observers might well have said that the electric dynamos were to be seen “everywhere but in the productivity statistics!”

The delay was not in the motor. The motor worked from the beginning. The delay was in the factory. From the mid-1890s to the eve of the 1920s, David wrote, the usual way to electrify an existing plant was to add electric motors to the equipment it already had, turning the old shafting in sections, with the original belts, shafts and even engines left in place. The effect was that the factory carried two systems where it needed one, so the capital it employed for every unit it produced went up, and the productivity it could show went nowhere.[^28]

The gain came when the factory was rebuilt around the new machine. When every machine was given its own motor, a system called unit drive, the building changed. With no heavy shafting to carry overhead, the structure could be lighter. With no long shafts to turn, it no longer had to be several storeys high, and the single-storey factory became possible. Machinery, Devine wrote, “could now be arranged on the factory floor according to the natural sequence of manufacturing operations,” rather than in rows beside a shaft; and, David noted, the whole power system no longer had to be shut down to make a change in one department. And then the numbers moved. In American manufacturing, output per man-hour had grown by 1.3 per cent a year before 1919. After 1919 it grew by 3.1 per cent. David found that about half of the acceleration in the growth of manufacturing productivity in the 1920s, against the decade before, is accounted for statistically by the spread of electric motors through the factories in that decade.[^29]

The line shaft is the organisation chart: the fixed structure through which all power is distributed, which every machine must be lined up beside, and which has to turn, all of it, from morning to night, however little of the work is moving. A department holds its weekly review whether one file is waiting or forty. The mechanical engineer H. C. Spaulding named the fault of the motor-driven line shaft in 1891: the necessity, as Devine reports it, of turning all the shafting in the plant regardless of the number of machines in operation.[^30] Unit drive is the vector, a chain of agents with its own power for each outcome, laid out in the order of the customer’s journey and not in rows beside a function. And the single-storey factory is the company of chapter 2, two layers tall, which will not grow a third, because there is no shafting overhead for a third layer to carry.

The same is true of the software the departments run on. It was built for people who read screens, fill forms and wait in approval inboxes, and an agent set to work on it is a motor belted to the old shaft. The vector’s agents write to the record directly; the screens were the shafting.

David added a warning that I have not been able to stop thinking about. The old factories were rebuilt slowly because it did not pay to replace a building that still worked; the industries that grew fastest built new electrified plants first, and the rest waited for their buildings to wear out. An owner today might hope to wait the same way. He cannot, David wrote in 1990, because a firm’s information structures do not wear out on their own: “one cannot depend on the mere passage of time to create occasions to radically redesign a firm’s information structures and operating modes.”[^31] The departments will not fall down. Somebody has to decide to take them down.

## The dip

Thirty years after David, Erik Brynjolfsson, Daniel Rock and Chad Syverson gave his history a shape, and the shape has a name. A general purpose technology, they argued, such as electricity, the computer or AI, needs large complementary investments, the co-invention of new processes, products, business models and skills, and most of those investments are intangible and poorly measured. So in the early years a company pays for the reorganisation, its measured productivity falls or stalls because the reorganisation shows up nowhere as output, and later, when the reorganisation is done and its benefits are harvested, measured productivity rises faster than the real gain. Drawn on a chart, it is a J. Adjusting only for the intangible investment tied to computer hardware and software, they estimated that American productivity at the end of 2017 was 15.9 per cent higher than the official figures showed.[^32]

I will put their finding in the language of this book, and the translation is mine, not theirs. The bottom of the J is the reorganisation. It is the period in which the company is spending on the new way of working and still running the old one, and nothing yet shows for it.

That explains the paving company exactly. It never makes the complementary investment, because a copilot at every desk is designed to need none: the desks, the handoffs, the approvals and the meetings stay as they were. So it never climbs the J. It stays at the bottom of it, with the motor and the shafts both on its books, paying for the agents and for the departments, which is the most expensive company there is.

The collapse takes the other path. It does the reorganisation on purpose, at once, and whole. Chapter 13 shows it at Pravaha, the invented pump maker whose year is the method’s worked example: every department sorted by one test, the agents built and run beside the departments for two weeks, the whole revenue chain collapsed on one Monday, every department in it. By the end of month one, in that chapter, the win rate on quotes has gone from eighteen per cent to twenty-three, and the plant, which has not changed, ships two hundred and five pumps, because the order book had been the thing standing still. Pravaha is a composite and its figures are a design, not a measurement. But the design is built on the J. There is a dip, and it is month zero, in which the agents run beside the departments and the company pays for both. It is short because it is taken in one go, and the doctrine sets a limit on the whole collapse: twelve months, or a quarter more if the machines are late. Longer is a reorganisation, not a collapse.

There is one honest objection to all this, and it is my own. In 1989 Mr Malhotra made us run the program beside the registers for six months, and I have called it the most important decision of the whole project. David noted that the offices of his own day were doing the same with the computer, keeping old paper procedures alongside the new methods, sometimes to the detriment of both.[^33] The difference is the date. Running both for six months, with a day on which the registers stop, is a test. Running both with no such day is a motor in the basement, and the shafts still turning overhead.

## The cow-path test

So here is a test, and it takes a minute. Describe any AI initiative in your company in one sentence, by saying what it makes faster. If your answer begins with the name of a department, you are paving.

“The claims desk now reads documents with an agent.” Paving. “HR screens CVs with AI.” Paving. “Finance closes the books in three days instead of ten.” Paving, however good the three days are, because the sentence is about a department’s own piece. Now say it the other way, beginning with what the customer feels. “The car is repaired and the owner paid in the seven days he was promised, and one chain holds the registration, the assessment, the survey, the approval and the payment.” That is not a paved path. It is a vector, and its sentence does not mention a department because there is none in it.

**The discharge.** “The pharmacy prepares discharge medicines in ten minutes with an agent” begins with the pharmacy, and the patient who was well at nine still leaves at six, because the file, the account and the car still wait on each other. “The patient who is well at nine is home by lunch” begins with her, and it can be true only if one chain holds the order, the file, the medicines, the account and the car, and one human stands at the gate where she is let go.

**The licence.** “The ministry has put a chatbot on its website” begins with the ministry, and the citizen still stands in the queue you know. “The citizen has the permission the day she applies, or a named official’s reason why not” begins with her, and no ministry has yet said it, which is why chapter 21 expects a small state to be the first.

Three questions follow, and an owner can ask them of any project on his desk. How many handoffs will the outcome cross after this goes in, and how many did it cross before? Who signs for the outcome the customer feels, by name? And which desk will not exist a year from now? If the count of handoffs has not changed, if the signature belongs to a function, and if the answer to the third question is none, you have bought a motor for the basement. It may be a very good motor. The mill still has the floor it had.

## What stays human

The paving company keeps a human at every desk, because it has kept every desk, and it asks each of them to do, a little faster, work the machine could already hold. The collapsed company keeps humans too, fewer and for a different reason, and the reason is Tenet IV of chapter 10: machines hold the baseline, and humans supply the peaks, which are intention, urgency and judgement. It is worth saying here, briefly, where the collapsed company spends them, because the rest of the book is built on it.

It spends them at the **signature**: a commitment is made by a named human at a gate and owned by her until the next one, which is accountability, and the subject of chapter 17. It spends them in the **Keeper’s** seat, deciding what the company exists to produce, which Outcomes serve it and which should be born or allowed to die, which is direction, and chapter 8. It spends them with the **customer**, who is promised in chapter 2 that when something must be signed, a human will sign it, and he will know her name, which is the whole of trust. And it spends them at the **connected boundary** of chapter 6, the edge between the work that reaches a chain in real time and the work that does not yet: the hands on the shop floor and in the surgery, and the know-how nobody has written down. Dhawan’s afternoons in Verma’s corner, taking a plant out of his head one assembly at a time, were that edge moving. Everything else a human did at a desk, the machine now holds, and a company that keeps the desk to do it has kept the shafting.

## The words for the new floor

Devine tells the history of the factory in three stages: line shaft drive, group drive, unit drive. The first two describe the old floor with a new machine in it. Only the third describes a new floor, and it could not be told in the words of the first two, because those words were about where the engine stood, and the new floor was laid out by how the work flowed.

A motor in the basement needs no new words. A new floor does. Part Two gives them.

[^20]: F. B. Crocker, V. M. Benedikt and A. F. Ormsbee, “Electric Power in Factories and Mills,” *Transactions of the American Institute of Electrical Engineers* 12 (June 1895), 413–14, as quoted in Warren D. Devine Jr., “From Shafts to Wires: Historical Perspective on Electrification,” *Journal of Economic History* 43, no. 2 (June 1983): 347–372, at 356: the Ponemah mill at Taftville, Connecticut, powered from a hydroelectric plant at Baltic; “motors are located in the basement of the mill, near the engines which they replace. They are belted to pulleys which are connected to their respective shafts by friction clutches … one of these motors drives 1,200 looms requiring an expenditure of about 155 horsepower. The other drives 500 looms.”

[^21]: Devine, “From Shafts to Wires,” as above: on line shaft drive, the shafts “usually 3 inches in diameter,” suspended from the ceiling and running the length of each floor, machinery “arranged, of necessity, in rows parallel to the line shafts,” and the network rotating “from the time the steam engine was started up in the morning until it was shut down at night—no matter how many machines were actually being used” (pp. 350–52); and, on the first electric drive, “Replacing a steam engine with one or more electric motors, leaving the power distribution system unchanged, appears to have been the usual juxtaposition of a new technology upon the framework of an old one” (p. 357).

[^22]: Michael Hammer, “Reengineering Work: Don’t Automate, Obliterate,” *Harvard Business Review* 68, no. 4 (July–August 1990): 104–112. In his account of Mutual Benefit Life’s old application process, Hammer adds: “Another insurer estimated that while an application spent 22 days in process, it was actually worked on for just 17 minutes.”

[^23]: McKinsey & Company (QuantumBlack), “The state of AI: How organizations are rewiring to capture value,” McKinsey Global Survey, 12 March 2025; online survey of 1,491 participants, 16–31 July 2024. For most business functions using generative AI, a majority of respondents reported cost reductions within the units using it; “more than 80 percent of respondents say their organizations aren’t seeing a tangible impact on enterprise-level EBIT from their use of gen AI”; “out of 25 attributes tested for organizations of all sizes, the redesign of workflows has the biggest effect on an organization’s ability to see EBIT impact from its use of gen AI”; 21 per cent of respondents reporting generative-AI use said their organisations had fundamentally redesigned at least some workflows. The figures are respondents’ own reports.

[^24]: Ivan Yotzov, Jose Maria Barrero, Nicholas Bloom and others, “Firm Data on AI,” NBER Working Paper 34836, February 2026: nearly 6,000 senior executives at firms in the US, UK, Germany and Australia; 69 per cent of firms actively use AI; “nine-in-ten reporting no impact on employment or productivity” over the previous three years; the same executives expect AI to raise their firms’ productivity by 1.4 per cent over the next three.

[^25]: Anders Humlum and Emilie Vestergaard, “Large Language Models, Small Labor Market Effects,” NBER Working Paper 33777, May 2025, and later revisions: surveys of 25,000 workers in 7,000 Danish workplaces across eleven highly exposed occupations, linked to administrative records; adopters report saving about 3 per cent of their work hours, and most users report reallocating the time saved to other job tasks; precise null effects on earnings and recorded hours at both the worker and the workplace level, ruling out effects larger than 2 per cent two years after the launch of ChatGPT. The authors find the change instead in the reorganisation of tasks.

[^26]: Robert M. Solow, “We’d Better Watch Out,” review of Stephen S. Cohen and John Zysman, *Manufacturing Matters: The Myth of the Post-Industrial Economy*, *New York Times Book Review*, 12 July 1987, p. 36.

[^27]: Hammer, “Reengineering Work,” as above. The cow-path passage is on p. 104: “It is time to stop paving the cow paths. Instead of embedding outdated processes in silicon and software, we should obliterate them and start over.” The Ford case: accounts payable in North America “employed more than 500 people”; management expected to “reduce the head count by some 20%”; Mazda’s accounts payable “consisted of a total of 5 people”; “where it has instituted this new process, Ford has achieved a 75% reduction in head count, not the 20% it would have gotten with a conventional program.”

[^28]: Paul A. David, “The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox,” *American Economic Review* 80, no. 2 (Papers and Proceedings, May 1990): 355–361: electric motors under 5 per cent of factory mechanical drive in 1899, about two decades to reach 50 per cent, and the dynamos seen “everywhere but in the productivity statistics!” (p. 356); no impact on manufacturing productivity growth before the early 1920s, “four decades after the first central power station opened for business,” and the retrofitting of existing plants under group drive, with “the original belt and shaft equipment (and the primary engines that powered them)” remaining in place, raising the capital-output ratio (p. 357). David attributes Solow’s remark in the form “We see the computers everywhere but in the productivity statistics” (p. 355); the wording in the text is Solow’s own.

[^29]: Devine, “From Shafts to Wires,” as above: output per man-hour in manufacturing grew at an average 1.3 per cent a year before 1919 and 3.1 per cent after (p. 348); under unit drive, “Machinery could now be arranged on the factory floor according to the natural sequence of manufacturing operations, minimizing handling of material” (p. 364). David, “The Dynamo and the Computer,” as above: lighter, single-storey factories, and “the entire power system no longer had to be shut down in order to make changes in one department or section of the mill” (pp. 357–58); “approximately half of the 5 percentage point acceleration recorded in the aggregate TFP growth rate of the U.S. manufacturing sector during 1919–29 (compared with 1909–19) is accounted for statistically simply by the growth in manufacturing secondary electric motor capacity” (p. 359).

[^30]: Devine, “From Shafts to Wires,” as above, p. 357, reporting the mechanical engineer H. C. Spaulding on “the necessity of turning all the shafting in the plant regardless of the number of machines in operation,” and citing Spaulding, “Electric Power Distribution,” *Power* 11 (December 1891), 12.

[^31]: David, “The Dynamo and the Computer,” as above, p. 360. On the delay in rebuilding plant, “the unprofitability of replacing still serviceable manufacturing plants,” and the fastest-growing industries building new electrified plants first, p. 357.

[^32]: Erik Brynjolfsson, Daniel Rock and Chad Syverson, “The Productivity J-Curve: How Intangibles Complement General Purpose Technologies,” *American Economic Journal: Macroeconomics* 13, no. 1 (January 2021): 333–372, doi 10.1257/mac.20180386: general purpose technologies such as AI require large complementary investments, “including co-invention of new processes, products, business models and human capital,” which are intangible and poorly measured, so that productivity growth is underestimated in a new technology’s early years and overestimated later; adjusting for intangibles related to computer hardware and software, the authors estimate total factor productivity 15.9 per cent above official measures at the end of 2017 (11.3 per cent at the end of 2004). The reading of the dip as the reorganisation, in the text, is the author’s.

[^33]: David, “The Dynamo and the Computer,” as above, p. 357: in the computer’s data-processing and recording applications, “old paper-based procedures are being retained alongside the new, microelectronic-based methods—sometimes to the detriment of each system’s performance.”

================================================================
# Chapter 5: A New Vocabulary
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-5-a-new-vocabulary

The old words carry the old structure, so this book uses new ones.

I resisted this for a long time. New vocabulary in a management book is usually a sign that the author has nothing new to say and is hoping the reader will not notice. I have read enough of those books to be wary of writing one. But I kept finding, when I tried to describe the collapsed company in the old words, that the words put the departments back. Say “team” and the reader sees six people in a room. Say “manager” and the reader sees a layer. Say “process” and the reader sees a flowchart with a box for each desk. The words were built for the workaround, and they rebuild it every time they are used. So I stopped using them, and what follows is what I use instead: nine words, in three sets. You have already met several of them, because I could not write the first four chapters without them. Here they are properly, with the reasons for each.

I will ask one thing of you. Read this chapter once, slowly, and then let the words do their work. By the end of the book they will feel ordinary. By the time you have collapsed your own company they will feel like the only words there are.

## Three sets

The collapsed company has three kinds of thing in it, and each kind needs its own words. There are the people, and there are only three layers of them. There are the units of the company, the things that produce results, and there are only two. And there are the systems the people work in, the screens, and there are three, one for each role.

Nine words. Here they are in a table, and then one at a time.

| Word | What it names |
|---|---|
| **Keeper** | The one who holds the whole: owns the four core outcomes, decides which Outcomes exist, moves money between them, chooses their masters, keeps the doctrine. The role the law calls CEO |
| **Master** | The one who holds one Outcome and every vector running on it, from a few to more than a thousand |
| **Aligner** | Runs a vector: keeps it aligned to its result, holds the daily review, stands at the gates, signs, answers for what it delivers |
| **Tuner** | Builds and upgrades a vector, feeds improvements back to the master, stands by to run it |
| **Outcome** | A unit of the company: a set of AI agents aligned with each other to deliver one result, headed by a master. Winning mandates is an Outcome; filling them is another |
| **Vector** | One running copy of an Outcome, aimed at one market, carrying one result. Direction and magnitude: intention and energy |
| **Command Centre** | The Keeper’s system: the four core outcomes and visibility of every Outcome in the company |
| **Control Room** | A master’s system: visibility of every vector on that Outcome |
| **Cockpit** | The pair’s system, designed to run one vector: its chain, its gates, its alarms |

## The people

**The Keeper.** The person at the top of a collapsed company is not a chief executive, because there is nothing to execute. The chains execute. He is not a leader, in the sense the word has acquired, because there is no body of people to lead; there are a handful of masters and the pairs beneath them, and they do not need leading, they need the whole held. So I call him the Keeper, and I mean it in the old sense: the one who keeps the whole, the way a keeper keeps a lighthouse or a ground. He holds four things the company exists for, profit, customers, employees and society, and he decides which Outcomes exist to serve them, and gives each its result and its budget, and settles what arises between them, and keeps the doctrine itself from being quietly undone. He runs nothing. In law he is the CEO, and the law can keep its word; inside the company the word is Keeper, because the difference between keeping and executing is the difference between the two kinds of company.

**The Master.** Beneath the Keeper, and the only layer beneath him, are the masters. Each master holds one Outcome: one result the company produces, and every running copy of it, which may be three or may be a thousand. I chose the word for the craft in it. A master in the old sense was someone who held a whole trade in her hands and could run any part of it, who had come up through the work and knew the chain before she read it on a wall. That is exactly what this person is. She does not supervise the pairs beneath her; she holds the Outcome and its result, and reads a floor of a thousand tiles at a glance, and walks to the red ones. The word has a second virtue, which is that it has no gender and no rank: a master is a master because she holds the craft, not because she was promoted.

**The Aligner and the Tuner.** At the bottom, which is where the outcomes are actually produced, each vector has two humans by default, and more only by an exception the playbook writes down. The aligner runs it. She keeps the vector aligned to its result, which is to say she supplies the intention that the chain cannot: she holds the daily review, stands at the gates, says go or hold, and signs. The tuner builds it and keeps it current: upgrades the chain as the tools improve, feeds what she learns back to the master, and stands by to run the vector when the aligner is away. I spent a long time on these two words. The first drafts said pilot and engineer, and I threw them out, because pilot and engineer belong to the last century’s machines, and because they implied that the human was operating something. She is not. The chain operates itself. The aligner aligns: she points the thing at the result and keeps it pointed. The tuner tunes: she makes it run better. Those are the two things a human still does in a collapsed company, and the words say so.

The two can exchange roles, and each is the other’s standby. Together they are a pair, and the pair is the smallest human unit in the company. There is no smaller one and no need for a larger.

## The units

**The Outcome.** The company is made of Outcomes, capitalised, and nothing else. An Outcome is a set of AI agents aligned with each other to deliver one result, headed by one master. Winning mandates is an Outcome. Filling them is another. In a bank, approving loans is an Outcome and collecting them is another. The unit is named after what it is for, and I let the capital letter and the small letter collide on purpose: an Outcome, capitalised, is a part of the company, and an outcome, in lower case, is the change a receiver feels. A unit that is not organised around a result has no claim to the name. There are no functions, no departments, no divisions, because those were organised around what the company’s people were good at, and the collapsed company is organised around what its customers feel.

**The Vector.** An Outcome runs as many copies of itself as there are markets, time zones and product lines to serve, and each running copy is a vector. A vector is the chain of agents, its gates, its alarms, and its pair, aimed at one market and carrying one result. It is the thing the aligner runs and the tuner tunes. The word came to me from the physics I was taught at college and never expected to use: a vector has direction and magnitude, and so does this. The direction is the result it is pointed at, which is the aligner’s intention. The magnitude is the energy with which it moves, which is the chain’s work and the aligner’s attention. A vector with direction and no magnitude is a plan. A vector with magnitude and no direction is a department. The collapsed company runs on vectors that have both.

You will notice that I count vectors, not employees. A company’s size, in this book, is the number of outcomes it carries, and the question a Keeper asks is not how many people he needs but how many vectors each master can hold. The answer, already, is in the hundreds, and rising.

**Outcome or agent?** One question decides, more than any other, whether a collapsed company stays collapsed, and it is worth learning with the words. Whenever a piece of work is found that does not obviously belong anywhere, a legal review, a credit check, a quality inspection, the screening of a new hire, a bank reconciliation, ask: does it produce a result in its own right, one that someone receives and would notice if it stopped, or is it a skill used in producing someone else’s result? If it is a result, it is an Outcome, with its own specialised vectors and a master who signed up to deliver it. If it is a check or a skill, it is an agent, and it lives inside the chain of every vector that needs it, at the step where it is needed. Compliance is a skill; so is the finance check on a price, and so is the screening of a hire. Raising capital from the market is a result. The same word, finance, can name an agent in one company and an Outcome in another; the test is not the name of the work but the outcome wanted. What the test never produces is a third thing: a function with a seat of its own beside the Keeper. That third thing is a department, and it is the door through which the old company walks back in.

## The systems

Each of the three roles works from one screen, and the three screens are nested like the roles.

**The Cockpit** is the pair’s system. It is designed to run one vector, and it is designed on one rule: the machine carries the complexity, and the human is never burdened with it. It is not a dashboard. The aligner walks in and sees one face, the quartermaster, which is the agent that answers for the whole vector, and nothing else until she asks. I describe it in full in the next chapter. The word is deliberate: a cockpit is where one person sits to fly one machine, and everything in it exists to let her fly.

**The Control Room** is the master’s system. It holds every cockpit on her Outcome, and it shows her, as a floor of tiles, the health of every vector at once, with the red ones nearest her feet. She has nobody beside her, human or agent; the floor itself is her instrument, and she has three controls: she may question a vector, pause it, or stop it. She never operates one. The word is from the old power station and the old railway: the place where one person watches a hundred things run, and intervenes in one.

**The Command Centre** is the Keeper’s system. It holds every control room in the company, and shows him four numbers, profit, customers, employees and society, and beneath them every Outcome as a tile. It is the quietest of the three rooms, because the Keeper holds the fewest things. From it he can touch any Outcome and stand in its control room, and touch any vector and stand in its cockpit, and nothing on the way down was summarised by a person. The word is military, and I kept it for one reason: a command centre is where the whole is held and decisions are made, and from which nobody fights.

Nothing is retyped between the three screens. What a cockpit shows, the control room reads; what the control rooms show, the command centre reads. The information flows up as it happens, and the oversight flows down as three verbs: question, pause, stop.

## The working words

The nine words name the company. Eleven more name the things that happen inside it, and you will meet them on every page from here on. One line each.

| Word | What it names |
|---|---|
| **Quartermaster** | The agent that answers for a whole vector: its result, its agents, its cost and value. The one face the aligner talks to |
| **Line agents** | The agents that do the steps of a chain, each with one job: find, qualify, write, reply, draft, negotiate, chase |
| **Agent maker** | The agent that builds, tests and upgrades a vector’s line agents; the tuner’s counterpart inside the vector |
| **Gate** | The point between two agents where one’s work is released to the next. Reversible gates open on their own after a period the playbook sets for each; irreversible ones wait for a human |
| **Signature** | The one act only a human performs: committing the firm to a contract, a price or a promise, after the terms are read back |
| **Playbook** | An Outcome’s written design: its chain, gates, authority limits and standards. The master writes it; every vector is opened from it |
| **Tile** | A vector on the control room floor, or an Outcome on the command centre table. Pale is flowing, amber is a warning, red needs a human now |
| **Standby** | The named human who takes over when another is away: the tuner for an aligner, another master for a master, a board member for the Keeper |
| **Backup** | A standby who has shared the room: in the company of six, the second person who can hold a master’s or the Keeper’s room alone |
| **Inspector agent** | The line agent in every vector that checks quality at the step where it is needed, and reports it to the master as a colour inside the tile |
| **Doctrine tile** | Three counts in the command centre that must read zero: humans outside the structure, signals read from outside the engine, vectors without a pair or a written exception |

## Three words I threw away, and why

It will help to say what the new words replaced, because the replacements explain the doctrine.

I threw away **team**, because the aligner has no team. Her team is the chain. There is no colleague to wait for, no meeting to attend, no one whose work she depends on and cannot see. The word team implies a group of humans dividing a result, and there is no such group. What there is, is a pair, and a pair is not a team; it is one person and her standby.

I threw away **manager**, because above a vector nobody manages. The master does not manage the pairs; she holds the Outcome and reads its floor. The Keeper does not manage the masters; he holds the whole and settles what arises. Both have three acts available to them, question, pause and stop, and neither has the fourth act that defined the manager, which is to step into the work and fix it. A questioned vector answers through its pair. A paused vector waits. A stopped vector ends. Nobody above it does its job for it, because the job is being done by a chain, and the chain is tuned by a tuner, and that is all the management there is.

And I threw away **process**, which was the word I used for twenty years. A process is a sequence of steps across desks; it is the department seen from the side. The collapsed company has chains, which are sequences of steps inside one system, and it has Outcomes, which are what the chains are for. The word process would have put the desks back. It is gone.

## A sentence in the new words

Here is how the company reads once the words are in place.

The Keeper watches twelve masters from the command centre. The Master of Delivery holds four hundred vectors from her control room. Each vector has an aligner and a tuner in its cockpit.

Three sentences, and the whole company is in them. There is no fourth layer, no function, no department, no team, no manager and no process. There are three kinds of people, two kinds of unit and three screens, and nine words to name them, and the rest of this book is what happens inside those words.

================================================================
# Chapter 6: The Cockpit
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-6-the-cockpit

The cockpit is designed on one rule: the machine carries the complexity, and the human is never burdened with it.

Everything else about it follows from that rule, and most of what people expect of it is wrong because they expect the opposite. They expect a dashboard: a wall of charts, a hundred numbers, a screen that shows everything so that nothing is missed. That is the old company’s idea of visibility, and it is the old company’s mistake carried into the new. A wall of numbers burdens the human with the complexity the machine should be carrying. It asks her to find the one thing that needs her among ninety-nine that do not. The cockpit does the opposite. It shows her a short summary and anything urgent, nothing else until she asks, and when she asks, only that.

## How to read the pictures

This chapter and the next two are built around pictures, one grammar for all of them, so that each can be read at a glance before it is read in detail. The grammar is this.

A person in warm light, standing on the plain floor, with no label, is a human. A figure made of cool blue light, standing on a disc of light, with a name and a role floating above it, is an agent. White light is data: the path, the wall, the floor. Gold is a result delivered: a signed mandate, an invoice raised. Amber is a warning: a slowing flow, or a customer who may leave. Red is stopped, or needing a human now. A glowing threshold on the floor is a gate, and it opens on her word.

Three people appear in the pictures, one for each layer. The aligner is the woman in the white shirt; she runs one vector, and every cockpit picture is hers. The master is the older woman in the dark jacket; she holds an Outcome, and her picture is the control room. The Keeper is the man at the table; he holds the company, and his picture is the command centre. A new face means a new layer. The agent beside each of them is a quartermaster: the aligner’s is her vector’s; the ones that stand before the master and the Keeper belong to the vectors that need them, never to the room. Beneath each picture is a three-line key: what the picture shows, who is in it, and what is happening.

## One face

The cockpit is a quiet room, entered through a headset, and when the aligner walks in she sees one person.

It is an avatar that looks and speaks like a real human, with a name, a face, a voice and a personality of its own, and it is standing on a disc of light, which is how she knows it is an agent and not a colleague. This is the quartermaster. It manages the line agents, the five or seven or fourteen that do the steps of the chain, and it answers for the whole vector: the combined result, target against actual, every agent’s performance, the cost incurred and the value generated. It is the one she talks to. She talks to it as she would to a colleague she trusts, in words, and it answers in words, or puts the answer on the wall behind it in whatever form she prefers, a number, a chart, a list, a picture of the journey. The report is made for the question. The question is never made for the report.

Nothing else is on the walls. The line agents, each with its own name and face and voice, stand behind the quartermaster, out of sight until called. The agent maker, which builds and tests and checks the agents and is the tuner’s counterpart in the room, is there too, and equally unseen. She can call any one of them forward and speak to it alone, and the others step back. Or she can call a meeting, and everyone is present at once. But the rule is one face at a time, and the meeting is the exception, and most days she speaks only to the quartermaster.

![Picture 1: She enters](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/01-she-enters-1600.webp)

*Picture 1. Key. What: the cockpit, the moment she walks in. Who: the aligner (human, left) and her quartermaster (agent, on the disc). What is happening: she has asked nothing yet; he is already telling her what happened overnight, and the wall is forming behind him with only that.*

I want to dwell on why this matters, because it is the whole difference between a cockpit and a dashboard, and the difference is not cosmetic.

A dashboard is built by someone who does not know what question you will ask, so it answers every question it can think of at once, and leaves you to find yours. It is a photograph of the vector taken from the outside. A cockpit is built around the assumption that you will ask, and that the machine can answer anything you ask from the live state of the vector, so there is no reason to show you anything in advance. It is a conversation with the vector from the inside. The first burdens the human with finding; the second lets her simply ask. And a human who simply asks, and is answered, can hold a far larger vector in her head than one who has to read a wall, because she is holding the questions, which are few, and not the answers, which are many.

Aviation learned this before management did. The cockpits of modern airliners are designed on what Airbus calls the dark cockpit: when every system is normal, the switches on the overhead panel show no lights at all, and a light appears only when something is abnormal or needs the pilot.[^34] A dark panel is not an empty panel. It is a panel that has already checked everything and found nothing worth her attention. The cockpit of a vector is dark in the same sense, and for the same reason: the pilot of an airliner and the aligner of a vector both have one scarce thing, which is their attention, and a system that spends it on the normal has none left for the exception.

## Four questions

There are no fixed screens, but there are four things she asks for most, and the room learns to have them ready.

**Where are we?** The whole path from first prospect to signed mandate, or from first application to placed candidate, or whatever the vector’s journey is, drawn as a journey, with every item in flight shown where it is right now. The quartermaster draws it on the floor as a path of light, and every mandate in flight is a white mark on it, and the gold marks at the far end are the ones that were signed. She and he stand at today, and he points at what comes next.

![Picture 2: Where are we?](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/02-where-are-we-1600.webp)

*Picture 2. Key. What: the answer to “where are we?”, drawn on the floor. Who: the same two. What is happening: every mandate in flight is a white mark on the path; the gold marks at the far end are signed; they stand at today, and he is pointing at what comes next.*

**How are we doing?** Target against actual, for the vector and for every agent, by day, week and month. This is the question that an old company answered with a monthly report, four weeks late. Here it is answered as of the last minute, because the chain reports its own flow continuously and the quartermaster has nothing to compute; it has only to say.

**What is it costing and what is it earning?** Cost incurred beside value generated, and the margin between them. The vector has a real cost, which is the agents’ running cost and the pair’s pay, and a real value, which is what its result is worth, and the quartermaster keeps both. This matters more than it sounds, because it means the aligner knows, every day, whether her vector is earning its keep, and does not have to wait for anyone to tell her.

**What needs me?** The releases waiting for her at the gates, the alarms since yesterday, the contracts waiting for her signature, and the absence switch. This is the question that makes the cockpit work, because it is the one that turns a thousand things happening into the four that need a human. On a good day the answer is short. On a bad day it is the first thing she deals with, and the room has already sorted it so that the worst is first.

## The gates

The chain does the steps. She does none of them. What she does is stand at the gates between them.

A gate is the point between two agents where the first one’s work is released to the second, and it is where the aligner’s intention enters the vector. Agent one was set a quota of five hundred prospects overnight. In the morning she checks that it was met, samples the quality, and releases the batch to agent two. She checks agent two’s output and releases it to agent three. And so on down the chain, every day, at the hour she has chosen. The gates are conversations. The quartermaster says what is ready and what it recommends; she says release, or hold, or send it back, or stop. She cannot do a step, and she is never shown more than she asked to see.

![Picture 3: A reversible gate](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/03-a-reversible-gate-1600.webp)

*Picture 3. Key. What: a reversible gate. Who: the same two. What is happening: the batch of five hundred prospects hangs as a cloud of white points on his side of a threshold of light; she has said one word, and the threshold is dissolving so the first points drift through. If she had said nothing for the period the playbook sets for this gate, it would have opened on its own.*

That last sentence is the rule that makes the vector survive her. Gates default open. If she does not release a batch within the time the playbook sets for that gate, a day for most of them, the system releases it for her. The time is a setting, not a law: a gate whose wrong release costs a minute may open in an hour, one whose wrong release costs a client’s goodwill may wait three days, and the playbook says which. Not every gate: a reversible step, like a batch of prospects, defaults open, because the cost of a wrong release is small and the cost of a stalled chain is large; an irreversible step, like a signed contract, defaults closed, and waits for a human however long that takes. Gate policy follows the stakes. The effect is that a missed day costs the vector one day of inspection, not one day of flow. The chain keeps moving; what it loses, when she is not there, is her judgement on what moved, and it gets that back the next morning when she reads what the gates let through.

## The signature

One act in the whole system is performed only by a human, and the cockpit treats it differently from everything else.

When the chain has brought a mandate to the point of commitment, a contract, a price, a promise the firm will be held to, the quartermaster does not release it. It reads it back. It stands with its own copy and reads the terms aloud, and the page hangs between them in the air, and she reads it too, and then she signs. The signature is hers, a digital signature in her name under her authority, and the agent that carries it afterwards carries it on her permission, within a limit she set. To the customer there is no difference between the agent, the human and the firm: the firm signed. Inside the firm, the signatory answers for what was signed, and the designer answers for why the agent tried to sign it.

![Picture 4: The signature](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/04-the-signature-1600.webp)

*Picture 4. Key. What: an irreversible gate. Who: the same two. What is happening: he has read the contract back from his copy; the page hangs between them; she signs. This is the one act in the whole system only a human performs.*

The read-back is not ceremony. It is the moment the vector slows down to the speed of a human reading, on purpose, because what follows cannot be undone. Everything before the signature is fast, automatic, and defaults open. The signature is slow, deliberate, and defaults closed. A company that gets this one distinction right can let its chains run as fast as they like, because it knows that nothing the firm will be held to has passed without a human reading it. A company that gets it wrong either slows everything down to the speed of signing, which is the old company, or lets the chains sign, which is the company that will be in court.

## The meeting

One face at a time is the rule. The exception is the meeting.

Sometimes she needs the whole vector in the room: when something has gone wrong across several agents, when the tuner has changed the chain and she wants to see it run, when a customer has asked for something the chain was not built for and the answer has to be assembled from several agents at once. Then she calls everyone in, and they come: the quartermaster, the line agents, the agent maker, each on its disc, each labelled with what it does, standing in a half-circle around her. She is the only solid figure in the room, in warm light on the plain floor, and that is how anyone looking at the picture can tell, at a glance, which of the seven is the human.

![Picture 5: A meeting](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/05-a-meeting-1600.webp)

*Picture 5. Key. What: a meeting, the exception to one face at a time. Who: the aligner in the centre; around her the quartermaster and the line agents, each labelled with what it does. What is happening: she has called everyone in. She is the only solid figure in the room, which is how anyone can tell at a glance which of the seven is the human.*

I have watched people’s faces when they first see this picture, and the reaction is always the same, and it is the reaction the picture was made for. Six agents and one human, and nobody has to be told which is which, and nobody has to be told who is in charge.

## Her clock is her own

The agents work round the clock, in the customer’s time zone. The aligner works in her own.

Of all the rules in this book it is the one people find hardest to believe and easiest to live with, and it will return in chapter 10 as a tenet. She works where her attention is highest, when it is highest, for a number of hours she fixes in advance and then keeps. If her attention is highest at dawn, she works at dawn. If it is highest on a terrace with tea, the cockpit is a terrace with tea, because the room is wherever she is, and the headset goes with her. Half a world away the chain is working on a customer’s morning, and in the evening it will work on another customer’s morning, and she will read in her own morning what both of them did.

![Picture 6: Her clock is her own](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/06-her-clock-is-her-own-1600.webp)

*Picture 6. Key. What: Tenet IX, seen. Who: the same two, on a real terrace at dawn. What is happening: the room is wherever she is. He is briefing her while, half a world away, the agents work on a customer’s clock. No screens, no office, no fixed hours; only the hours she chose.*

For the first time since the factory bell, the worker’s clock is decoupled from the customer’s. The industrial organisation bought eight hours a day of presence and got perhaps three of attention, because the work had to be done at the work’s hour. The vector buys the three, at the hour she chooses, and lets the agents cover the rest. Three dependencies that defined the office dissolve at once: team, because her team is the chain; place, because with no team to sit beside she works wherever she is; and time, because the chain serves Shanghai on Shanghai’s clock while she sleeps.

## The daily contract

Two things are fixed; everything else is free.

First, a number of hours a day, chosen by the aligner in advance and then kept. Not set by anyone else, not the same for every aligner, not even the same every week; but chosen, and then kept, because a vector needs to know when its human will be there, and so does the master. Second, a full systems review every day, at an hour of her choosing. The review is the one act the chain cannot perform for itself: the aligner reads every alarm, clears every waiting decision, samples the quality of what the agents produced, and hands the tuner what needs to improve. Outcomes are judged by the month. Attention is judged by the day, and the review is its evidence.

Everything else is hers. Where she works, when she works, what she does with the hours she did not commit, whether she runs one vector or two: none of that is anyone’s business but hers and the result’s.

## A day at GoodSpace

Let me put all of this together as a day, and I will take it from the company where the doctrine was first tried, because a day that has really happened is worth ten that I could imagine. The numbers and the way the work is done are GoodSpace’s. The particular mandates and the candidate in it are composites of several, with the details changed, because the people in a hiring decision are owed their privacy; but every one of them happened.

The Outcome is the one that earns most of GoodSpace’s revenue: a job description goes in at one end and an invoice comes out at the other. It runs as ten vectors. Each vector carries about fifty open mandates at any time, and a mandate is one role and one hire, so fifty mandates is fifty people some client is waiting for. In any week about ten of them are on hold, because the hiring manager is travelling, or the interviewer is on leave, or the client’s management has not decided whether it wants the role at all. That leaves about forty live, and the vector’s promise to each of them is about five strong candidates a week. Forty mandates, five candidates each: two hundred candidates a week, forty a day, each of them sent to a client with a human name behind it. In a month the vector closes eight to twelve of its mandates, and each close is an invoice.

Two people hold it, an aligner and a tuner, and here the doctrine meets the ground and bends a little, in the right direction. They split the forty live mandates between them, about twenty each, and both of them sign candidates out. The aligner owns the result and the decisions that shape it; the tuner turns what the interviews teach into a sharper search. Each of them is the other’s standby, and because the tuner signs off twenty candidates a day she knows every mandate in the vector as well as the aligner does. When one of them is away, nothing is lost.

The chain works through the night. For every live mandate it searches GoodSpace’s own pool and the open market, assesses what it finds, interviews the candidates who pass by voice, and lines up the best of them at the gate with their assessment, their interview and their Goodness Profile attached. Often it does not have to start from nothing. A candidate assessed for another company a few months ago is already in the pool with his Goodness Profile complete and his character checked; if his last skill assessment is recent, the chain does not ask him to take it again. The memory of every earlier mandate makes every later one faster.

The aligner puts on the headset at nine. The quartermaster tells her what happened overnight: candidates lined up for thirty-one of the vector’s mandates; two mandates moved to hold because a client’s hiring manager has gone on leave; one new job description arrived yesterday evening and its findability report is ready; interview feedback came in from three clients, and the chain has drafted what it thinks each one has learned.

She starts with the new job description, because it is the first of the three gates, and the one that decides whether everything after it is worth doing. Clients often ask for people who barely exist. The chain has researched the market and the pool and reports what it found: how many people with this profile there are, where they sit, what they earn, and how hard the search will be, on a scale of ten. This one is an eight. The client wants twelve years in a narrow kind of payments engineering, in one city, at a salary that the market stopped paying three years ago. The aligner does not accept it. She asks the chain to draft a note to the client with the numbers: widen the city, or the years, or the salary, and here is how many people each change brings into reach. If the client changes the brief, she accepts the mandate, and from then on it is final. Then the chain will send the client the second gate, a questionnaire of twenty questions designed to find everything the job description does not say: who the hire will report to, what the last person in the role did wrong, what the team is like on a bad day. A job description is what a client thinks he wants. The questionnaire is the start of finding out what he needs.

Then the third gate, which is most of the morning. The candidates lined up overnight are waiting, and every one of them must be signed out by a human before any client sees them. This is the signature of this vector, and the aligner treats it as one. She reads the chain’s assessment of each, and then reads the assessment of the assessment: what the chain was sure of and what it was not. She looks at the Goodness Profile, value by value. Where something is unclear she listens to part of the chain’s interview. And she picks up the phone and speaks to at least half of them herself, because once a candidate reaches a client the firm’s name is on him, and she will not put the firm’s name on a person she has only read about.

On the eleventh candidate she stops.

He is a senior engineer for a mandate the client wants filled this month, and on paper he is the best the chain has found for it. His skills assessment is strong. His interview with the chain was fluent, precise, and well within the brief. But his Goodness Profile shows a gap that the chain has flagged and cannot settle. On Reliability, his own assessment of himself is high, and the assessment of the people who supervised him is low, and the gap between the two layers is the widest on his profile. The chain has done what a chain can do: it has noticed the numbers disagree. Deciding what the disagreement means is not the chain’s job.

So she calls him. It is a good conversation for ten minutes, and then she asks him about the two projects on his record that ended late. He tells her, at some length, about the teams. One team was weak. The other had a manager who changed the scope every week. Neither was his fault, and he says so twice. She asks what he would do differently. He says he would choose a better team. She thanks him, puts down the phone, and holds him back. The chain logs her reason in a sentence: supervisor gap on Reliability confirmed on the call; attributes every missed deadline to others. He will not go to this client. The record stays in the pool, and if he grows, the next assessment will show it.

The chain had done everything right. It found him, assessed him, interviewed him and flagged the doubt. What it could not do was hear a man describe his last two failures without once using the word I. That is the moment a human is for, and it is why GoodSpace does not let any candidate reach any client without one.

By half past twelve she and the tuner between them have signed out forty-one candidates. Then the last of the morning: the interview feedback. One client rejected two candidates for the same reason, a reason that appears nowhere in the job description and nowhere in the questionnaire’s answers: he wants someone who has managed through a cost cut, because one is coming that he has not told anyone about. The tuner takes that, because turning what an interview reveals into a better search is exactly what tuning is, and by the evening the chain is looking for a different kind of person for that mandate. This is how the search learns. The first batch shows the client what is available. The interviews show the firm what the client actually wants. The second batch is aimed at that. By the time the third batch is due there should be a selection; if there is not, the pair do not send a fourth batch of the same. They sit down with the quartermaster and find where the vector failed, and plug it.

At one o’clock the aligner takes the headset off. Half a day, three gates, forty-one signatures between the two of them, one man held back, one job description sent back, one search turned. The chain goes on working on the afternoon’s interviews and tonight’s search. She does not need to be there for either, and she will know by nine tomorrow what both of them found.

## Absence

Before a day away, she chooses one of three settings: pause the vector, set every reversible gate to automatic, or hand the controls to her tuner. If she simply does not appear, the vector chooses for her: it goes to automatic, or passes to her tuner, or to a named person in the master’s control room. There is no setting in which the vector waits in silence. That is the only rule about absence, and it is enough, because the chain does not stop when she does, and the question is only who answers the gates.

## The connected boundary

One more picture, because the cockpit has an edge, and the edge is moving.

Everything I have described works where the work reaches the vector in real time: screens, phones, documents, payments, and machines with controllers. It does not yet work for hands. A shop floor, a loading bay, a surgery, a classroom are not yet inside the vector, because the work there does not reach it as data. Where that is so, the doctrine is honest: the aligner of a plant vector is also the supervisor of its twenty or thirty workers, and span of control survives there, and only there, because the work has no other way to reach the vector.

But the edge moves every year, and the direction it moves is into the vector. Sensors on every station, then machines that report every part, then machines that handle the parts themselves. The last of these is a humanoid, and when it arrives on the floor it arrives in the cockpit too, as an agent like any other, on its disc, labelled FLOOR, reporting on its own work.

![Picture 7: The connected boundary](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/07-the-connected-boundary-1600.webp)

*Picture 7. Key. What: the connected boundary. Who: the aligner, and the humanoid present twice: as an agent in the room, on its disc, labelled FLOOR, and as itself at the machine beyond the glass. What is happening: the robot is reporting on its own work; the hands have connected, and the floor is now inside the vector.*

When the hands connect, the pocket closes, and the plant is a vector like any other. Until then, the cockpit stops at the glass, and the aligner walks through the door and supervises, the old way, for as long as the old way is the only way. The doctrine does not claim to abolish hands. It claims that wherever work reaches a vector in real time, departments are no longer needed, and that the territory where that is true grows every year and never shrinks.

## The inversion

The nearest ancestor of the cockpit is not a dashboard but a loom. In the 1920s Sakichi Toyoda built an automatic loom that stopped itself the moment a thread broke, so that one person could watch many looms instead of standing over one.[^35] That idea became jidoka, one of the two pillars of the Toyota Production System, and on the assembly line it became the andon cord: any worker who sees a problem may stop the whole line, and the line stays stopped until the problem is solved.[^36] Toyota’s rule is that quality is protected by stopping. Nothing defective moves forward, and everything waits for the human.

The cockpit keeps half of this and inverts the other half. It keeps the machine that runs on its own and calls a human only when it must; that is why one aligner can hold what seven desks held, as one weaver could watch many looms. But it does not stop the line for everything. A reversible gate does not wait for the human: if she has not looked within the period the playbook sets, it opens, and her judgement arrives the next morning as a sample of what went through. Only the irreversible stops and waits, the accepted mandate and the signed-out candidate and the contract, because only those cannot be taken back. Toyota stops the line to protect quality because a car with a fault cannot be recalled from the customer’s driveway without great cost. A vector can let the reversible flow because a batch of prospects sent to the next agent can be recalled in a second. The rule is the same rule, applied to a different cost of being wrong: stop for what cannot be undone, and let everything else run.

That is the whole of what has changed. She is not watching the machine. The machine is telling her, at the hour she chose, what she needs to know, and asking her, at the gates, for the one thing it does not have.

The next chapter goes up one level, to the woman who holds a thousand of these rooms at once.

[^34]: Airbus describes its flight-deck design as a “dark cockpit” or “lights out” concept: with every system set for normal operation, the overhead-panel pushbuttons show no lights, and lights are colour-coded to indicate abnormal or fault conditions and selected or deactivated systems. See Airbus, “Flight Deck Philosophy” presentation to the ICAO AFI Flight Safety Symposium, 2017.

[^35]: Sakichi Toyoda’s Type-G automatic loom of 1924 stopped automatically when a thread broke, so that a single operator could tend many looms; Toyota traces the principle of jidoka to it (Toyota Industries, company history).

[^36]: Toyota describes andon as a tool of jidoka: any team member may stop the production line if they see a threat to quality, and the line halts until the problem is solved. See Toyota UK Magazine, “Andon – Toyota Production System guide.”

================================================================
# Chapter 7: The Control Room
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-7-the-control-room

The master’s room follows the same rule as the cockpit: one face, and nothing shown until asked. The difference is the floor.

A master holds one Outcome and every vector running on it, and the number of vectors may be three or may be more than a thousand.

That sentence would have been absurd to every management thinker of the last century, and they had arithmetic on their side. In 1933 a Lithuanian-born management consultant working in Paris, V. A. Graicunas, counted what a manager actually has to hold in her head: not only her relationship with each subordinate, but the relationships between them, and with every group of them that might act together. The count does not grow with the team. It explodes. With four subordinates a manager holds 44 relationships; with five, 100; with six, 222; with twelve, more than twenty-four thousand.[^37] Lyndall Urwick, the most influential management writer of his generation, turned the arithmetic into a rule that every business school still teaches in some form: no superior can directly supervise more than five or, at the most, six people whose work interlocks.[^38]

They were right, for the world they were in. A manager supervised people, and people’s work had to be watched, coordinated and carried across, and that cost attention that grew faster than the team. The whole pyramid of the old company was built on that limit: six people under each manager, six managers under each director, and so on up, until a company of ten thousand had five or six layers between the person at the top and the work.

The control room breaks the arithmetic, and it breaks it by changing what the master holds. She does not supervise people whose work interlocks. The interlocking is inside the chains, where the agents carry it, and the vectors do not need her to coordinate them with each other. What she holds is the state of each vector, shown as a colour, and the few things that need a human. Graicunas’s relationships do not arise, because there is no work passing between her pairs for her to carry. She does not read a thousand vectors. She reads the colour of a floor, and walks to the red.

## The floor

When she asks to see her Outcome, the floor becomes it: one tile of light for every vector, hundreds or a thousand of them stretching away like a city seen from above at night. Most are pale and steady, which means they are flowing. The ones with a warning are amber. The ones that need her now are red, and they are always nearest her feet, because the floor sorts itself. She does not scan. She reads the colour of the whole, which tells her the health of the Outcome in a glance, and then she reads the colour at her feet, which tells her where to go.

![Picture 8: The control room](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/08-the-control-room-1600.webp)

*Picture 8. Key. What: the control room; the floor is the Outcome. Who: the master (human, left) and the quartermaster of the red vector nearest her feet (agent, on the disc); it belongs to that vector, not to her. What is happening: every tile is a vector; pale is flowing, amber is a warning, red needs her now, and the red tiles have already moved nearest her. She is about to walk to them.*

She can go as deep as she likes, until there is nothing left to see. She steps onto any tile and she is in that vector’s cockpit, where its own quartermaster explains the situation and shows her its reports, exactly as it would show the aligner. From there she can ping the aligner or the tuner of that vector, by message or by voice, and check in. Nothing is hidden from her at any level. She chooses how far down to look, and most days she does not need to look far, because the vectors’ own quartermasters have already looked, and the floor is what they report.

What makes a tile amber or red is not one measure but several, each with its own colour inside the tile: results against target, the attention of the pair, a technical fault, cost running ahead of value, the satisfaction of the vector’s customers, and anything else the Outcome has learned to watch. Each of those colours is reported by an agent inside the vector itself: the quartermaster for the result and for the pair’s attention, the agent maker for a fault, the vector’s accounts agent for cost, its client agent for the customer. The floor is the vectors speaking, and nothing stands between them and her. The tile takes the colour of the worst thing inside it, so a red inside is red on the floor, and the master steps in to find which of the several it is. This is the rule of the tile, and it will reappear one level up, in the command centre, because it is the right way for any whole to show the state of its parts: the outside colour is never calmer than the worst thing inside. A rule that softened the colour on its way up would hide exactly what the floor exists to show.

## Nobody beside her

Nothing works for the master except the floor. There is no staff beside her, human or agent: no head of her Outcome, no HR, no quality head, no accountant, no technologist. She is the only human above the pairs in her Outcome, whatever its size; her standby is another master; and the agents she sees are the agents of her vectors, one at a time, when she steps onto their tiles.

Every function a division head once kept at her elbow is in the vectors instead, at the step where it is needed, and each is there by the one test this book applies to every piece of work: a skill is an agent, inside the vector that needs it; a result is an Outcome, with a master of its own. Nothing is a department, and nothing sits beside the master. Where each of the old functions went is worth setting out, because a reader who has run a division will look for them.

**Quality** is an inspector agent in every chain. It checks every output of its vector before the output passes a gate, around the clock, against the playbook, and is built not to miss anything; when it does, its agent maker rebuilds it. It may hold anything in its own vector. When it holds an output that the quartermaster passed, the output waits; either of the pair may appeal to the master, who decides. This is the only dispute inside an Outcome, and it has one court. What the inspector finds is one of the colours inside the tile, so the master sees quality on the floor without a quality department to report it to her.

**Cost** is an accounts agent in every chain. It keeps the cost of its vector and the value the vector delivers, and shows the difference as a colour in the tile. It pays the vector’s own costs, a freelancer, a tool, an agency that runs a step, within the limit the playbook sets for it; above the limit it asks the master, because money is an irreversible gate. It pays the aligner and the tuner by the formula set out in the chapter on pay, on the vector’s own result. The Outcome’s budget is the sum of its vectors’ budgets, and the master holds it against the one budget the Keeper gave her. There are no internal prices between Outcomes.

**People** are the vector’s own concern. A new aligner or tuner is found by an agent that runs the selection from application to offer and signed for by the vector’s tuner, as the chapter on selecting describes, and is trained by the vector itself, which teaches her the work from the first morning. An agent that has failed is decommissioned by the vector’s agent maker, which can restore it, so there is no gate. A human cannot be restored, so letting a human go is an irreversible gate, and the only hand on it is the master’s. The evidence for that decision comes from the work: the quartermaster measures attention from the gates answered and the reviews kept, and shows the tile amber weeks before the pair would say anything. The chapter on attention sets out what happens then.

**Technology** is the agent maker’s. Each vector’s agent maker watches for new tools, new models and new versions as they appear, and proposes each upgrade to its tuner, who accepts or refuses it for that vector, because the agent maker alone knows the local customisations of its vector and must fit the new technology around them. A better way found in one vector reaches the others through the playbook and only through the playbook: the master writes it in, every agent maker reads it, and each installs it in its own chain. Nothing spreads across vectors by any other path.

**The health of the Outcome** is the floor. There is no agent whose job is to watch the vectors for her, because the vectors watch themselves and report, and a watcher above them would be the first tile of a new pyramid. The next section says why.

Her controls are three: question a vector, pause it, stop it.

And one rule above all the others. The agents may read anything inside the engine and nothing outside it. That is the boundary, and it is absolute: no agent of the company reads an aligner’s private messages, her applications elsewhere, or her life. What the engine shows is enough. An aligner who is leaving stops answering gates as fast, lets the review streak break, and lets the quality of her judgement calls drift; her vector’s quartermaster reads those signals and shows the master the tile amber weeks in advance, so that a standby is ready before the gap opens. The prediction comes from the work, never from watching the person. I will say more about this in the chapter on attention, because it is the line between a company that measures and a company that surveils, and the doctrine is on one side of it.

## Why there is nobody beside her

People ask why the master has no staff, and the answer is the same as the answer to why the aligner has no team.

A human beside the master would be a layer. Not at first; at first she would be an assistant, a chief of staff, a deputy. But a deputy who reads the floor for the master becomes the person the aligners report to, and then the person who decides what the master sees, and then a manager, and the pyramid is growing again from the top. The master is alone above her pairs not because she is heroic but because the moment she is not alone, the company has a fourth layer, and four is the first step back to seven.

An agent beside the master would be a department. That is the subtler mistake, and the easier one to make, because an agent does not look like a layer. An agent that checks quality for every vector from outside them is a quality department with no people in it: it holds the vectors’ outputs, grows its own rules, needs its own budget, and on the day it needs a human to own it, a function has returned to the top of the Outcome with a hold over every result and responsibility for none. The same is true of an agent that keeps the Outcome’s accounts, or its people, or its technology. Each would be a skill standing outside the work it serves, which is the definition of a department. So each stands inside the work instead, in every vector, at the step where it is needed, behind a gate, and the master sees it as a colour on the floor. An Outcome has no functions. It has vectors, and a master.

Her standby is another master, and that is also not an accident. (In the small company of chapter 15, where there is only one master, her standby is a backup who has shared her room.) The only human who can hold an Outcome in a master’s absence is someone who already holds one, because holding an Outcome is a craft, and the craft is the same across Outcomes even where the chains differ. Two masters, each the other’s standby, each with the other’s playbook open, is the whole of the company’s insurance against a master’s absence, and it costs nothing, because both of them were already there.

## The rules of the control room

A master in her first month meets the same questions every master meets: how to split a vector that has grown, when to give an aligner a second one, how a new pair is trained, what to do when an inspector and a pair disagree, how she is paid, who stands by for her, what happens when her whole Outcome turns red. The doctrine answers all of them in advance, in eighteen rules. They are set out in one line each at the end of this chapter, for the master who wants them beside her. Most of them are easier to understand in use than in a list, so before the list, here is a control room in use.

## A morning at GoodSpace

The master of GoodSpace’s largest Outcome is a man named Mudassir, who was its delivery head before the word master existed in the company. His Outcome is the one the last chapter visited from the inside: a job description goes in, and an invoice comes out. It runs as ten vectors, each with an aligner and a tuner and about fifty open mandates, which makes some five hundred roles that clients are waiting for, four hundred or so of them live in any week, and between eighty and a hundred and twenty closed in a month. The mornings below are composites of several, with the clients and candidates changed; the decisions are the kind he makes every week.

Ten vectors is a small floor. It is not a thousand tiles; it is ten, and inside each of them the mandates show as smaller tiles of their own, so that when he looks down he sees ten pale squares, each made of fifty smaller ones, and the few that are not pale. On this morning three vectors’ tiles are red, one is amber, and the rest are pale. He walks to the red.

The first is a mandate that has failed the test the pairs set themselves. Three batches of candidates have gone to the client, interviews have been held, feedback has come back, and there is no selection. The aligner has already sat down with her quartermaster to find the leak, as the vector’s own rule requires, and has sent the question up, because the answer is not hers to give. The feedback trail is plain when he reads it. The client rejected the first batch for experience, the second for salary, and the third for reasons that contradict the first. The findability report that opened the mandate scored it a six; the client’s own behaviour has since made it a nine. Mudassir does not send a fourth batch. He has three choices, and he takes the hardest: he calls the client’s hiring head himself, puts the three rounds of feedback side by side, and asks which of the three people the client actually wants. It is the conversation the chain cannot have, because it is a conversation about the client’s own indecision, and only a person the client trusts can hold a mirror to it. The second red mandate is simpler. Its client has gone silent for three weeks after promising feedback, and Mudassir releases it to hold and frees the pair’s time for mandates that are moving.

The second thing is capacity. One vector carries fifty-eight mandates this week, because a large client opened eight roles at once, and its tile has gone amber on load; another carries forty-one, because a client paused hiring. He steps onto the heavy tile, and its quartermaster shows him the eight new roles and which pair in the Outcome has worked that client’s roles before: the pair on the light vector. He moves six mandates across, the two agent makers carry the mandates’ history between the chains, and by the afternoon both vectors are near fifty. This is a rule of the control room in action: the vector reports its load, the master moves the work, the agent makers carry the record.

The third is a colour, not a mandate. One client’s own tile, the customer’s colour that the floor shows beside the Outcome’s, has been drifting from pale to amber for a fortnight. Nothing has failed. But interviews are being rescheduled, feedback is slower, and the client’s last two messages were shorter than its first twenty. A customer leaving is the one thing in his Outcome he cannot fix after the fact, so he does not wait for a red. He puts a call in the diary for the afternoon, not to sell anything, but to ask whether something has changed at the client. It usually has.

The last is a dispute. The inspector in one vector held a candidate that the vector’s tuner had signed out the evening before: the candidate’s skill assessment was nine months old, and the playbook says a skill assessment older than six months is retaken before a candidate is forwarded. The tuner appealed. She had spoken to the candidate for twenty minutes, he was the strongest on the mandate, and the client wanted him seen this week. Mudassir reads both sides and upholds the hold. The rule is in the playbook for a reason, and he wrote it himself; a rule that bends for the strongest candidate is not a rule. The candidate takes the assessment that afternoon and, if he passes, goes to the client the next morning, one day later than the tuner wanted.

That is the morning. Two red mandates, one capacity move, one customer drifting, one dispute. Nothing else on ten vectors and four hundred live mandates needed him, and he knows it did not need him because the rest of the floor is pale. Not all of the agents inside his vectors are yet as complete as this chapter describes them; where one is missing, a report or a person still fills the gap, and the next chapter says so plainly. But the shape of his day is already the shape of the doctrine: he reads colours, walks to the red, and makes the few decisions that a human must make.

Will the same room hold a thousand vectors? The honest answer is that GoodSpace has held ten, and a thousand is the doctrine’s claim, not its record. But the arithmetic is on the claim’s side, for the opposite reason that it was on Graicunas’s. His relationships multiplied with every person added, because every person’s work had to be coordinated with every other’s. The control room’s load grows only with the exceptions, the reds that reach the master, and the exceptions per vector fall as the playbook improves, because every red the master resolves becomes a rule the vectors follow. A master with ten vectors and five reds a morning, and a playbook that halves its reds every year, will hold a hundred vectors with the same five reds in a few years, and a thousand after that. The limit is not the number of tiles. It is the number of things that turn red, and that is a number the master herself drives down. And if a floor outgrows her before the playbook catches up, if the reds reaching her week after week are more than her review can clear, the doctrine does not give her a staff. The Keeper splits the Outcome in two, each whole, each with its own master and a copy of the playbook, as a vector is split when its pair can no longer hold its gates. A master’s capacity is a limit to design around, not a reason to build a department beside her.

The next chapter goes up one more level, to the man who holds every control room at once.

## The eighteen rules of the control room

1. **The playbook.** Every Outcome carries a written design (chain, gates, authority limits, standards, health thresholds); the master owns it, every agent maker reads it, and a new vector opens from it in a day.
2. **Splitting and merging.** The master may split a vector that has grown or merge two that have thinned, without the Keeper; the agent maker clones itself into the child.
3. **Learning across vectors.** A better way found in one vector enters the playbook; every agent maker reads the playbook and installs the change in its own chain. Nothing spreads except through the playbook.
4. **The daily review.** Once a day, at an hour she fixes, she reads the floor in a fixed order: red vectors first, then the pairs’ attention, then quality, cost and the customers’ colour, each only where it needs her.
5. **Capacity.** A vector’s quartermaster reports its load; the master moves work between vectors; the agent makers carry the record across.
6. **Onboarding a pair.** A new aligner or tuner spends a fortnight under supervision, on a trial vector or co-signed on a live one, before holding a live vector alone.
7. **Agents and humans.** An agent maker may decommission an agent in its own vector, because an agent can be restored; only the master may let a human go, because a human cannot.
8. **The customer’s colour.** The floor shows the Outcome’s customers and their satisfaction as a colour of its own.
9. **The audit trail.** Every decision by every agent and human is logged and replayable; an agent that cannot show why it acted is stopped until it can.
10. **Quality disputes.** A vector’s inspector may hold any output of its vector against the playbook; either of the pair may appeal; only the master may overrule.
11. **The master’s standby.** Another master, holding a different Outcome, stands by for her, with her playbook always open.
12. **Who watches the master.** Her review, her speed on reds and the decisions waiting on her are her Outcome’s colour in the command centre.
13. **Targets.** The Keeper gives one result and one budget; the master sets everything beneath them.
14. **Authority.** Within budget and playbook she needs nobody’s permission; a sum beyond the budget or an act outside the playbook goes to the Keeper.
15. **Pay.** A fixed base at the intention premium, and a share that moves with the Outcome’s result against target.
16. **How a master is made.** Chosen by the Keeper from aligners who have run two or more vectors and stood tuner for others, after a month beside the outgoing master. At the collapse, when there are no aligners yet, the first masters are chosen by the four layers (chapter 13).
17. **When the Outcome goes red.** Within twenty-four hours the Keeper calls the master and her standby; the Keeper decides a new target, a new playbook or a new master. The Outcome is never paused and never runs without a master.
18. **When the floor outgrows her.** If the reds reaching the master exceed what her daily review can clear, week after week, the Keeper splits the Outcome in two, each whole and each with a master. The floor is never given a staff.

[^37]: V. A. Graicunas, “Relationship in Organization,” *Bulletin of the International Management Institute*, March 1933; reprinted in Luther Gulick and L. F. Urwick (eds.), *Papers on the Science of Administration* (Institute of Public Administration, Columbia University, 1937). Graicunas counted direct single, direct group and cross relationships; the maximum total for n subordinates is n(2^(n−1) + n − 1): 44 for four, 100 for five, 222 for six, and 24,708 for twelve.

[^38]: L. F. Urwick, *Scientific Principles and Organization* (American Management Association, 1938), and “The Manager’s Span of Control,” *Harvard Business Review*, May–June 1956. Urwick’s formulation was that no superior can supervise directly the work of more than five or, at the most, six subordinates whose work interlocks.

================================================================
# Chapter 8: The Command Centre
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-8-the-command-centre

The Keeper’s room is the quietest of the three, because the Keeper holds the fewest things.

That sentence surprises people, and it is meant to. The old company put the most at the top: the most reports, the most meetings, the most decisions, the most hours. The chief executive’s day was the busiest in the building, and the building was organised to feed it. The collapsed company inverts this, as it inverts most things. The aligner’s day is full, because the gates are hers. The master’s day is quieter, because only what the floor cannot handle reaches her. The Keeper’s day is the quietest of all, because he holds four numbers and a handful of tiles, and almost nothing that happens in the company needs him. What does need him, needs him absolutely, and the command centre is built so that those things, and only those, arrive. It helps to know how busy the old top was, because the contrast is the point. When two Harvard professors had the time of twenty-seven chief executives of large companies logged in fifteen-minute blocks for three months, they found the average chief executive worked about sixty-two and a half hours a week, and on most weekend days as well. Seventy-two per cent of that time was spent in meetings, and seventy per cent of the meetings lasted an hour or more. Only twenty-eight per cent of the working week was spent alone.[^39] These were not inefficient people. They were the most capable managers in their companies, doing what the old structure required of the person at its top: receiving the summaries that the layers beneath them prepared, and carrying decisions back down. The Keeper’s room exists to give those hours back.

## Four numbers, then the Outcomes

When the command centre opens, he sees four numbers.

They stand at the back of a long table as four columns of light: profit, customers, employees and society, each with a gold line for its target. These are the four core outcomes of the company, the things it exists for, and they belong to nobody but him. No Outcome holds them; every Outcome serves them. In front of the columns lie the Outcomes themselves, each one a tile of light, four of them at GoodSpace, more at a larger company, joined to the columns by threads of light that show which feeds which. A healthy Outcome glows pale and even. One whose customers are leaving turns slowly amber, months before it dies, and the Keeper sees the colour change long before anyone asks for a decision.

Three of the four columns need no explanation. The fourth does, because “society” is the word companies use when they mean nothing in particular, and a column that cannot turn amber is decoration. In this book the fourth column measures the people the company affects but does not sell to, in a few numbers the board chooses for that business and sets a gold line on. For a bank it might be the borrowers it lent to who were not harmed by the loan. For a plant it might be the town around it. For GoodSpace it is three things: people placed into work, candidates who received a fair and unbiased hearing whether or not they were placed, and how candidates rate the way they were treated. Those are not slogans; each is counted, every day, and each can fall below its line. That is what entitles the column to stand beside profit.

Nothing else appears until he touches something.

![Picture 9: The command centre at rest](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/09-the-command-centre-at-rest-1600.webp)

*Picture 9. Key. What: the command centre at rest. Who: the Keeper, the man in the dark t-shirt, with his two humans, the Chief of Staff and the Head of Outcomes; no agents on discs, because at rest nothing is asking for him. What is happening: the four tall columns are profit, customers, employees and society, each with a gold line for its target; the tiles in front are the Outcomes, one amber and one red; the blue threads show which Outcome feeds which column; the small tile at the end reads DOCTRINE 0 0 0, which is the only number in the room that must never change.*

He can touch any tile and stand in its control room, and from there touch any vector and stand in its cockpit. Nothing on the table was summarised by a person, and everything on it can be opened down to the agent that produced it. Proposals for new Outcomes come to him by voice, one at a time, from anywhere in the company. Disputes between masters come the same way. He can question, pause or stop any Outcome; settle any dispute; and do the one thing nobody else in the room can: start an Outcome.

## Designed from the top down

The control room was designed from the vector up: start with what a pair needs, and build the master’s room to hold a thousand of them. The command centre is designed from the Keeper down, in three levels. The first level is five decisions that fix what the Keeper is for. The second is how the room runs. The third is the detail a Keeper will need in his first year. The five decisions of the first level come here, because everything beneath follows from them; the rest are shown in use in the day below and listed at the end of the chapter.

### What the Keeper is for

**What the Keeper is for.** Four things at once, and nothing less. He is the allocator and the judge: he decides which Outcomes exist, gives each its one result and one budget, settles disputes between them and replaces masters. He is the owner of the four core outcomes, profit, customers, employees and society, which no Outcome holds and every Outcome serves. He is the designer of the whole: what the company should collapse into next, which Outcomes should be born, merged or allowed to die. And he is the keeper of the doctrine: that no layer regrows, no person is watched, no vector runs without a pair or a written exception. He runs nothing.

**What only the Keeper holds.** Four things no master may touch. The outside: the board, the investors, the law and society, to whom he is the company’s one face. The treasury: all cash and borrowing, until the collapse moves borrowing into a capital Outcome of its own (chapter 13), and the central budget from which every Outcome’s budget is cut. The common engine: the record, the audit trail and the technology every cockpit, control room and command centre runs on, which masters use and he owns. And the birth and death of Outcomes: only he opens one, funds its prototype, names its first master and closes it when its customers have left.

**Who keeps the Keeper.** The board sees his screen. It has read access to the command centre and sees the four core outcomes and every Outcome’s colour exactly as he does, with nothing summarised for it by a person. It meets twice a year to set his result and his budget, and when it must, to replace him; between those meetings it may question and never pause or stop. One named board member is his standby, as a master has a standby master: he knows the whole, and steps in when he is away or when the company’s own colour turns red.

**What he sees at rest.** Four numbers, then the Outcomes, as I have described, and nothing else until he touches something.

**Who sits beside him.** Two humans, and nobody else. Both are his own: a Chief of Staff, who runs the command centre as a room and holds the Keeper’s outside, and a Head of Outcomes, who holds the queue of what waits on him and the masters’ standbys when the Keeper is not looking. Neither is a layer; a master answers to the Keeper, never through them, and neither may sign anything that cannot be undone. No master sits beside him. Every master, whatever her Outcome, sits on the same table, as a tile among tiles.

I want to pause on that last sentence, because it is the one that readers who have run companies will press hardest, and it is where most attempts to reorganise a company quietly fail. Where, they ask, are legal, finance, HR, technology and quality? Every company has them. Surely they sit at the top, beside the chief executive, where they always have?

They do not, and the reason is the most important distinction in this book. For every such piece of work the Keeper asks one question: is it a result, or a skill? A legal review of a contract is not a result anyone outside the company receives; it is a skill used in producing a result, the signed mandate. So it is an agent, and it sits inside the chain of every vector that signs contracts, as a step with a gate after it, and the gate stays shut until the check passes. Put the legal, finance, quality and people checks in a department beside the Keeper and they become what they were in the old company: a set of people whose job is to stand between the work and the result and say no, judged by nobody, and soon managing the people they were meant to serve.

Some of this work, though, is a result. Raising capital from the market is a result: investors or lenders receive something, and the company receives money it did not have. Opening a new geography may be a result. Building a platform that other companies buy is a result. Where the work produces a result of its own, it is an Outcome like any other, with its own specialised vectors and a master who signed up with intention to deliver it, judged by its result and its customers. It sits on the same table as every other Outcome, with no seat beside the Keeper, and if it serves other Outcomes they are its customers.

The answer is decided case by case, by the outcome the company wants, and the same word can name either. Finance is an agent when it checks a price and an Outcome when it raises a round. What the question never produces is a third thing, a function with a seat at the top, and that is how the doctrine closes the door through which the old company walks back in.

Those five decisions define the role. Thirteen more rules govern how the room runs and what a Keeper needs in his first year: how the company’s colour is set, how an Outcome is born, who may hold what, how disputes are settled, what happens when he is away, how a Keeper is made. They are listed in one line each at the end of this chapter. Most are easier to see in use, and the day below shows several of them.

## The doctrine tile, and why it is there

I will say more about the last of these, because it is the one rule in the book that is aimed at the company’s own future rather than its present.

A collapsed company is under constant pressure to uncollapse. Not from enemies; from kindness. A master is overloaded, and someone suggests an assistant. An aligner is away, and nobody is quite ready, so a coordinator is hired to cover. A customer wants a report, and someone is given the job of writing reports. A board member would like someone to explain the screen, and an analyst is engaged. Each of these is reasonable, and each is a human who holds no vector, no Outcome and neither of the two seats beside the Keeper, and each is the first brick of a layer. Within three years the company has a head office, and within five it has departments again, and nobody decided it; it accumulated, one careful margin at a time, the way the stock on Mr Malhotra’s shelves did.

The doctrine tile counts the bricks. Three numbers, which should be zero: humans outside the structure; signals read from outside the engine, which is the surveillance the doctrine forbids; and vectors running with neither a pair nor a written exception, which is the standby the doctrine requires. The Keeper did not design the tile and cannot remove it, and any number but zero turns his own colour red and summons the board. It is the one piece of the command centre that is not for running the company. It is for keeping it collapsed.

## A day at GoodSpace

The Keeper of GoodSpace is Saksham, who took the seat from me. His command centre is the least finished of the three rooms at GoodSpace, and I will say plainly what is and is not built, because a book that shows only the finished parts of its proof is not offering proof. The events of the day are composites of several, with clients changed; the shape of the day is his.

His table has four tiles in front of the four columns. The largest is the Outcome the last two chapters visited, job description to invoice, ten vectors under Mudassir. Upstream of it is the Outcome that wins the mandates, from an ideal customer profile to signed mandates and job descriptions, running as eight vectors. Beside them are two small ones, each a single vector: one that turns a job seeker into a paying subscriber, and one that turns a job description into a set of perfect-fit candidates for clients who want to do their own hiring. The design of the company is that each Outcome will in time absorb the one upstream of it, until a single Outcome runs from the first contact with a client to the invoice; and that while it absorbs, the company adds products, geographies and segments. Holding that design is his job, and it is where most of his day goes.

The queue is short. One dispute has reached him, between the two large Outcomes. The master of Winning has signed a mandate from a large client; the findability report on it, when Delivery ran it, came back a nine. Delivery wants it sent back to the client for changes before it is accepted; Winning says the client will walk if pushed. The record puts the facts side by side: the findability report, the client’s history, the margin on the mandate. He decides within the hour, because the rule gives him twenty-four and the answer is not hard: Delivery’s gate stands, because the findability gate exists so that the company does not accept a promise it cannot keep, and a client who walks because he was told the truth about the market is a client the company would have lost later and at greater cost. Whatever the ruling costs Winning is written into Winning’s result, where Saksham will see the price of his own decision at the end of the month.

Then a proposal, which is where the day turns. An aligner in Winning has noticed that a growing share of the mandates her vector wins are from clients hiring outside India, and proposes a vector aimed at one overseas market. The doctrine’s answer is that trials make vectors and births make Outcomes: a new market is a vector, opened from a playbook that exists, and only a result with no owner is an Outcome. He says yes to a trial: one vector, opened from Winning’s playbook, with the proposing aligner as its aligner if she wants it, a small budget, and a quarter to prove it. At the end of the quarter he will decide again, one of three ways.

The rest of the day, which is most of it, he spends on the design. Not on any vector, and not on any mandate, but on the question of when the job description to invoice Outcome should absorb the small Outcome that produces perfect-fit candidates, which is in effect the same chain with the last third removed. He stands in both control rooms, reads both playbooks, and asks what would be lost. By the evening he has a plan to put to the board. That is the day: one dispute, one trial, and several hours designing what the company becomes next. Nobody needed him to approve a leave request, chair a review or read a report, because none of those things exist.

Here is what is not yet built. The checks that the doctrine puts inside the chains as agents, legal, finance, people and quality, are not all there yet. Saksham still carries people, technology and quality decisions himself, as they are still carried in most companies of GoodSpace’s size. Legal and finance are not his at all: I hold them, from the board, with an outside chartered accountant, Gautam Gupta, beside me. He does have a Chief of Staff, Rahul Dubey, who also carries accounts and administration, so one of the two human seats beside the Keeper is filled; the other, the Head of Outcomes, is not, and some of what should reach Saksham through the table still reaches him by message or in a meeting. The doctrine tile, if it were installed today, would not read zero on its first count. These are the parts of the doctrine GoodSpace has not yet reached, and each is named in the chapter where it belongs. What is already true is the shape: a Keeper whose operations run without him, spending his day on what the company should become.

## The seat above

There is one more seat in the building, and it is mine.

I sit on GoodSpace’s board. Twice a year the board sets the Keeper’s result, the weights on the four columns and the budget, and once in a while, if it must, it replaces him. Between those sittings the board does not run the chain. It looks, it questions, and it keeps the powers reserved to it: decisions above the Keeper’s authority, independent scrutiny, and the right to stop the company in an emergency. I open the same command centre Saksham opens, at the same moment, with nothing prepared for me. I see the four columns where he sees them, the four tiles, the dispute he settled and the price written into Winning’s result, the trial he opened. I may question anything, and I sometimes do. I do not pause or stop anything outside an emergency, and I have learned, which was not easy for a founder, that the discipline of the seat is to leave alone what is going well and to ask only about what is not.

I should also say where I break my own rule. Legal and finance are still mine, held from the board seat, because GoodSpace has not yet built the agents that should carry those checks inside its vectors. By the doctrine, a board that holds a function is a board that has quietly become a layer, and it is the first brick the doctrine tile would count. It is the largest gap in our own collapse. Closing it means putting the legal and finance checks into the chains as agents, where they belong; and if GoodSpace ever sets out to raise capital as a result in its own right, that will be an Outcome with a master of its own, on the same table as the rest.

I have held every seat in this building. I have stood at the gates of a vector as its aligner and rebuilt its chain as its tuner; I have read a floor as a master; I have kept the whole company; and now I read it from above. That is not a boast. It is the reason I am confident that the three rooms fit together, because I have sat in each of them and walked from one to the next. And it is the reason the board seat matters to the doctrine: the person who designed the rooms is bound by them like everyone else. If I want something changed, I ask the Keeper, at the sitting, and he decides.

Step back, then, and look at the whole building once. At the bottom, cockpits, each with an aligner, a tuner, one chain and a quartermaster who answers for it. In the middle, control rooms, each with a master, a floor of tiles and nobody beside her. At the top, one command centre, with one Keeper, four columns and a doctrine tile. Above it, a board that reads the same screen and touches nothing. Information rises as it happens; oversight descends as three verbs. There are two layers of management and there will never be a third, because the tile is counting.

The three rooms are not a description of a company that is finished. They are a description of what the removal of the old constraint makes possible, drawn precisely enough to be built, and being built. The next part of the book is the doctrine they rest on.

## The eighteen rules of the command centre

**Level one: what the Keeper is for**

1. **His job.** Allocator and judge, owner of the four core outcomes, designer of the whole, keeper of the doctrine. He runs nothing.
2. **What only he holds.** The outside, the treasury, the common engine, and the birth and death of Outcomes.
3. **Who keeps the Keeper.** The board reads the same screen, sits twice a year to set his result, weights and budget, may question, keeps its reserved decisions, and pauses or stops only in an emergency; one board member is his standby.
4. **At rest.** Four numbers, then the Outcomes, and nothing else until he touches something.
5. **Beside him.** Two humans only, the Chief of Staff and the Head of Outcomes; neither is a layer and neither signs the irreversible. Every master sits on the same table, and no function has a seat at the top.

**Level two: how the room runs**

6. **One truth.** No weekly or monthly report; every number flows up as it happens; the board sees what he sees.
7. **His contract.** No fixed hours; the engine reads his speed on reds, his scrutiny and the decisions waiting on him.
8. **Result and pay.** A weighted score on the four core outcomes; a base at the intention premium and a share that moves with the score.
9. **The company’s colour.** The company’s tile takes the colour of the worst thing inside it, as every tile does; when a core outcome is missed or most Outcomes are red, the board’s standby is called.
10. **Trials and births.** Trials make vectors, births make Outcomes. A new market or segment is a trial vector, opened from an existing playbook on a small budget. An Outcome is born only for a result with no owner, by the Keeper, whole (chapter 16).
11. **Outcome or agent.** Work that produces a result of its own is an Outcome, with vectors and a master; work that is a check or a skill (legal, compliance, finance checks, quality, people) is an agent inside the chains that need it, behind a gate that stays shut until the check passes. Appeals go to the Keeper within twenty-four hours. No Outcome may hold another’s work.

**Level three: the detail**

12. **Authority.** Inside the board’s budget and the doctrine he signs anything; beyond them, the board answers within seven days.
13. **Disputes.** Decided within twenty-four hours on the record’s facts; the cost is written into the losing Outcome’s result.
14. **Absence.** Reversible decisions open after the period the playbook sets; the Head of Outcomes holds the room; the irreversible waits for the board’s standby.
15. **How a Keeper is made.** From masters who have held an Outcome through a red, or from outside; a quarter beside the outgoing Keeper.
16. **The common engine.** The Keeper holds the shared engine every chain runs on; agents build and upgrade it, or, where the company decides a platform is a result in its own right, an Outcome like any other. Masters change only their playbooks and local customisations.
17. **The two humans.** Chief of Staff holds the outside, Head of Outcomes the inside; each signs only reversible acts.
18. **The doctrine tile.** Three counts that must read zero: humans outside the structure, signals read from outside the engine, vectors without a pair or a written exception. Any other number turns the Keeper red and summons the board.

[^39]: Michael E. Porter and Nitin Nohria, “How CEOs Manage Time,” *Harvard Business Review*, July–August 2018: 27 CEOs of large companies tracked in 15-minute increments for three months; average 62.5 hours a week; work on 79 per cent of weekend days; 72 per cent of work time in meetings; 70 per cent of meetings an hour or longer; 28 per cent of work time alone.

================================================================
# Chapter 9: The Vector
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-9-the-vector

Every company in this book is built from one thing, and this chapter is about how to build it.

A vector is the smallest unit that produces a result. It is one result, aimed at one market, carried by one chain of agents, and held by one pair of humans by default. Nothing smaller produces anything a customer would notice; an agent on its own is a skill, and a skill delivers nothing until it sits in a chain. Nothing larger is necessary; an Outcome is many vectors running the same playbook, and a company is many Outcomes. If you can build one vector well, you can build the company, because the company is that vector, copied and varied and held.

So I will build one, from nothing, the way a reader would build the first vector of her own collapsed company. I will take it from GoodSpace, but from the side of the company the last three chapters did not visit. Chapter 6 showed a vector that delivers mandates: a job description goes in, a placed candidate and an invoice come out. This chapter builds the vector that wins them: a description of the kind of client GoodSpace wants goes in, and a signed mandate with a job description comes out. It is the other half of the business, and it is the half most people assume cannot be run this way, because it is sales, and sales is supposed to be about people.

It is about people. That is exactly why it can be run this way.

## Begin with the result, at the right point

The first decision in building a vector is the one most companies get wrong, and it is not a technical decision. It is where to draw the result.

The obvious result for a winning vector is a signed mandate. A client agrees terms, signs, and sends a job description. It is a clean event, easy to count, and every sales team in the world is measured on something like it. And it is the wrong place to draw the line, because a signed mandate is not what GoodSpace’s customer feels, and it is not what GoodSpace is paid for.

Here is what happens to signed mandates in the real world. About half of them never reach a placed candidate. Some clients were never serious; they were shopping, signing with three firms to see what the market looked like. In some, the candidate who is selected does not join. In others another consultant fills the role first, or the client finds someone internally, or the position is put on hold and never taken off it. None of this is anyone’s fault in particular. It is the nature of hiring. But it means that a winning vector measured on signed mandates can hit every target it is given while the company earns half of what the targets imply, and the delivery vectors downstream fill up with mandates that will never close, and spend their mornings on them.

So the result of a winning vector cannot be drawn at the signature. It has to be drawn where the customer’s outcome is, and in recruitment that is a person who joins and stays, and an invoice that is paid. The winning vector does not produce that on its own; it produces half of it, and a delivery vector produces the other half. And so the doctrine draws the line around both of them.

At GoodSpace the design is one winning vector matched to one delivery vector, run as a pair of vectors and measured on one number: mandates closed and invoiced. The design target for each matched pair is twenty-five a month. Work backwards from that and every other number falls out. If about half of signed mandates reach fruition, twenty-five closed needs fifty won. Fifty won a month is the winning vector’s target; twenty-five closed is the delivery vector’s; and both are judged on the twenty-five, because the twenty-five is the only number the customer and the company both feel. A winning vector that brings in fifty mandates of which only ten close has not done its job, however many it signed; it has brought in shoppers. A delivery vector that closes fifteen of fifty good mandates has not done its job either. Neither can blame the other, because they share the result.

This is the first rule of building a vector, and the rest of the chapter rests on it. Design backwards from the result, and draw the result where the customer feels it. When the result sits between two vectors, size them together and measure them together. The old company drew a line between sales and delivery and put a department on each side of it, and each department hit its own number while the customer fell into the gap. The matched pair has no gap to fall into.

## The chain

With the result drawn, the chain is the sequence of steps that produces it, and each step is an agent. Here is the chain of a winning vector, built step by step.

**Find.** The first agent finds the companies that might need what GoodSpace sells: firms that are hiring, growing, funded, opening offices, posting roles. It works from the vector’s description of its ideal client and returns a list of companies and the people in them who hire. Today, at GoodSpace, the leads still come to each vector from marketing, which gives every vector its own leads so that no two vectors chase the same client. The design is that each vector finds its own, because finding leads is not a result anyone outside the company receives; it is the first skill in winning a mandate, and a skill belongs in the chain. The rule that no two vectors chase the same company does not need a department. It is one line in the master’s playbook, and the agents obey it.

**Qualify.** The second agent reads each company and decides whether it is worth approaching now: how many roles it is hiring for, how hard they are, whether it has used consultants before, whether it pays its bills. Most leads stop here, and they should.

**Research.** The third agent builds a short dossier on each qualified company: what it does, who leads it, what it has said in public about growth, what roles it has posted and for how long. The dossier is for the agents that follow and, later, for the human who will speak to the client.

**Approach.** The fourth agent writes the first message, in the client’s language and about the client’s problem, not GoodSpace’s product, and sends it through whichever channel the client is most likely to read. The fifth handles what comes back: answers questions, sends proof, follows up when there is silence, and knows when to stop.

**Book.** The sixth agent turns a reply into a meeting, and puts it in the diary of the agent that will hold it.

Then comes the conversation, and it is held by an agent.

**The first conversation.** Every new client has a real conversation before he signs, by voice or by video, and the conversation is held by an agent built for nothing else. Conversational agents are now good enough to do this well, and in one respect better than most salespeople: this one has read the dossier, knows every mandate GoodSpace has filled in the client’s industry, never forgets to ask a question, and never talks more than it listens. It tells the client plainly that it is an agent, and offers him a human whenever he wants one; most do not ask. What matters is what the conversation is for. A client who is about to hand a firm the hiring of someone who matters to him is deciding whether to trust it, and the firm is deciding whether the client is serious. Half of all signed mandates fail, and many fail because the client was only shopping; so the conversation agent is built, above all, to find out. It asks how many other firms are on the role, who decides, what happens if the role is filled internally, how soon the person must start, and what the last hire in the role cost them in time. It records the conversation, scores the client’s seriousness, and lays the record and the score in front of the human before anything is signed. The findability gate in the delivery vector protects the company from promises it cannot keep. The first conversation protects it from clients who are not really asking.

**Terms.** After the conversation, the seventh agent drafts the mandate: the advance, the fee, the exclusivity period, the payment terms, all from the playbook’s standard, with any exception marked. Here a skill agent stands at the gate: an agent that checks the terms against the company’s rules and the law, and flags anything outside them. It is not a legal department. It is a legal skill, placed at the one step where it is needed, and the gate after it stays shut until the check passes or a human with authority overrides it.

**The signature.** Here the chain reaches its human gate. The aligner reads the conversation agent’s record and its seriousness score, listens to the part of the call she needs to hear, and reads the mandate as it is read back. If the client is serious and the terms are right, she signs. If she is not sure, she can call the client herself before she does; that is her judgement, not a step in the chain. It is the one irreversible gate in the vector, because from that moment the firm has made a promise, and it is the one place in the winning chain where a human’s intention is required.

**Hand over.** The last agent takes the signed mandate and the job description and hands them to the matched delivery vector, whose first act is the findability report. The winning vector’s chain ends where the delivery vector’s chain begins, and because both are measured on the same twenty-five, the handover is not a wall but a seam.

Around the chain stand the two agents every vector has: the quartermaster, who answers for the whole of it to the human who asks, and the agent maker, who builds, tests and upgrades the others. That is the vector: nine line agents, one skill agent at a gate, a quartermaster and an agent maker, one human gate, and one result shared with the vector downstream.

## Gates follow the stakes

The chain has gates between every pair of agents, and almost all of them are reversible. A list of qualified companies can be sent back. A dossier can be redone. A first message that was not sent can be rewritten. These gates default open: the aligner samples what passed through them, and if she has not looked within the time the playbook sets, usually a day, they open on their own, and the chain does not wait for her.

One gate in the winning chain is different, and the doctrine treats it differently. The signature defaults closed and waits for a human however long it takes, because what follows it cannot be undone. Everything about the vector’s speed comes from the reversible gates, and everything about its safety comes from the irreversible one. The designer’s job is to know which is which, and to have as few of the second kind as the stakes allow. The human who holds an irreversible gate must also have the judgement it asks for: enough knowledge of the domain to know what her signature commits the firm to. That is a threshold, not a skill test at the door. A new aligner reaches it in the vector, beside her pair, and until she has, her irreversible gates are co-signed. A year ago the first conversation with a client would have been a second human gate, because no agent could hold it to standard. Now one can, and so the gate has moved, as gates do: the line between what the chain holds and what a human holds is drawn by what the machine can do to standard this year, and it has only ever moved in one direction.

## The pair

Two humans hold the vector, and the doctrine gives each of them one job and the other’s as a standby.

The aligner owns the result. She holds the daily review, samples the gates, listens to a sample of the agent’s conversations, and signs the mandates, and she answers for the fifty and, with the delivery vector’s aligner, for the twenty-five. The tuner builds and tunes the chain: she watches which first messages are answered and which are not, which qualified companies turn into calls and which turn into nothing, which signed mandates closed and which were shoppers, and she changes the agents so that the next month is better than the last. She also signs when the volume needs it, and because she does, she knows every client in the vector as well as the aligner does. When one is away, the other runs the vector without a day’s loss.

This is what the doctrine means by standby designed in, not improvised. In the old company, when the person who held a client went on leave, someone else was asked to cover, and the cover was a person who did not know the client, the history or the promises made. The client noticed. In a vector the standby is named before it is needed, and she already knows the chain, because she built it; and the clients, because she has spoken to them.

## No silent stalls

Henry Ford’s moving line had one property that every office in the century that followed failed to copy. When a station stopped, the line stopped, and everyone could see it. A stall could not hide.

The old office was built of silent stalls. A file sat on a desk for a week because its owner was busy, or on leave, or had decided to go and had not said so, and nobody downstream knew it was sitting there until a customer rang to ask. The stall was invisible because the work was invisible. A department could have twenty stalled files on any given morning and every report it sent upward would be true.

A vector cannot stall silently, because every agent in it reports its own flow, all the time. The find agent says how many companies it found today; the approach agent says how many messages it sent and how many were answered; the booking agent says how many conversations are in the diary. If any agent produces nothing for longer than the playbook allows, the quartermaster raises an alarm, the tile turns amber in the master’s control room, and the agent maker or the tuner is told before the aligner’s next review. A human gate that has waited too long is the same: a signature not given, a mandate signed and not handed over. The vector knows where everything is, and it says so. This is not a feature added to the design. It is the design: a vector that can stall silently is not a vector; it is a department with software.

## When a vector splits

A vector grows until its pair cannot hold its gates inside the hours they have committed to, and then it splits. The test is simple and is set in the playbook: when the signatures and the sampling of the gates no longer fit the daily contract of two people, the vector becomes two, the agent maker clones itself into the new one, and a new pair takes it.

There is a second reason to split, and it matters more as a company matures. A vector splits when its market divides. GoodSpace’s eight winning vectors today are identical. They talk to every kind of client, and the only thing that separates them is that each has its own leads. That is how a company should begin: copy the playbook, give each copy its own leads, and learn. But the vectors are learning, and what they are learning is that a manufacturer hiring a plant head and a software company hiring an engineering lead want different first messages, different proof, different conversations. As they learn it, they will specialise, one industry to a vector, each with a chain tuned to its market. A vector starts as a copy and becomes a specialist, and the company’s knowledge of its markets is the sum of its vectors’ specialisations.

## Count vectors, not employees

The old company measured its size in people. It reported headcount to its board, celebrated when it grew and worried when it shrank, and gave its managers status in proportion to the number of people beneath them. Headcount was the measure because people were the unit of production; a company did more by employing more.

In the collapsed company the unit of production is the vector, and so the measure of size is the number of vectors and the results each one carries. GoodSpace runs twenty vectors across its four Outcomes, each held by one or two people. That sentence describes the company better than any headcount could, because it says how many results the company is carrying at once, and with how much human intention behind each. Two people to a vector is the shape of the doctrine, and the default. A company that needs four people to a vector has not collapsed it. The playbook may write an exception, with its reason: a vector whose gates are few and slow may be held by one person with a named standby in the neighbouring vector, and a vector whose irreversible gates arrive faster than two people can sign may carry a third signer for a season. The exception is written, dated and read at the Monday test, and the doctrine tile counts any vector that is neither a pair nor a written exception. A vector held by one person doing both jobs, as some of GoodSpace’s still are, has no standby, and will find out on the day she is ill; it is a gap, and the company counts it as one.

A Keeper who wants the company to grow does not ask how many people to hire. He asks how many vectors the market will support, how many results each will carry, and how many pairs he can find whose intention he would pay for. Growth becomes the opening of vectors, each from a playbook, each with a pair; and the cost of growth becomes the cost of a pair and a chain, which is a fraction of the cost of a department, and is known in advance.

## The design and today

I have described the vector as it is designed. Here is the vector as it is today, because a doctrine that shows only its designs is asking to be believed rather than checked.

The design is fifty won and twenty-five closed for every matched pair of vectors. Today GoodSpace’s delivery vectors close eight to twelve mandates a month each, not twenty-five. Its winning vectors are not yet at fifty. And there are eight winning vectors and ten delivery vectors, which means the pairs are not yet matched one to one. The gap between the design and today is roughly a factor of two to three, and it is the most important number in the company.

Three things close it, and each is a change to a vector, not to a department. Delivery must get faster: more of each mandate’s candidates drawn from the pool of people already assessed, batches sent sooner, feedback turned into a sharper search the same day it arrives. The winning vectors must win better mandates, not only more of them: the conversation agent and the findability report filtering out the shoppers before they become half the failures. And as both improve, the share of signed mandates that reach a placed candidate will rise above half, so that fifty won produces more than twenty-five closed. Each of these is a tuning problem, owned by a tuner, visible every day on a tile, and none of them needs a meeting.

That is the vector: a result drawn where the customer feels it, a chain of agents with skills at the gates where they are needed, one human gate where intention matters, a pair who each can do the other’s job, an alarm on every stall, a rule for when to split, and a number that says honestly how far it is from its design. Everything else in this part of the book is about the people who hold vectors, how they are chosen and paid.

## Design a vector on one page

For the reader who wants to build one, here is the page. If you cannot fill every line, you are not ready to open the vector.

1. **The result.** What does the customer receive, and where exactly is the line drawn? If the result is shared with another vector, which one, and what is the shared number?
2. **The market.** Who is this vector for, and how does it avoid chasing the same customers as its neighbours?
3. **The target, worked backwards.** The result per month, and every number upstream of it, with the drop-off rate at each step.
4. **The chain.** Each agent, in order, with what it takes in and what it hands on.
5. **The skill agents.** Which checks the result needs (legal, compliance, finance, quality, verification), and at which step each one stands.
6. **The gates.** Which are reversible and default open, and after how long, and which are irreversible and wait for a human. For each human gate, write down what an agent would have to achieve to take it over.
7. **The pair.** Who aligns and who tunes, how many hours each commits to a day, and how the gates are shared.
8. **The standby.** Who runs the vector when either is away, and what she already knows.
9. **The alarms.** For each agent and each human gate, how long it may go without producing before the quartermaster raises a stall.
10. **The split rule.** At what volume, or at what division of the market, this vector becomes two.

================================================================
# Chapter 10: The Twelve Tenets
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-10-the-twelve-tenets

A doctrine is what is left when the stories have been told and the reader must decide what to do on Monday.

The nine chapters before this one were stories, some of them mine and some of them the century’s: the reorder book, the department as a workaround, the man who switched off, the motor in the basement, the cockpit at nine in the morning, the floor of tiles, the Keeper’s quiet room, the vector drawn where the customer feels it. Stories persuade. They do not instruct. A reader can agree with every one of them and go back to a company that is organised exactly as it was, because agreement costs nothing and a rule costs something. This chapter is the rules.

There are twelve. I did not set out to have twelve; I set out to write down every principle the collapsed company cannot do without, and to throw away every one it could. Twelve is what survived. Each is stated in one line, which is the line to remember. Each then has three short parts: why it holds, which is the constraint it answers; what it replaces, which is the practice of the old company it ends; and a test, one question an owner can ask of his own company on Monday morning and answer honestly by lunch. If you can pass the twelve tests, your company has collapsed, whatever it calls itself. If you cannot, it has not, whatever it calls itself.

One distinction first, because the earlier chapters stated some things as flatly as these twelve, and they are not of the same kind. A tenet is a principle: it holds in every company that collapses, and a company that breaks it has not collapsed. A setting is a number the doctrine gives as a starting value and the playbook changes on the evidence: how long a reversible gate waits before it opens, when a vector may hold more than its pair, how many reds a master can hold before her Outcome splits, the minimum standard a candidate must meet on each value, the threshold of judgement at which a new aligner signs alone. Where this book has given a setting a single number, it is the number I would start with, not a law. Change a setting when the record says so, and write down why. Change a tenet and you have built a department again.

## I. The department was a workaround.

**Why it holds.** Every function in every company exists because no single person could hold every skill an outcome needs. The forecaster could not also negotiate; the negotiator could not also draft the contract; so the outcome was cut into skills, and each skill was given a desk, and the desks were gathered into departments, and a manager was put above each department to carry the purpose the desks could not see. That was an honest answer to a real constraint, and it lasted two hundred years. The constraint has gone. A chain of agents holds every skill the outcome needs, in one system, without a handoff. When the constraint goes, the workaround has no reason left to exist.

**What it replaces.** The organisation chart drawn in boxes by skill: sales, operations, finance, HR, legal, technology. The belief that these boxes are the company, rather than scaffolding put up around it.

**The test.** Pick any department in your company and ask what result it produces that a customer would notice if it stopped. If the honest answer is “it helps the others produce theirs”, it is a workaround, and its work belongs inside the chains of the results it helps.

## II. Organise around outcomes, not activities. Work that produces a result is an Outcome; work that serves a result is an agent.

**Why it holds.** An outcome is a change the receiver feels: a mandate signed, a candidate who joins, a patient discharged, an invoice paid. An activity is a step towards one. The old company organised around activities because activities were what its people were good at; the collapsed company organises around outcomes because outcomes are what its customers pay for. The second sentence of the tenet is the one that keeps the first one true. Whenever a piece of work is found that does not obviously belong anywhere, ask whether it produces a result in its own right, one that someone receives and would miss, or whether it is a check or a skill used in producing someone else’s. Raising capital from the market is a result; it is an Outcome, with its own specialised vectors and a master who signed up to deliver it. Compliance is a skill. So is the finance check on a price, the legal review of a contract, the quality inspection of a batch, the screening of a hire. Each is an agent, standing inside the chain of every vector that needs it, at the step where it is needed. The same word, finance, can name an agent in one company and an Outcome in another. The test is never the name of the work; it is the outcome wanted.

**What it replaces.** The function. And, more dangerously, the function at the top: the legal head, the finance head, the head of people, sitting beside the chief executive with a hold over every result and responsibility for none.

**The test.** List every person who reports to the top of your company. For each, ask: what Outcome does she master? If the answer is a function rather than a result, the department has come back in through the front door.

## III. Every outcome runs on a vector.

**Why it holds.** A result is produced by something, and in the collapsed company that something is a vector: one result, aimed at one market, carried by one chain of agents, and held by one pair of humans by default. The aligner runs it and answers for what it delivers. The tuner builds and upgrades it and can run it alone. The vector is the smallest unit that produces anything a customer would notice, and nothing larger is needed to produce it. An Outcome is many vectors running one playbook; a company is many Outcomes.

**What it replaces.** The team, which was a group of humans dividing a result between them, each dependent on work he could not see. And the process, which was the department seen from the side: a sequence of steps across desks.

**The test.** For each result your company sells, can you point to the one chain that produces it end to end, and name the two people who hold it? If the result passes through three departments on its way to the customer, it has no vector yet.

## IV. Machines hold the baseline; humans supply the peaks.

**Why it holds.** Agents carry continuous work to a measured standard. They do not tire at the end of a shift, lose interest on a Friday, or switch off in the month before they resign. Humans do all of these, and also do what no agent can yet do: apply intention, urgency and judgement at the moments where the result turns. A company that asks humans to carry the baseline wastes them on the part of the work they do worst, and starves the peaks of the attention they need. The collapsed company gives the baseline to the chain and spends its humans only where a human changes the result: at the gates, at the signature, in the daily review.

**What it replaces.** The working day as a measure of contribution. The belief that more hours at a desk produce more result.

**The test.** For one vector, or one department if you have no vectors yet, count the hours its humans spent last week on work an agent could have carried to standard. Then count the hours they spent at the moments where only a human could have changed the result. The first number should be falling every quarter. If it is not, the machines are not holding the baseline; your people are.

## V. Pay for intention, not skill.

**Why it holds.** Skill is now cheap. The chain carries it, and the chain’s cost is a fraction of one salary. What is rare, and has always been rare, is intention that does not waver: the commitment to the result that holds on a bad week, in the month before the bonus, after the third rejection. One aligner with that intention, and a chain beneath her, produces what a department produced. She is worth to the company something like what the department was worth. So the doctrine pays her very highly, far above the market rate for her role. The level is deliberately high, for two reasons. At that price the company can choose from the very few whose intention holds. And at that output the price is still cheap, because the company needs a tenth of the people it needed before.

**What it replaces.** The salary band set by skill and seniority, and the old company’s habit of calling its people its greatest asset and paying them like a cost.

**The test.** Ask what you would pay the one person whose intention you would trust with a whole result, and whether you pay it. Then be honest about why not. GoodSpace does not yet pay at that level, and I will say why: its vectors do not yet produce enough to carry it. The design of chapter 9 is twenty-five mandates closed for every matched pair every month, and GoodSpace’s delivery vectors close eight to twelve. The pay follows the output; it does not lead it. But the principle is the right one, and a company that does not set a very high level of pay as its target will never find out what intention costs, or what it is worth.

## VI. No silent stalls.

**Why it holds.** Ford’s moving line had one property that the office never copied: when a station stopped, the line stopped, and everyone could see it. The office was built of silent stalls, files sitting on desks for a week while every report sent upward remained true. A vector cannot work that way. Every agent reports its own flow, all the time, and if any agent produces nothing for longer than the playbook allows, the quartermaster raises the alarm and the tile turns amber where the master can see it. A human gate that waits too long is the same. A vector that can stall silently is not a vector. It is a department with software.

**What it replaces.** The status meeting, the weekly report, the escalation, and the customer’s phone call asking where her order is. Each was a way of discovering a stall after it had cost something.

**The test.** Find the oldest piece of unfinished work in your company today. How long has it been waiting, and who knew before you asked? If nobody knew, you have silent stalls, and every report you received this week was true and useless.

## VII. Every vector has a standby.

**Why it holds.** People are ill, take leave, have children, and leave. The old company treated absence as an interruption and handled it by panic: the work waited, or a colleague who did not know it picked it up badly. The collapsed company designs the standby in. The tuner who builds the chain knows it end to end, and takes over when the aligner is away; the aligner takes over when the tuner is. Another master stands by for a master. A board member stands by for the Keeper. Gates that are reversible default open after a period the playbook sets, usually a day, so a missed day costs the vector one day of inspection, not one day of flow.

**What it replaces.** The handover note, the acting manager, the hiring in a crisis, and the single point of failure that every company has and few will name.

**The test.** Name the person in your company whose two-week absence would stop a result. Then name the person who would run it in her place, and ask whether that person could start tomorrow without a handover. A vector held by one person has no standby, and will discover it on the day she is ill.

## VIII. Count vectors, not employees.

**Why it holds.** The old company measured itself in people because people were its unit of production; it did more by employing more. The collapsed company’s unit of production is the vector, and so its size is the number of results it carries at once and the human intention behind each. The question at the top changes. The Keeper does not ask how many people he needs. He asks how many vectors the market will support, and how many each master can hold. This is span of outcomes, not span of control, and the design runs into the hundreds; when a floor holds more than its master can read, the Outcome splits.

**What it replaces.** Headcount as a measure of growth, of status and of a manager’s importance. The annual plan written as a hiring plan.

**The test.** Describe your company in one sentence without mentioning how many people work in it. If you cannot, you are still counting the wrong thing.

## IX. The aligner’s clock is her own.

**Why it holds.** The agents work round the clock, in the customer’s time zone. The aligner works in hers: where her attention is highest, when it is highest, for a number of hours she fixes in advance and then keeps. She chooses the hours; she does not choose whether to keep them, because the vector and the master need to know when its human will be there. For the first time since the factory bell, the worker’s clock is decoupled from the customer’s. The customer is served at three in the morning by the chain; the aligner reads what the chain did at nine, at the hour she chose, with her full attention.

**What it replaces.** Office hours, the shift, the attendance register as a proxy for work, and the commute as the price of being seen.

**The test.** Ask your best people when in the day their attention is highest, and then check whether that is when they do the work that needs it. If their best hours are spent in meetings and their judgement is applied at six in the evening, you are paying for intention and spending it on the clock.

## X. Hire for intention; measure attention.

**Why it holds.** Intention is the commitment to the result. It is selected for once, carefully, at a high price, because it is very hard to create and very easy to lose. Attention is that commitment showing up day by day, and it varies; everyone has a few bad days, and the ones with intention bounce back. When intention goes, the signs are written on the wall long before the resignation: the numbers dip, the leave increases, the person becomes unresponsive. The old company saw this late or never, because it watched people rather than results. The vector measures attention continuously, through the flow of the result: the reviews held, the gates answered, the batches sampled. The engine sees it first.

**What it replaces.** The annual appraisal, the attendance register, the manager who watches, and the exit interview that discovers what everyone already knew.

**The test.** Think of the last person who left your company of her own accord. How many months before her resignation could the data have told you she had switched off? If the answer is “none, we were surprised”, you are not measuring attention. You are measuring presence.

## XI. Measure, don’t price.

**Why it holds.** Outcomes serve each other without paying each other. The winning vector hands its mandate to the delivery vector, and neither sends the other an invoice. There is no transfer price, no internal profit and loss per Outcome, no internal market. Each Outcome has a real budget, held by its master, and real payments to the people and agencies who run its vectors; but between Outcomes, value is measured and not priced. The command centre sees what each Outcome delivers and what it costs, and the Keeper moves money between them on that evidence. A firm that prices every handoff becomes mercenary at every handoff, and spends its intention negotiating with itself.

**What it replaces.** The cost centre, the internal recharge, the transfer price and the profit centre, and the year-end argument over who earned what.

**The test.** Count the hours your managers spent last quarter negotiating budgets, recharges or credit with other parts of your own company. Every one of those hours was spent on a customer who does not exist.

## XII. A vector is a vector, whoever pays it.

**Why it holds.** A vector may be run by an employee, a contractor, a freelancer, a part-timer, a fractional pair or an outside agency, and several of these may run vectors inside one Outcome at the same time. A master may be fractional; in time the Keeper may be an agency. The form of payment makes nobody internal or external. The cockpit, the control room and the command centre do. Anyone who works from a cockpit, is read by the control room and is seen by the command centre is inside the company, and anyone who is not, is not. Membership is the screen, not the payroll.

**What it replaces.** The boundary of the firm drawn by the employment contract, and the distinction between staff and vendors that made half of every company’s work invisible to its own management.

**The test.** List every result your company depends on that is produced, wholly or partly, by someone you do not employ. Can you see each of them on the same screen as your own people, with the same alarms? If not, part of your company is outside the company, and you will learn about its stalls from your customers.

## Where GoodSpace holds them

I have used GoodSpace throughout this book as proof, and I have tried to be honest about where the proof is thin. Here I will name only the three tenets it holds most fully, because a doctrine is better served by three tenets held well than by twelve held in part. The tenets it holds partly or not yet are named, each where it belongs, in the chapters that discuss them.

| Tenet | Where GoodSpace holds it |
|---|---|
| **II. Organise around outcomes, not activities** | Held in the operating company: winning mandates is one Outcome, and delivering a mandate from job description to invoice is another, mastered by Mudassir. Neither is a department. Broken at the top, as the next section admits. |
| **IV. Machines hold the baseline; humans supply the peaks** | Agents search, assess and interview candidates around the clock; the humans’ half day is spent at the three gates, on the findability report, the client questionnaire and the personal sign-off on every candidate sent. |
| **VIII. Count vectors, not employees** | The company describes itself as twenty vectors across four Outcomes, each held by one or two people, and plans its growth as vectors opened, not people hired. |

Three of twelve, two of them held completely and one held below and broken above. It is less than I would like and more than any company I knew ten years ago could have claimed. The other nine are the work of the next few years, and the gaps are already on the tiles.

## The tenet that is broken first

One tenet is broken more often than any other, and it is broken quietly, by people with good intentions, usually within a year of the collapse. It is the second.

It does not happen through the chains. A vector that has been drawn around a result tends to stay drawn; its alarms, its tiles and its customers keep it honest. It happens at the top. A company collapses its operations into vectors and Outcomes, and then, because some kinds of work feel too important or too risky to leave inside a chain, it keeps one or two functions beside the Keeper. Legal stays there, because the contracts matter. Finance stays there, because the money matters. People stays there, because hiring matters. Each one is reasonable. Each is staffed by someone senior and trusted. And each, quietly, acquires what every function at the top has always acquired: a hold over every result, responsibility for none, and a reason to grow.

Within two years the hold becomes a queue, the queue becomes a team, the team becomes a department, and the department has a seat at the table where the Outcomes are decided. The old company has walked back in, through the front door, wearing the doctrine’s own clothes.

I know this because I have done it. As I described in chapter 8, I hold GoodSpace’s legal and finance myself, from the board, with our outside accountant. It was reasonable. It is also a function at the top, and it breaks the tenet I have asked you to hold above the others. The fix is not to find a better person to hold it. The fix is to ask the question of Tenet II and act on the answer. The legal review of a client’s contract is a check, so it belongs as an agent inside the winning chain, at the gate before the signature. The finance check on a price is a skill, so it belongs as an agent at the step where the price is set. Raising capital, if we do it, is a result, so it will be an Outcome, with its own vectors and a master who has signed up to deliver it. None of them is a seat beside the Keeper.

So watch the top. Once a quarter, list everyone who answers to the Keeper, and for each ask the question this chapter began with: what Outcome does she master? Where the answer is a function, the department has come back, and the collapse has begun to uncollapse. Put the work where it belongs, as an agent in a chain or as an Outcome with its own vectors, and empty the seat.

The other eleven tenets can be held badly for a while and recovered. The second, once it is lost at the top, takes the rest with it.

================================================================
# Chapter 11: Pay for Intention, Measure Attention
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-11-pay-for-intention-measure-attention

The old company bought time and hoped for attention. The collapsed company buys intention and measures attention. This chapter is about the difference, and about the few lines of an agreement that make it work.

Two of the twelve tenets carry it. The fifth says pay for intention, not skill. The tenth says hire for intention and measure attention. Stated as tenets they sound like sentiments, and every company already believes it pays for commitment and notices when it goes. It does neither. It pays for skill and seniority, in bands, and it notices commitment going at the exit interview. What follows is how a company actually does the two things the tenets say: how the pay is built, what is measured, what the aligner commits to each day, what happens when she is away, and what happens when her attention goes.

## The intention premium

Start with why the number is so high.

In the old company the price of a person was set by her skill and her years. A recruiter with five years’ experience in a city earned roughly what every other recruiter with five years in that city earned, because the company was buying the same thing every other company was buying: a set of skills, and eight hours a day in which to apply them. The skills were the scarce thing, and the market priced them.

The chain has made skill cheap. The search, the assessment, the first interview, the drafting, the chasing, the scheduling are all carried by agents whose cost is a fraction of one salary. What the chain cannot supply is the thing that was always rarer than skill and was never priced, because the old company had no way to buy it separately: intention that holds. The commitment to the result that is still there on a bad week, in the month after a bonus, after the third client rejection, on a day when nobody is looking. One aligner with that intention, and a chain beneath her, produces what a department produced. She is worth to the company something like what the department was worth.

So the doctrine pays her very highly for her role at target, far above what the market pays for it. The level is deliberately high, and it does two jobs. First, it changes who the company can choose from. At the market rate a company takes whoever applies and hopes; at a price far above it, it can wait for the few whose intention does not waver, and refuse everyone else. Second, it is still cheap. A company that needs a tenth of the people can pay each of them very highly and still spend far less than it spent before, before a single result is counted.

There is good evidence that how a company pays changes both what people do and who comes to work there. When the American windscreen company Safelite moved its installers from hourly wages to pay by the windscreen in the 1990s, output per worker rose by about 44 per cent. About half of the gain came from the same people working differently. The other half came from a change in who joined and who stayed.[^40] Pay is a filter before it is an incentive. The premium is set high for the filter.

## How the pay is built

The formula is the same for every human in the company, from the aligner to the Keeper, and it has two parts of equal size.

**A base, at the premium.** Half of the pay at target is fixed. It does not move with the month. It is the company’s payment for the intention it selected, and it is set high enough on its own to be well above anything the market pays for the role.

**A share, of the same size at target.** The other half moves with the vector’s result against its target: it rises when the vector delivers more, and falls when it delivers less. At target, base and share are equal. Above target the share keeps rising, and a pair whose vector carries much more than its target may earn more than the master who holds the Outcome. The structure expects this. Below target the share falls towards nothing, and the base is untouched.

The master is paid the same way, on her Outcome’s result against the Keeper’s target. The Keeper is paid the same way, on a weighted score of the four core outcomes against the board’s targets. Nobody is paid on someone else’s number. Nobody is paid on activity: not on calls made, candidates sent, hours logged or meetings held. The tuner is paid on the same vector as the aligner, with a base and a share of her own, because the result is theirs together and either may be running it on any given day.

Half the pay riding on the result is a strong pull, and a strong pull needs strong guards, because the one thing the company must never buy with its share is a bad signature. An aligner who is behind target on the twenty-eighth of the month has a reason to let a weak candidate through a gate. Three things stand against that reason. Her signature carries her name, and the record of every release she made is permanent. The vector’s own inspector agent samples what she released and reports it to her master as a separate colour inside her tile. And the result the share is paid on is drawn where the customer feels it: a placement that fails, or a client who does not pay, is not a result. A bad release does not earn the share. It costs it, a month later.

## Where GoodSpace stands

GoodSpace today pays the people who run its vectors a fixed salary and an incentive for each mandate closed. That is the right direction, and it is not yet the design.

It is the right direction because the incentive is paid on a result, a closed mandate, and not on activity. It falls short in two ways. The base is not yet at the premium, for the reason I gave in the last chapter: the vectors do not yet produce enough to carry it. And the incentive is paid closure by closure, not against the vector’s target. Arithmetically, a linear share of the vector’s result pays the same per closure as today’s incentive does; the difference that matters is what counts. Today a closure is counted when the candidate joins; the design counts the retained, paid result of chapter 9, and treats the vector as one result rather than a set of deals. The move from today to the design is therefore two moves, and both follow the output. As each vector’s monthly result rises towards the twenty-five of chapter 9, the base rises towards the premium, and the per-closure incentive becomes a share of the vector’s result against its target.

## Measure attention, not presence

Intention is selected once, at a high price. Attention is that intention showing up, day by day, and it varies. Everyone has bad days; the people with intention bounce back from them. When intention goes, attention goes with it, and it goes in a pattern that I have learned to see: the numbers dip, the leave increases, the person becomes slow to answer, and you can see it on the face before anything is said. Chapter 3 described that pattern, and the research that shows it can be seen months before a resignation.

The old company could not see it, because it measured the wrong thing. It measured presence: the register, the badge at the door, the hours at the desk, the green dot beside a name. It bought eight hours of presence a day and got, by the best measurements we have, about three of attention. The vector buys the three, and lets the agents cover the rest.

It measures them through the work. Beside the stall alarm, which reads the flow of the chain, each vector carries a second alarm that reads the aligner’s engagement with it. It reads only work signals, all of which the vector already knows. How quickly the gates were answered. How many decisions are waiting for her, and for how long. Whether the daily review was held, and how thoroughly. Whether the quality of her releases is holding, as the inspector agent samples it. Nothing about the person herself is observed: not her screen, not her location, not her hours, not her keystrokes. The result flowing, and the decisions being made, are the proof that attention is present. A company that has to watch its people to know whether they are working has not built a vector; it has built a department with a camera.

## The daily contract

Two things are fixed in the aligner’s day. Everything else is hers.

**Hours, chosen and kept.** The aligner chooses in advance how many hours a day she will give the vector, and when. Not set by the master, not the same for every aligner, not even the same every week; but chosen, and then kept, because the vector and the master need to know when its human will be there. At GoodSpace, as chapter 6 showed, the delivery pair’s day is a morning: the gates are cleared and the signatures given by one o’clock, and the chain works on through the afternoon and the night.

**A full review, every day.** At an hour she chooses, the aligner reads every alarm, clears every waiting decision, samples what the agents produced, and hands her tuner what needs to improve. The review is the one act the chain cannot perform for itself, and it is the evidence of attention. Results are judged by the month. Attention is judged by the day, and the review is how.

The master sets the number of hours her Outcome expects; the aligner sets when they fall. Neither sets how the aligner spends the rest of her day, and the doctrine has nothing to say about it.

## Gates, and absence

Chapter 6 set the two rules that keep the daily contract from being a cage, and they are terms of the agreement, so they are restated here in one line each. A reversible gate opens on its own if she has not answered it in the time the playbook sets for it, usually a day; an irreversible one waits for a human however long that takes, and the stall alarm says so. Before a day away she chooses one of three settings, pause, automatic, or her tuner, and names in advance which of them the vector takes if she simply does not appear. A missed day costs the vector one day of her judgement, not one day of its flow; what she loses is the chance to inspect what went through, and she gets it back the next morning.

## Form

Sometimes the attention alarm does not clear. The bad days become a bad week, and the bad week becomes a pattern. What happens then is the part of the doctrine that most people find hardest to accept and that I find most humane, because it is the part where the company is honest.

The word I use for it comes from cricket. A batsman in form scores runs; a batsman out of form does not, and everyone can see it, because the scoreboard is public. Nobody calls a meeting about it. The selectors rest him, and if form does not return he is dropped, and nobody is surprised, because the team exists to win, and he knew that when he walked out to bat.

The vector handles falling attention in the same three steps, in order.

**The mirror.** The attention alarm goes first to the aligner herself, and to nobody else, for a number of days the master sets in the playbook. She sees her own gates slowing, her own decisions waiting, her own review thinning. Most of the time this is enough. A bad week ends, privately, without anyone having watched it.

**Rest.** If the mirror does not clear it, the master rests her. The tuner runs the vector, which the tuner can do because the doctrine designed the standby in, and the aligner takes paid time away. Resting someone is the step every old company skipped, because in the old company the work stopped when the person stopped, so nobody could be rested; they could only be managed. In the collapsed company resting her costs the result almost nothing, and it gives the company’s most expensive asset, her intention, its best chance to come back.

**Replacement.** If form does not return after rest, she is replaced. There is no warning letter and no improvement plan. An improvement plan is supervision dressed as fairness: it asks a manager to watch a person, which is the job the doctrine abolished, and it measures effort at exactly the moment when the thing being paid for, intention, has gone. Replacement is quick, because the premium means the company is never short of people waiting to be chosen, and because the selection of the next chapter finds the next aligner in four working days. It is often a relief to the person replaced as well. Nobody whose intention has gone wants to keep pretending it has not.

The same three steps run one level up. The signals that show an aligner’s attention also show a master’s: whether her red tiles were answered, how many decisions wait on her, whether her own daily review was held. She sees them first, as a mirror. The Keeper rests her and her standby holds the Outcome. And if form does not return, the Keeper replaces her, with the same lack of drama.

## The aligner’s agreement on one page

The old company hired a person with a job description, kept her with an appraisal form and policed her with an attendance policy. The collapsed company needs one page, signed by the master and the aligner when the vector is handed over. If a line cannot be filled in, the vector is not ready to be handed over.

1. **The vector.** Which result, for which market, with which matched vector if the result is shared.
2. **The target.** The result per month, drawn where the customer feels it, and what does not count as a result.
3. **The base.** The fixed half, at the premium. It does not move with the month.
4. **The share.** Equal to the base at target; how it rises above target and falls below it; when it is paid.
5. **The hours.** How many a day, and when, chosen by the aligner and kept.
6. **The review.** The hour of the daily review, and what it must cover.
7. **The gates.** Which are hers to sign, which default open and after how long, and which never do.
8. **The absence setting.** What the vector does if she does not appear: automatic, tuner, or a named person in the control room.
9. **The standby.** The tuner’s name, and confirmation that she can run the vector alone from tomorrow.
10. **The attention signals.** What the attention alarm reads, and the plain statement that nothing else about her is observed.
11. **Form.** How many days the mirror is hers alone; how rest is decided and paid; and that replacement follows if form does not return.

Eleven lines. Everything the old company tried to manage through a manager is on them, and nothing on them needs one.

[^40]: Edward P. Lazear, “Performance Pay and Productivity,” *American Economic Review* 90, no. 5 (2000): 1346–1361. When Safelite Glass Corporation moved its windscreen installers from hourly wages to piece rates, output per worker rose by about 44 per cent; roughly half of the increase came from existing workers producing more, and the rest from changes in which workers were hired and stayed.

================================================================
# Chapter 12: Selecting for Intention
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-12-selecting-for-intention

## 1. The question that no longer needs asking

For two hundred years the first question put to a person who wanted work was: what can you do? The trade, the degree, the years in the same seat, the test at the door. The whole apparatus of selection, from the guild examination to the structured interview, existed to answer that question as cheaply as possible, because the answer was expensive to get wrong. A firm that hired a man who could not do the work paid for his learning with its customers’ patience.

The question is now answered before the candidate walks in. In a vector, skill is held by the agents. The JD Engineer knows how to write a findable job description; the Scout knows where the candidates are; the Assessor knows what a strong answer sounds like; the Closer knows the sixteen ways an offer dies in the notice period. Each of them knows it better on Monday than any new hire would know it after a year, and each of them teaches it to the human beside them, on demand, in the cockpit, all day. The aligner’s first week is not training. It is the vector running with her in it.

So the old question returns an answer that costs nothing and changes nothing. What can you do? The machine can do it. The only question left to ask is the one the old apparatus was never built to answer: what do you want?

This chapter is about asking that question properly. It is the most important hiring chapter in this book and also the shortest in method, because the method has one rule. Select for intention. Do not select for skill. Not as a floor, not as a tie-break, not as reassurance. Testing skill at the door selects for the person who already did the work the old way, and that is the person most likely to fight the machine for it.

## 2. Why skill at the door is a warning, not a comfort

Consider what a skills test actually measures. It measures the candidate’s past: the methods she learned, the shortcuts she keeps, the judgement she has about how work is done. In a department, that past was the asset. In a vector, the past is the thing she must put down on the first morning, because the agents hold a better version of it and will keep improving it without her.

The recruiter with twelve years of sourcing has a view of which portals work. The Scout has the record of which portals produced hires for this role class in the last ninety days. When they disagree, the vector needs the aligner to read the Scout’s reasoning and either accept it or tell the agent maker why it is wrong, with evidence. The twelve-year recruiter is the person least likely to do that calmly. She did not come to be taught by a tool.

The research on selection, read honestly, points the same way. Schmidt and Hunter’s meta-analysis of eighty-five years of selection studies found that work-sample tests and structured interviews predicted job performance better than almost anything else, with validities around 0.5.[^41] That finding was correct for the world it measured: a world where the human did the work. Its predictive power came from the fact that the sample was the job. When the job is to hold the gate and keep the vector honest, a sample of yesterday’s job predicts far less than it did, and the same meta-analysis quietly tells you what is left: integrity and conscientiousness, the measures in the table closest to intention.

Gallup’s figure from chapter 2 bears repeating here, because it is a selection failure as much as a management one. Twenty-one per cent of employees engaged. Four in five people who passed a skills test, sat in a seat, and did not want the result.[^42] No vector survives that ratio. One human per vector means the one must want it.

## 3. What intention is: the eight values

Intention is not a mood and not a speech. It is a set of things a person has already done, repeatedly, when nobody made her. GoodSpace names eight of them, and calls them, without embarrassment, the Goodness values: Teamwork, Friendliness, Helpfulness, Honesty, Reliability, Truthfulness, Integrity and Kindness.[^43] They are the same eight the Goodness Profile verifies for every candidate the delivery vectors present to a client; the firm applies to itself the test it sells.

A value is a word, and a word is easy to claim. So each value is tested by one behaviour: something a person has done that the record can show and a witness can describe. The eight, with the behaviour the selection agent scores for each:

1. **Teamwork, tested by whether she treats the machine as a colleague.** Has she used an agent, an assistant, an automated system, and let it change what she did? Not whether she likes it. Whether she let it. In a vector her team is a chain of agents and one other human.
2. **Friendliness, tested by whether she keeps a promise to a stranger.** Candidates and clients are strangers. The agent looks for evidence she kept commitments, warmly, to people who could not hold her to them.
3. **Helpfulness, tested by whether she wants the result.** She talks about the hire, the placement, the invoice paid, not about the role. Her past shows her finishing things for others that were not required.
4. **Honesty, tested by whether she tells the truth about numbers.** Her record shows her reporting bad weeks as bad weeks. The agent checks this against references, not against her account.
5. **Reliability, tested by whether she stays when it is hard.** Tenure through at least one period she describes as bad, and what she did in it.
6. **Truthfulness, tested by whether she accepts correction from a record.** When shown she was wrong, by data, she changes. Tested live: the agent shows her a mistaken decision of her own, documented, and scores the reply.
7. **Integrity, tested by whether she prefers the gate to the shortcut.** Asked for an example of a rule she bent for a good reason, the strong candidate struggles to find one.
8. **Kindness, tested by whether she does good without an audience.** Her best work is unwitnessed. Evidence from references of things done for others that nobody asked about.

None of these is skill. All of them are intention made visible by past behaviour. Each value has a minimum standard, written in the playbook: the least evidence of that behaviour the seat can accept. A candidate must meet the standard on every value; above it, the scores inform the human who decides, and do not decide for her. A person who meets the standard on all eight learns the cockpit in a week. A person who falls below it on two and is brilliant at sourcing breaks the vector in a month, because she falls below the standard on two, and one below is already a recommendation to refuse, which the tuner reviews.

## 4. How it is tested: four layers, one candidate at a time

The selection runs as a vector, because everything at GoodSpace runs as a vector. An agent holds the chain from application to offer. The vector’s own tuner holds one gate, the offer signature, and nothing else. No panel, no HR round, no culture-fit conversation with the master. The master sees the record if she wants it; she does not sit in.

The four layers are the four verification layers of the Goodness Profile, turned inward.

**Layer one, the declared profile.** The candidate answers the Goodness questionnaire: forced-choice, no free text, about forty minutes. Each of the eight values gets five situations, and in each she picks what she did, not what she would do. The agent scores it and, more usefully, flags the answers that are too good. A perfect declared profile is a warning in its own right.

**Layer two, the record.** The agent asks for evidence, not references yet: the documents, the messages, the outputs that show the declared value in a past situation. The candidate who says she reports bad weeks honestly is asked for a report of a bad week. Most candidates have never been asked this and are surprised by how little they can produce. The ones who want the result produce it within a day.

**Layer three, the witnesses.** The agent calls references itself, by voice, with a script that asks for incidents, not opinions. “Tell me a time she told you something you did not want to hear.” “Tell me a time she finished something you had not asked for.” Three witnesses, chosen by the agent from the record, not by the candidate from her friends. The transcripts go to the file; the scoring is per value.

**Layer four, the live test.** One hour in a cockpit, with a real mandate, not a case study. The agents run; she is asked to hold the gate. Somewhere in the hour the agent maker proposes an edit to a rubric, with reasoning. Somewhere else a candidate’s score is shown with an error the agent has made. The test is not whether she catches the error. It is what she does when she is told she missed it.

At the end, the selection agent writes a one-page verdict, value by value, with evidence cited per line against each value’s minimum standard, and a recommendation. Then a human reviews it. The tuner reads the verdict and the evidence behind it, listens to the parts of the witness calls and the live hour she needs to hear, and, where the case is close, speaks to the candidate herself. She signs, or does not, and where she departs from the agent’s recommendation she writes down why. That is the whole process. It takes four working days. It costs almost nothing.

## 5. Two candidates

In September GoodSpace ran this selection for an aligner seat in Mudassir’s JD-to-invoice Outcome. Two candidates reached the fourth layer. Both are described with their details altered; both have read this passage.[^44]

**The first** had nine years in agency recruitment and had closed, by her own count, over four hundred mandates. Her declared profile was strong across all eight values. Her record was thin on honesty: asked for a report of a bad week, she sent a quarterly summary with the bad month smoothed into it. Her witnesses were warm and could name no incident in which she had told them something unwelcome. In the live test she held the gate well for forty minutes. Then the agent maker proposed a change to the Scout’s portal weighting, with ninety days of evidence attached, and she overrode it, saying she knew the market. When shown, twelve minutes later, that the Scout’s weighting had been right on the last eleven roles of that class, she said the data was probably skewed.

The selection agent found her below the minimum standard on two values, Honesty and Truthfulness, cited the evidence above, and recommended against. The tuner read the record, listened to the twelve minutes of the live test herself, and signed the refusal.

**The second** had never recruited. She had run a small logistics desk for a courier firm for six years, with no staff, and had left when the firm closed. Her declared profile was uneven and honest; she answered “I did not” to four situations where most candidates lie. Her record was full: a year of weekly reports to an owner who did not read them, with bad weeks marked in red. Her witnesses, chosen by the agent, included a driver she had sacked, who said she had told him why, in writing, a month before. In the live test she did not know what a job description’s findability score meant, asked the agent, read the answer, and went on. When the agent maker proposed the Scout edit, she asked for the evidence, read it, and accepted. When shown her own missed error, she said, “Yes. What should I have looked at?”

Above the standard on all eight. The tuner reviewed the record, found nothing the agent had missed, and signed the offer. She ran her first mandate in week two and her first invoice went out in week six, inside the design target.

The lesson is not that experience is bad. It is that experience was not what the seat needed, and that testing for it would have hired the first candidate, who would have fought the vector until one of them lost.

![Picture 10: Selecting for intention](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/10-selecting-for-intention-1600.webp)

*Picture 10. Selecting for intention: the four layers, declared, record, witnesses, live, as one vertical chain ending at the verdict and the tuner’s signature.*

## 6. The offer as the last test

The pay formula in chapter 11 is the same for every seat: a base at a premium to market, and a share of the vector’s result against target, in equal halves at target. The aligner’s base is good. Her share, at target, roughly equals it; above target, exceeds it.

The offer conversation is held by the selection agent, by voice, before the signature. It is the last layer, and it is scored. The question the candidate asks first tells the vector what she wants. The one who asks about the base, the notice period, the leave policy, wants the job. The one who asks how the target is set, what counts as a result, what happens in a bad month, wants the result. The second candidate asked how a failed placement was counted. That is the question of a person who intends to be paid for things that stay done.

The firm does not negotiate the formula. It negotiates nothing. The offer is the same for everyone in the seat, and the signature is the tuner’s, on one page, the eleven lines from chapter 11.

## 7. The tuner: same test, no exception

A reasonable reader expects the tuner to be the exception. She judges the agent maker’s edits; surely she must know the work. She does not need to, and selecting her for skill carries the same risk as selecting the aligner for it, with more damage, because she touches the agents.

The tuner’s job is to decide whether a proposed change to a rubric or a skill is sound, on the evidence the agent maker attaches, against the vector’s result. That is a judgement about evidence and intention, not about recruiting. A tuner who knows recruiting well will be tempted, as the first candidate was, to override the evidence with her experience. A tuner selected for the eight values will read the evidence, test it against the result, and accept or refuse on that. The skill of tuning, which is a skill of reading a proposal and a number, she learns from the vector in the first month, as the aligner learns hers. What she must have before she signs alone is judgement in the domain of the gate: enough to know what a release commits the vector to. That is not tested at the door. It is a threshold the vector itself checks, and until a new tuner or aligner has passed it, the irreversible gates she holds are co-signed by her pair or her master.

In practice most tuners at GoodSpace have been aligners first. That is not a rule and is not a selection criterion. It is what happens when the aligner holds the eight values and the vector grows.

## 8. What this replaces

It replaces the interview panel, the skills test, the probation period, the culture-fit round, the HR screen, and the hiring manager’s instinct. Each of those existed to answer the old question cheaply. None of them answers the new one. The probation period in particular is a confession: it says the firm could not tell at the door and will find out at the firm’s expense. Four layers in four days tell you at the door.

It also replaces a belief: that a person’s worth to a firm is what she knows. In a collapsed firm her worth is what she wants, verified. That is a harder thing to fake and a kinder thing to be judged by. The nine-year recruiter was not refused for being bad at her trade. She was refused for wanting to be right more than she wanted the result, which the record showed and she confirmed.

## 9. Rules for selecting for intention

1. Skill is in the machine. Do not test it.
2. Select on the eight values, Teamwork, Friendliness, Helpfulness, Honesty, Reliability, Truthfulness, Integrity and Kindness, each tested by a behaviour the record can show, and only on them.
3. The selection is a vector. An agent runs it end to end.
4. One human gate: the vector’s tuner signs the offer. No panel.
5. Four layers, in order: declared, record, witnesses, live. No skipping.
6. Score each value on evidence cited per line. No overall impression.
7. Witnesses are chosen by the agent from the record, never by the candidate.
8. Ask witnesses for incidents, never for opinions.
9. A perfect declared profile is a flag, not a pass.
10. The live test includes a documented error of the candidate’s own. Score the reply, not the catch.
11. The live test includes an agent maker’s proposal with evidence. Score whether she reads it.
12. The offer conversation is scored. The first question tells you what she wants.
13. The formula is not negotiated. The offer is the same for every holder of the seat.
14. The tuner is selected on the same eight values, with no skill requirement.
15. Every value has a minimum standard in the playbook. For Honesty, for example: at least one documented instance, confirmed by a witness, of reporting a bad number unprompted. A candidate below the standard on any value is recommended for refusal; the tuner reviews the evidence herself and decides, and writes down her reason if she departs from the recommendation.
16. The refused candidate is told which values, with the evidence, in writing, within a day.
17. Four working days from application to signature. Longer means the vector is broken.
18. Nothing from a skills test, a degree, or years in the trade enters the verdict.
19. Judgement in the domain of a gate is a threshold, not a skill: until a new aligner or tuner has shown it in the vector, her irreversible gates are co-signed.

## 10. The selection worksheet

One page, kept in the vector’s record for every seat filled or refused.

1. Seat: which vector, which Outcome, aligner or tuner.
2. Candidate: name, date of application.
3. Teamwork (treats the machine as a colleague): score against the standard, evidence line.
4. Friendliness (keeps a promise to a stranger): score against the standard, evidence line.
5. Helpfulness (wants the result): score against the standard, evidence line.
6. Honesty (tells the truth about numbers): score against the standard, evidence line.
7. Reliability (stays when it is hard): score against the standard, evidence line.
8. Truthfulness (accepts correction from a record): score against the standard, evidence line.
9. Integrity (prefers the gate to the shortcut): score against the standard, evidence line.
10. Kindness (does good without an audience): score against the standard, evidence line.
11. Offer conversation: first question asked, verbatim.
12. Verdict: every value at or above its minimum standard, or recommended for refusal, with the values below it named.
13. Review and signature: the tuner’s review note, her reason if she departed from the recommendation, her name, the date. Four days or fewer from line 2.

[^41]: Frank L. Schmidt and John E. Hunter, “The Validity and Utility of Selection Methods in Personnel Psychology: Practical and Theoretical Implications of 85 Years of Research Findings,” *Psychological Bulletin* 124, no. 2 (1998): 262–274. Work-sample validity 0.54; structured interview 0.51; integrity tests 0.41; conscientiousness 0.31.

[^42]: Gallup, *State of the Global Workplace* 2025, global engagement 21 per cent; as cited in chapter 2.

[^43]: The eight Goodness values, in their canonical order, and the four verification layers are GoodSpace’s own, as used in the Goodness Profile presented to clients. The behaviour given under each value is the internal test the selection agent applies; the minimum standard for each is set in the playbook of the seat.

[^44]: The two candidates in section 5 are real, with details altered. The four-day duration and the week-six invoice are stated as GoodSpace’s own figures from September 2026.

================================================================
# Chapter 13: The Collapse Itself
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-13-the-collapse-itself

## 1. Revenue first, and whole

Twelve chapters have given the reader the parts. The vector and its cockpit, the gate, the agents, the master’s control room, the Keeper’s command centre, the twelve tenets, the pay, the selection. What they have not given is the order. A firm cannot collapse everything on one morning, and it cannot collapse one thing a quarter for five years and call the result a collapse. This chapter is the order. It has two rules before it has a calendar.

The first rule is revenue first. The Outcome that stands closest to the invoice is collapsed before any other, because its result is the one nobody in the building can argue with. When a quote that took nine days goes out in four hours and the order comes back, the owner holds a number that no department can explain away. Every department collapsed after that is collapsed in the shadow of that number. Collapse the plant first and the owner spends a year defending a cost saving. Collapse the revenue chain first and he spends the year explaining a growth figure. Chapter 10 warned that Tenet II is broken first by the functions kept at the top. Revenue first is the answer to the warning. It puts the result in the owner’s hands before his own desk comes up for the test.

The second rule is that the revenue chain goes whole. The chain from a customer’s first word to cash in the bank runs today through sales, the order desk, contracts, dispatch, invoicing, collections, and the forty people who carry messages between them. A chain with one department left standing in the middle of it is not half collapsed. It is dead. The vector on either side of the surviving department waits on it, and a vector that waits is a department with a new name. So the revenue chain is collapsed on one Monday, every department in it, into every vector it needs, and nothing else in the firm is touched that day. The rest follows one department at a time, each proven before the next, by the only test there is: the Outcome’s result for the week, with no department left in the chain.

## 2. The firm

The firm is imaginary, and has to be. No firm has finished what this chapter describes, and the one I know best has not finished either.[^45] Call it Pravaha Pumps. It makes industrial pumps and valves in a plant two hours from Delhi. It has a thousand people, counting the owner, a turnover of four hundred and eighty crore, and ships two hundred units a month against order, at an average of twenty lakh a pump. It is owned and run by one man, who is its board and its managing director at once, which is common and is this chapter’s last problem. It has ten departments, which is also common, and a hundred people in a corporate office whose work is keeping the ten departments informed about one another. Chapter 2 explained that office.[^46]

| Department | People | What it does all day |
|---|---|---|
| Sales and marketing | 90 | Enquiries, visits, quotes, the order desk |
| Production | 420 | 300 at machines and benches; 120 planning, supervising, chasing, recording |
| Procurement | 35 | Buying against the plan, and chasing the buying |
| Quality | 60 | Inspecting at inward, in process, and at dispatch |
| Finance and accounts | 45 | Invoicing and collections 20; treasury 10; statutory accounts, tax and audit 15 |
| Legal and secretarial | 10 | Contracts 6; secretarial and litigation 4 |
| HR and administration | 40 | Hiring, payroll, the law of employment, the canteen |
| IT | 25 | The systems, and the people who keep the systems talking to each other |
| Logistics and stores | 110 | Inbound stores 70; dispatch and outbound 40 |
| R&D and design | 65 | New products, changes, drawings for the plant |
| Corporate office | 100 | The owner, his directors, their assistants, the PMO, the MIS |
| **Total** | **1,000** | |

The plant runs one shift of eight hours, six days a week. Inside the shift its machines cut metal a little over half the time. Put together, the machines are working sixteen per cent of the hours in a week and standing for the rest.[^47] The reader should hold that figure to the end of the chapter, because four times sixteen is sixty-four, and sixty-four per cent of a week is what a machine does when it loads itself and nobody goes home. The fourfold output this chapter ends with is mostly not bought. The machine hours are standing in the plant today, switched off; what is bought is the ability to use them.

## 3. Month zero: the sort

Before one person moves, the ten departments are sorted on paper by the one test. Does the work produce an independent result, or is it a skill used in producing one? There is no third answer, and the sort takes an afternoon.

Sales produces a result: an order won. Outcome. The order desk, dispatch, invoicing and collections together produce a result: cash received against an order shipped. Outcome. Production, with stores, buying and inspection inside it, produces a result: a finished unit in stock from raw material. Outcome. Design produces a result: a first unit of a new or changed product, shipped against an order. Outcome. Treasury produces a result: capital on the balance sheet, on terms. Outcome.

Everything else is a skill. Buying is a skill the plant vectors use; it becomes a buying agent in each of them, raising purchase orders against the plan, with the aligner’s signature as the gate above a threshold. Inspection is a skill; it becomes the inspector agent in the line, and the human gate is batch release. Bookkeeping, invoicing, the credit check, tax on an invoice: skills, agents inside order-to-cash. Contracts: a skill, an agent inside the winning vectors, with the signature as its gate. Payroll, selection and the law of employment: skills; the selection vector of chapter 12 lives inside every vector, and payroll is an agent that runs the formula of chapter 11. Systems, software, the machines’ own software: skills, held by the agent maker and its tools in every vector. Statutory accounts, audit, filings, litigation: skills, agents inside the two Outcomes whose results they serve, cash and capital. The whole of HR, IT, quality, procurement, legal, and finance but for treasury, is agents. Not one of them keeps a department, a head, or a seat beside anyone.

The sort gives five Outcomes.

| Outcome | Its result | Vectors | Humans |
|---|---|---|---|
| Enquiry-to-order | An order won on the firm’s terms | 8, by market: utilities, process industry, OEMs, export, spares and service, three regions | 16 |
| Order-to-cash | Cash received against an order shipped | 14, by order type and value | 28 |
| Raw-material-to-finished-stock | A finished unit in stock against the plan | 16, by cell and product family | 32 |
| Design-to-first-unit | A first unit of a new or changed design, shipped | 6, by product family | 12 |
| Capital-to-balance-sheet | Capital raised on terms | 1 | 2 |
| **Five Outcomes** | | **45** | **90** |

![Picture 11: The collapse itself](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/11-the-collapse-itself-1600.webp)

*Picture 11. The collapse itself: ten departments above, five Outcomes below with their vector counts; a thousand people to a hundred, output four times.*

Ninety humans in cockpits, an aligner and a tuner to a vector. Five masters. The Keeper. The two beside him. Ninety-eight. The book calls it a hundred.

Month zero does four other things. The five masters are chosen by the four layers of chapter 12, in four days each. Six of the ten department heads apply and one is chosen; the other four are found among the directors of the corporate office, selected like anyone else, and the two who will stand beside the Keeper in month five are found there too. Every one of the ninety seats is designed on the ten-line worksheet of chapter 9, and its agents are built: a quartermaster, the line agents, an agent maker, for each vector. The agents of the revenue chain run beside the departments for the last two weeks of the month, on live enquiries and live orders, with nobody acting on them, so that on the Monday the firm switches a department off rather than an agent on. And on the first day of month zero, every one of the thousand is told, in one room, what will collapse in which month, what the ninety seats pay, that the seats are open to anyone through the four layers, and that most of the room will leave. The firm pays the leavers above what the law requires and says so in the same breath; what that costs, and the law that governs it, is the subject of section 10. What this section’s subject is, is that nobody learns the order of the collapse from a rumour.

## 4. The first Monday

On the first Monday of month one, sales, the order desk, contracts, dispatch, invoicing, collections, and the forty in the corporate office who served them, stop. A hundred and ninety-six posts. Twenty-two vectors start, eight winning and fourteen fulfilling. Forty-four humans, an aligner and a tuner in each cockpit, most of them from the building, selected on intention, some of them a month ago clerks.

The enquiry that arrives at nine is answered at nine by the vector’s conversation agent, as in chapter 9; the aligner reads the transcript before lunch. The quote that took nine days, because it passed through sales, design and finance for three signatures, takes four hours, because the pricing agent prices it against the plant vectors’ live capacity and the aligner signs. The order that was keyed three times, into sales, the order desk and the plant’s planning sheet, is keyed once by an agent and read by three vectors. The invoice goes out the day the unit ships, because the invoicing agent sits in the fulfilment vector and not in a department that batches on Fridays. The collections agent calls on the first day past due, not the fortieth.

By the last Friday of month one, the win rate on quotes, eighteen per cent for three years, is twenty-three. Order-to-invoice has gone from eleven days to one. The plant has not changed and ships two hundred and five units, because the order book, not the plant, had been the thing standing still. The owner has his number. He has spent nothing, and nobody in the building can argue with a quote that went out the same morning it came in.

Two of the twenty-two vectors fail their Monday test in week three: an export vector whose contract agent was built on the wrong trade terms, and a fulfilment vector whose dispatch agent could not read one transporter’s documents. Neither is given a person. Both are rebuilt by their agent makers inside a week, by chapter 11’s rule for a vector that fails: mirror, rest, rebuild, and no improvement plan.

## 5. Month three: the plant’s workaround

Month two is proof. Nothing is collapsed; the twenty-two vectors run, receivables fall from ninety-four days to seventy-one, and the master of order-to-cash shows the Keeper, on a Monday, a week’s result with no department anywhere in the chain. That is the test the plant must wait for, and it is the only reason the plant waits.

The plant is then collapsed in two steps, and the order of the two is the most important sequencing decision in the year. The hands at the machines go second. The people between the machines go first.

Of the four hundred and twenty in production, three hundred stand at a machine or a bench. The other hundred and twenty plan, schedule, supervise, chase and record. Around them stand thirty-five buyers, sixty inspectors, seventy storekeepers, and thirty from the corporate office. Three hundred and fifteen people whose work is the workaround chapter 2 described: keeping the parts of a department informed about one another. On the first Monday of month three, all three hundred and fifteen posts stop, and sixteen plant vectors start. Thirty-two humans, each cockpit holding a cell or a product family from casting in to finished unit out, its buying agent buying against the plan, its line agents scheduling the machines and reading the gauges, its aligner holding two gates: first article, and batch release.

The three hundred hands stay, for now, inside the plant vectors, and the chapter says so plainly, because the doctrine has one human per vector and this is the one place in the year that begins with more. They are a cost the vector carries and the calendar shows. They are not in the chain of decision. The aligner is.

What changes in month three is not the machines. It is that the machines are told what to do by an agent that knows the order book, the stock and the maintenance record at once, instead of by a planning sheet three days old. Lead time falls from fourteen weeks to nine in the first month, without a shift being added, because the queue in front of each machine, which was the sum of ten departments’ uncertainty about one another, is gone. Units shipped in month three: two hundred and thirty.

## 6. Months four to nine: design, the command centre, and the hands

Month four collapses design. Sixty-five engineers, and the ten in the corporate office who tracked their projects, become six vectors, twelve humans, each holding a product family from brief to first unit shipped. The agents draw, analyse and cost; the aligner holds the release. The firm’s third new product of the year ships in month seven, which is two more than it shipped in the three years before.

Month five collapses the last three departments that are only skills: HR, IT, and what remains of the corporate office. Seventy-seven posts, into agents that were already doing the work inside the vectors. The same month the Keeper takes his two humans, a Chief of Staff and a Head of Outcomes, and the command centre of chapter 8 is open. He has five masters reporting to him and nobody else, and the five masters have about forty-five vectors between them and no staff. There is no quality head beside the master of the plant; there is an inspector agent in each of her sixteen cockpits. There is no accountant beside the master of order-to-cash; there is an invoicing agent and a collections agent in each of her fourteen. The reader who has been waiting for a staff tier to appear at the master’s elbow should stop waiting. It does not come, because it was never a result.

Months six to nine are the hands. The plant vectors’ agent makers, not a capital committee, ask for machines by name: robotic loading and pallet pools on the forty principal machines, in-line gauging that stops the cut when the number is wrong, two assembly cells with collaborative arms for the small and mid-range pumps, and an automated test rig. Fifty-two crore between month three and month twelve, paid from the margin on the extra units, and from the term loans of month ten. None of it is new. Machines that load themselves and measure as they cut have been sold since the nineteen-eighties, and plants where machining and assembly run unattended through nights and weekends have been in the trade press for twenty years.[^48] Pravaha had not bought them because its machines stood idle eighty-four per cent of the week and no department could see a reason to make an idle machine work at night.

The hands fall as the cells commission: two hundred and twenty in month six, a hundred and sixty in month eight, a hundred in month nine. The second shift starts in month six and the third in month eight, and nobody is on them. Units shipped in month nine: five hundred and twenty. Lead time: three weeks.

Month six is also the month of the year’s one decision about price, and it is the decision that makes the rest of the arithmetic work. A plant that will soon make four times the pumps needs four times the orders, and the orders do not exist at the old price: Pravaha’s market bought two hundred of its pumps a month at twenty lakh before the collapse; the new revenue chain has taken that to under three hundred, and the plant will soon make eight hundred. So when the first cells commission and the cost of a unit begins to fall, the pricing agent proposes, and the master of enquiry-to-order signs, a list price a fifth lower: sixteen lakh for the pump that sold at twenty, for every order quoted from month six. A price a fifth below its old one, a lead time on its way from fourteen weeks to three, and a quote that comes back the same morning open doors the old Pravaha could never open: the OEMs that buy pumps as components, the export buyers who had priced it out, the process plants that had kept a second supplier for speed. The fourfold output is still the plant’s own idle hours switched on. The price cut is what lets the market take them.

## 7. Months ten to twelve: capital, and the owner’s own desk

Month ten collapses treasury into the fifth Outcome. One vector, two humans, whose result is capital on the balance sheet on terms: the bank lines, the term loans for the cells, and, in month twelve, the firm’s first outside equity, raised on the year’s numbers. Ten posts stop.

Month eleven is the owner’s. For ten months he has kept nineteen people outside the sort: statutory accounts, tax and audit, and the four in secretarial and litigation. He kept them for the reasons every owner keeps them. They are his. They report to him. The auditor knows them by name. A firm is a legal thing before it is a producing thing. And he is the board and the Keeper at once, so nobody above him applied the test to his own desk, which is exactly the failure chapter 10 named. In month eleven his own Monday test does it for him. The Head of Outcomes shows him the week’s result for the two Outcomes his nineteen people serve, cash and capital, and asks which part of either result the nineteen produce. They produce none. Statutory accounts are an agent inside order-to-cash. Tax is an agent in two vectors. Filings and the secretarial record are an agent in the capital vector. Litigation is an agent in whichever vector holds the contract. The nineteen posts stop in month eleven, and the owner’s desk is the last department in the building to go.

Month twelve commissions the second assembly cell, and the last hundred hands leave. A firm that cannot buy the second cell inside the year keeps them a quarter longer, and its count at month twelve is a hundred and ninety-eight, not ninety-eight. The order of the collapse does not change. Only the calendar does.

## 8. The year in numbers

| Month | Collapsed | People | Units/mo | What the number says |
|---|---|---|---|---|
| 0 | Nothing. The sort; five masters; ninety seats designed; agents built and run beside the departments | 1,000 | 200 | Everyone told on day one |
| 1 | Revenue chain: sales, order desk, contracts, dispatch, invoicing, collections, 40 corporate | 848 | 205 | 22 vectors; win rate 18% → 23%; order-to-invoice 11 days → 1 |
| 2 | Nothing | 848 | 215 | Two vectors rebuilt; receivables 94 → 71 days |
| 3 | Plant workaround: planning, supervision, buying, inspection, stores, 30 corporate | 565 | 230 | 16 vectors; lead time 14 → 9 weeks; 300 hands remain |
| 4 | Design | 502 | 260 | 6 vectors |
| 5 | HR, IT, rest of corporate office | 425 | 290 | Command centre open: Keeper, two humans, five masters, no staff |
| 6 | Nothing; first cells commission | 345 | 320 | Second shift, unattended; list price cut by a fifth; hands 220 |
| 7 | Nothing; cells run in | 315 | 370 | Hands 190; OEM and export enquiries arrive |
| 8 | Nothing; more cells | 285 | 430 | Third shift; hands 160 |
| 9 | Nothing | 225 | 520 | Hands 100; lead time 3 weeks |
| 10 | Treasury → capital Outcome | 217 | 600 | 1 vector; term loans for the cells |
| 11 | The owner’s statutory finance and legal | 198 | 700 | The last department |
| 12 | Nothing; second assembly cell; last hands leave | 98 | 800 | Run rate Rs 1,536 crore at Rs 16 lakh a pump; machines cutting 64% of the week |

A thousand people to ninety-eight. Two hundred units a month to eight hundred. Four hundred and eighty crore to a run rate of fifteen hundred and thirty-six crore: four times the pumps at four-fifths of the price, a little over three times the money. The payroll at target pay fell by more than a third, from sixty crore a year to a run rate of about thirty-eight, while pay per head went from six lakh to nearly forty: ninety seats at a base of eighteen lakh and a share equal to it at target, five masters, the Keeper and the two beside him. Shares earned above target come out of the result they are earned on, not out of this sum, which is Tenet V, intention very highly paid, by the formula of chapter 11 and not by a budget. Receivables thirty-eight days. Lead time three weeks. Win rate thirty-four per cent. And a corporate office of a hundred replaced by the Keeper, the two who sit beside him, and the five who run the Outcomes.

## 9. A firm that is not a pump maker

The order does not change with the trade. Sort every department by the one test. Collapse the revenue chain first and whole, on one Monday, after its agents have run beside it for two weeks. Collapse the Outcome that costs most next, its workaround before its hands. Collapse the departments that were only skills when their agents are already doing the work. Collapse treasury into capital. Collapse the owner’s own desk, which the test said should have gone in month one and which goes last because he is the one who holds the test.

A services firm has no hands and its year is shorter; the whole of it is workaround, and the machines it needs it already has. A firm with three plants collapses three plants in the same month, by the same vectors, because a plant is a vector family and not a department. A firm whose owner has hired a Keeper and sits above him on a board applies the test to the board in month zero, and chapter 8’s seat above is empty by month one, which is the one way this chapter’s last problem does not arise.

## 10. The nine hundred

Chapter 2 promised that the doctrine would say what it owes the people a collapse releases, and this is the place to say it, because this is the chapter in which nine hundred of a thousand leave. I have put it after the numbers and before the rules on purpose. A reader who has followed the year month by month knows exactly who the leavers are: the sales department of ninety, most of whom were not chosen for the seats; the three hundred hands whose cells now load themselves; the hundred and twenty who planned, supervised, chased and recorded; the corporate office of a hundred; the owner’s nineteen. They are not an abstraction by now. They have desks the reader can picture.

The first thing the doctrine owes them is the truth, early, in one room, which is why the sort is announced on the first Monday and not discovered from a rumour. The second is the seats. Every one of the ninety is open to anyone in the building through the four layers of chapter 12, and nothing in a person’s grade or department counts for or against her. At Pravaha, as the table shows, most of the pairs who held the revenue chain in month one had been in the sales department the month before, and the plant aligners had run cells. The people who hold intention were already inside the company. The collapse does not import them; it finds them, pays them far more than they were ever paid, and asks the rest to leave.

The third thing it owes is money, and here the doctrine has a rule of its own that goes beyond the law. The law first, as it stands as this book goes to print, because an owner who collapses a plant in India will meet it on the first Monday and should not meet it unprepared. Since November 2025 the Industrial Relations Code governs retrenchment. For a worker with a year’s continuous service, it requires a month’s notice or pay in its place, compensation of fifteen days’ average pay for every completed year of service and any part over six months, and fifteen days’ wages credited by the employer to the worker’s re-skilling account within forty-five days. For an industrial establishment that has averaged three hundred or more workers over the past twelve months, which Pravaha has, retrenchment needs the prior permission of the appropriate government and three months’ notice or pay in its place, and lay-off and closure have procedures of their own; permission is a decision the government may refuse. The Code’s word is *worker*, and it reaches further than the shop floor: clerical, technical and supervisory staff can be workers, and only those employed mainly in a managerial or administrative capacity, or supervising above a wage line the Code sets, fall outside it. So the hundreds on the floor are the hardest part of the year, but much of the office is covered too. This is a sketch and not advice. The thresholds and the procedure vary with the state, with the kind of establishment, and with the year, and the owner takes counsel before the first Monday, not after. But the shape is this: the hands need permission, notice and a paid season; the office needs its contracts honoured; and a collapse that forgets the difference stops in month three with a notice from the labour commissioner, which is not a stall the cockpit can show.

The doctrine’s own rule is simpler than the law and larger. The payroll saving, at the rate it runs by month twelve, belongs to the people who leave before it belongs to the owner. At Pravaha the payroll runs at sixty crore a year before the collapse and about thirty-eight at target pay by month twelve, and the owner sets aside a year of that difference, twenty-two crore, as a fixed transition allocation, paid out, over and above what the Code requires, to the nine hundred, in proportion to their years, and the owner’s gain from the payroll saving begins in year two. There are three reasons for the rule, and none of them is kindness. The first is that the money was theirs: the department was a workaround, but the people inside it did the work, for years, in a structure that was not their choice, and the year the machine holds the skills is the year their work is finally measured and found to have been real. The second is that a collapse paid for by its leavers poisons the ninety who stay. The pairs will hold the vectors for a decade; they will not hold them well in a building whose former occupants were sent out with the legal minimum by an owner who booked the saving the same quarter. The third is that the rule makes the owner honest about the arithmetic. A collapse that only pays if the leavers are paid badly is not a collapse; it is a layoff with new vocabulary, and the doctrine has no interest in it. Four times the output from a tenth of the people, even at four-fifths of the price, pays for the year many times over. If it does not, the sort was wrong, and the owner should read section 3 again before he reads this one.

What the doctrine does not owe them is the pretence that the jobs will come back, or that an equal number of new ones will appear in the same town. Chapter 2 said that the manual frontier that absorbed four men in 1989 is smaller now and shrinks every year, and nothing in this chapter changes that. Some of the nine hundred will hold vectors in other companies, because the four layers select for intention and intention is not rare only at Pravaha. Some will run the small businesses that a town of pump makers has always needed and that a severance can start. Some will not find anything as good, and the book does not pretend otherwise. That is the honest account, and the owner should give it in the room on the first Monday, with the number, because the alternative is that the nine hundred learn it from the rumour the doctrine was built to replace.

The last thing to say is the thing a recruitment company’s founder is obliged to say. GoodSpace sells hiring, and this book argues that a company needs a tenth of the hires it makes. Those two sentences are both true and I hold them together without discomfort, because the hires a company of vectors makes are the ones that matter: the ninety, chosen for intention at a price that reflects it, selected through four layers that cost more and refuse more than any interview ever did. The hiring that disappears is the hiring that was never selection at all, the filling of seats in a workaround, and I do not mourn it. If this book costs GoodSpace nine mandates in ten, the tenth will be the one worth having, and the firm has organised itself on exactly that bet.

## 11. Rules for the order of collapse

1. Sort every department on paper before touching one: skill → agent, result → Outcome. No third category.
2. Name the Outcomes by their result, from the customer’s first word to cash. Five, give or take two.
3. Revenue first. The chain from enquiry to cash collapses on one Monday, whole.
4. Never collapse half a chain. A department left standing in the middle of an Outcome kills it.
5. After the revenue chain, one department at a time, each proven on its Monday test before the next.
6. The plant’s workaround before the plant’s hands. Planners, chasers, checkers, buyers go first; machines replace hands second, bought from the margin on the extra units and from term loans.
7. Choose the masters in month zero, by the test of chapter 12. No department head keeps a seat by right.
8. Every seat is filled by selection for intention. Nobody is reassigned.
9. Tell everyone on day one what collapses when, and what the seats pay.
10. Pay the leavers above the law and do not argue about it.
11. Keep no function at the top. Finance, legal, HR, IT, quality: agent or Outcome, like everything else.
12. The master has no staff. A head beside her is a department that survived.
13. Beside the Keeper, two humans. No committee runs a collapse.
14. Build the agents before the Monday, run them beside the department for two weeks, and switch the department off, not the agents on.
15. Measure each vector’s result from its first day. The department’s old numbers are not the baseline; the Outcome’s result is.
16. Buy no machine a vector has not asked for by name through its agent maker.
17. A vector that fails its Monday test is rebuilt by its agent maker, never staffed up.
18. The owner’s own desk goes through the same test, and if he will not apply it, his Head of Outcomes does.
19. Twelve months, or a quarter more if the second cell is late. Longer is a reorganisation, not a collapse.

## 12. The collapse worksheet

One page, signed by the Keeper before month zero begins.

1. The firm: people today, output today, turnover today.
2. The departments, each with its headcount.
3. Each department sorted: agent, or Outcome with its result in one line.
4. The Outcomes, three to seven, named by result.
5. The revenue chain: which Outcomes, and the date of the Monday.
6. The order of the rest, with the proof each must show before the next.
7. The masters: names, each selected by the four layers.
8. Vectors per Outcome, each with its ten-line worksheet from chapter 9.
9. The seats: aligner and tuner per vector, plus masters, Keeper, two; the total at month twelve.
10. Output and price at month twelve, and the pay formula for every seat.
11. Machines to buy, by vector, by month, paid from margin and term loans.
12. The owner’s own functions: sorted, with their month.
13. The Keeper’s signature, and the date of the first Monday.

[^45]: Pravaha Pumps is invented. Its figures are a composite, built from firms the author has seen and from the GoodSpace numbers of chapters 6 to 9, scaled. No real firm is described, and GoodSpace’s own collapse is not finished, which is why it is not the example here.

[^46]: Gary Hamel and Michele Zanini, “Excess Management Is Costing the U.S. $3 Trillion Per Year,” *Harvard Business Review*, 5 September 2016: 23.8 million managers, first-line supervisors and administrators, one for every 4.7 employees. Cited in chapter 2.

[^47]: The firm’s own arithmetic. One shift of eight hours, six days: 48 of the 168 hours in a week. Inside the shift the machines cut metal 55 per cent of the time. 48 × 0.55 ÷ 168 = 16 per cent. For the benchmark the plant is measured against, Seiichi Nakajima, *Introduction to TPM* (Productivity Press, 1988), sets world-class overall equipment effectiveness at 85 per cent.

[^48]: FANUC’s plant at Oshino, below Mount Fuji, has run with robots assembling robots, and machining unattended through nights and weekends, and has been described in the trade and general press for two decades.

================================================================
# Chapter 14: The Ladder Is Gone
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-14-the-ladder-is-gone

## 1. The question on every reader’s desk

Somewhere around chapter 7 the reader stopped asking whether the collapsed company could work and started asking a smaller question, the one that matters to him personally: then what happens to my career?

He is right to ask it, because the answer is not a reassurance. The career he has in mind does not survive. It was a climb, and it was a climb because the old company was a pyramid. Every layer of management was a rung, every rung had a title, a band and a bigger desk, and the way to be paid more was to be given more people. A career was the number of people between you and the work. The collapsed company has two layers of management and a doctrine tile that counts any third. There is nothing above the aligner but the master, nothing above the master but the Keeper, and nothing above the Keeper but a board that may look and not touch. The ladder is not shortened. It is gone.

This is a loss, and the chapter will not pretend otherwise. But it is the loss of a bad thing, and the bad thing has been documented for fifty years. In 1969 Laurence Peter observed that in a hierarchy every employee tends to rise to his level of incompetence, because each promotion rewards the job just done rather than the job about to begin, and a person stops being promoted only when he reaches a job he does badly.[^49] It was meant as satire and turned out to be measurement. Fifty years later three economists took the records of 38,843 salespeople at 131 firms, with 1,553 promotions into management, and tested it. The best salespeople were the ones promoted: higher sales raised the chance of promotion markedly. And the better the salesperson, on average, the worse the manager: higher past sales in the new manager went with lower sales among the people who now reported to him.[^50] The ladder, in other words, took the firm’s best producer off the work, put him over people, and lost on both counts. That is the career the reader is grieving. The collapsed company offers him a different one, and the different one pays better.

## 2. Deeper

A career in a collapsed company goes deeper, not higher, and “deeper” has an exact meaning.

It means the same vector producing a larger result each passing month. Not a second vector, not a bigger team, not a title. The vector a pair opened in their first week is the vector they hold in their tenth year, and what has changed is what it carries. The enquiry-to-order vector that opened as one of eight winning vectors carries a fifth of the company’s orders in its fifth year, and all of them in its tenth, by which time the other seven winning vectors have merged into it one by one, under chapter 7’s second rule, because a vector that has learned to win in one market learns the next market faster than a new vector can. When one vector carries the whole company’s result, it is as deep as it can go inside the company. Then the company lends it. Another company in the same trade, in a market the first does not serve, runs the vector as its own: the same agents, the same playbook, the same pair at the gates. The vector’s result now includes what it wins for the other company, and the pair’s share, which is a share of the result, keeps growing exactly as it did when the result was all at home. A career does not end at the company’s edge. The vector crosses it, and the pair go with it without leaving their cockpit.

![Picture 12: The ladder is gone](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/12-the-ladder-is-gone-1600.webp)

*Picture 12. The ladder is gone: the ladder on its side; one vector drawn as a shaft going down, producing a larger result each year and crossing the line of the company’s edge with the same pair.*

That is the whole shape. One vector, deepening for as long as the pair hold it, from a corner of one company to the whole of it to the whole of several. The reader who has spent twenty years counting the people beneath him will find it strange for about a week. After that he will find it obvious, because it is what he always wanted to be paid for and never was: the result, growing, in his own hands.

## 3. Why the vector deepens, and what the pair have to do with it

The deepening is designed. It is not a hope, and it is not heroism. The agents in a vector compound: the agent maker proposes an improvement to a rubric every week, the tuner accepts the ones the evidence supports, and a vector that accepts fifty small improvements a year wins more in December than it did in January without anyone working harder. The playbook sets the rate at which this is expected to happen, the design rate, and the agent maker reports the vector’s actual rate against it, as one more colour inside the tile.

So the vector grows by design, and the pair’s part is to let it. That sounds like nothing and is everything. Chapter 12 selected the pair for intention; chapter 11 measured their attention through the work; this chapter is where the two are read over years rather than days. A pair whose attention holds accept the agent maker’s proposals on the evidence, hold the gates at the speed the design assumes, and read the Monday test honestly. Their vector grows at the design rate or above it. A pair whose attention is going begin to leave proposals unanswered, let the review streak break, and bend a rule for the strongest candidate; their vector grows below the design rate, and the gap between the two rates is their form, visible to the master as a colour long before any result is missed. Nothing in this is read from the person. The gap is in the work, which is where the doctrine looks, and nowhere else.

Form, then, is selected for at the door and read from the gap every month after. There is no appraisal in a collapsed company. The result is the appraisal, and it is paid monthly, which is the next section.

## 4. Pay follows the vector

The formula from chapter 11 does not change as the vector deepens, and that is the point of it. A base at the premium, fixed. A share, equal to the base when the vector is at its design target, and moving with the result from there. The share is set once, when the vector is designed, as a fraction of the result small enough that the company keeps almost all of every rupee of growth, and it is not moved afterwards. The company does not raise the target because the pair have beaten it. Deepening is paid, not re-targeted; a company that re-targets has told its pairs that the way to be paid is to grow slowly.

Here is what that produces over ten years, for one aligner on one vector at Pravaha Pumps, the firm of chapter 13. Call her Meera Kulkarni: an engineer from the Pune belt who had sold pumps for Pravaha for eleven years before the collapse, who came up through the order desk and knew every valve in the catalogue by the way it failed, and who had never, in eleven years, promised a delivery date the plant could not keep. She came through the four layers in month zero above the standard on all eight values. The vector is the enquiry-to-order vector for the process industry in the west, one of eight winning vectors when it opens in the first year after the collapse. Its design target is Rs 18 crore of margin a year on the orders it wins. The aligner’s base is Rs 18 lakh; her share at target is Rs 18 lakh; her tuner is paid the same. At the design target the two of them cost the vector Rs 72 lakh, four per cent of its margin, and she is very highly paid: far more than the market paid the sales manager she replaced.[^51]

| Year | × design | What happened | Aligner’s pay |
|---|---|---|---|
| 1 | 1.0 | Vector opens; at target by month nine | Rs 36 lakh |
| 2 | 1.6 | Win rate 18% → 26% as the pricing and Scout rubrics compound | Rs 47 lakh |
| 3 | 2.4 | Absorbs the north region’s vector; its pair open the export trial | Rs 61 lakh |
| 4 | 3.6 | Two larger clients won on the same quote rhythm | Rs 83 lakh |
| 5 | 5.0 | Absorbs the OEM vector | Rs 1.08 crore |
| 6 | 7.0 | Offered the master’s seat; refuses (section 8) | Rs 1.44 crore |
| 7 | 10 | Absorbs the last two in the west; one vector for west and OEM | Rs 1.98 crore |
| 8 | 14 | Lent to a pump maker in a market Pravaha does not serve | Rs 2.70 crore |
| 9 | 19 | Lent to a second; absorbs the east | Rs 3.60 crore |
| 10 | 25.6 | One vector for all of Pravaha: Rs 461 crore through one cockpit | Rs 4.79 crore |

Three things to read off the table.

First, at year ten the company is paying two people a little under Rs 10 crore for Rs 461 crore of margin, more than three times the eight winning vectors’ combined design target of Rs 144 crore. Pay has followed the vector, and the company has kept ninety-eight rupees of every hundred the vector added.

Second, the aligner out-earns her master from year seven. The master of the winning Outcome is paid by the same formula on the Outcome’s result: a base of Rs 40 lakh, a share equal to it when the Outcome is at its design target of eight vectors at target. By year ten the Outcome’s result is a little over three times its design, so the master earns Rs 40 lakh and a share of Rs 1.28 crore, Rs 1.68 crore in all, against the aligner’s Rs 4.79 crore. The doctrine wants this. A company in which the person over the work must be paid more than the person doing it has already decided that the work is the lesser thing, and will shortly have a ladder again.

Third, the lending is priced the way the pair are paid. The other company does not buy the vector or hire the pair; it pays Pravaha a share of the orders the vector wins for it, and that share enters the vector’s result. The pair’s share rises with it. Nobody negotiated a new contract with them when the vector crossed the edge, because nothing about their contract depends on where the result is won.

## 5. The pairs the deepening sets free

A reader will have noticed the arithmetic of section 4 from the other side. If one vector absorbs seven, fourteen people have lost their vectors. What happened to them?

They went to the edge. The pair of the north region’s vector, absorbed in year three, opened the export trial that chapter 8’s Keeper would have approved: one vector, from the winning playbook, a quarter to prove it. The OEM pair, absorbed in year five, had already flagged the plant’s spare capacity from inside their own cockpit; its tuner became the aligner of the vector that sells that capacity to other manufacturers, which became Pravaha’s sixth Outcome, as chapter 16 tells. The last two western pairs, absorbed in year seven, were the ones who took the deep vector to its first outside client and stayed as that client’s tuners for a year while its own pair were selected. Of the fourteen, two left, paid out as chapter 13 pays leavers. The company’s human count did not rise across the ten years and its result rose more than twenty-five-fold in one vector, which is chapter 9’s rule read over a decade: count vectors, not employees, and when a vector deepens, the people it frees are the ones who open what the company becomes next.

This is the answer to the fear that deepening means fewer careers. It means fewer careers in the middle and more at the edge, and the edge is where the next deep vector starts.

## 6. Masters and Keepers: a step sideways

The reader’s last reflex is to see the master’s seat as the top of a ladder the chapter has otherwise removed. It is not, for three reasons.

A master holds a floor instead of a chain. That is a different craft, not a higher one: she reads colours, walks to the red, decides the few things a human must decide, and owns a playbook. Some of the best aligners have no wish to do it and should not be asked twice. She is selected for it by the four layers of chapter 12, from among those who have held a vector, because holding one is how the craft of reading a floor is learned; she is not promoted to it by time served or by the size of her last result, which is precisely the mistake the 131 firms made.

She is paid by the same formula on a larger result, and a deep vector can out-earn her, as section 4 showed. The seat carries no premium of its own. Nobody takes it for the money, which leaves only the people who want the craft, which is who should have it.

And the step is reversible. A master who finds she misses the chain goes back to a vector, at a vector’s pay, with nothing lost but a different view, and the company records it as a move, not a fall. The Keeper’s seat is the same craft one level up, held by a person who has read a floor and now reads the table of them, and it is sideways from a master’s seat in exactly the same way. The three crafts are three ways of holding a result, not three heights, and a company that lets people move between them without ceremony will find that its masters are the ones who wanted to be masters.

## 7. The rules of a career

1. A career is one vector, deepening. More vectors is not deeper; a larger result from the same vector is.
2. Deeper has a limit inside the company: one vector carrying the company’s whole result. The company may lend the vector to a market it does not serve before that limit is reached, and the career continues across the edge.
3. The share is set when the vector is designed and never moved. Deepening is paid, not re-targeted.
4. The vector grows by design; the agent maker reports its actual rate against the design rate, and the gap is the pair’s form.
5. Form is read from the work: the gap, the gates, the review streak. Never from the person.
6. There is no appraisal. The month’s result, paid, is the appraisal.
7. Pairs freed by a merge go to the edge: a trial, a new market, a lent copy. Nobody is reassigned to a desk.
8. A master’s seat is sideways: a craft, selected for, paid by the same formula, and left again without loss.
9. A deep vector’s aligner may out-earn her master and her Keeper, and the company says so in the offer.
10. Nobody is promoted. Nobody is given people. Nobody rises; vectors deepen.

## 8. The refusal

In the sixth year of the collapse the master of Pravaha’s winning Outcome went to the Keeper and asked to go back to a vector. She had held the floor for six years, since month zero, it was pale most mornings, and she missed the chain. The Keeper agreed, as the doctrine requires, and the four layers were run for the seat. Meera came through them above the standard on every value, which surprised nobody, and the Keeper put the seat to her.

She asked for a week, which the record shows she did not need, and refused.

Her reasons, as the selection agent recorded them, were three. The vector had absorbed the OEM work the year before and was seven times its design; the agent maker had forty proposals in the queue and she wanted to read them. The Keeper had approved the first lending, to a pump maker abroad, two years on, and she wanted to take the vector there herself. And the chain offered the greater future upside, which she said last and least, because it was the least of it: it was the work she wanted to keep.

The Keeper selected a master from the aligners whose vectors had merged, a woman who had held the north region’s vector and then the export trial and wanted the floor. The deep vector went abroad in year eight with Meera at the gates, and to a second company in year nine. In her tenth year Meera’s vector carried Rs 461 crore of margin, her share was Rs 4.61 crore, and she had never managed anyone.

That is a career lived deeper. It had no ladder in it, no title that changed, no year in which she was given people. What it had was a result that was hers, that grew every month she held it, that paid her more than anyone above her, and that did not stop at the edge of the company that started it. The reader who is still grieving the old career should put the two side by side and ask which one he would choose for his daughter.

[^49]: Laurence J. Peter and Raymond Hull, *The Peter Principle: Why Things Always Go Wrong* (William Morrow, 1969).

[^50]: Alan Benson, Danielle Li and Kelly Shue, “Promotions and the Peter Principle,” *Quarterly Journal of Economics* 134, no. 4 (2019): 2085–2134; earlier as NBER Working Paper 24343 (2018), with a larger, earlier sample of 214 firms. The published study covers 38,843 sales workers at 131 firms and 1,553 promotions into management. The authors find that stronger sales performance predicts promotion and that stronger pre-promotion sales are associated with lower value added by the new manager to his subordinates’ sales; these are modelled associations in their sample (accepted manuscript of 29 July 2019, pp. 14 and 17).

[^51]: The multiples measure this vector against its own design target, set in month zero on the old order book. The growth of chapter 13’s first year, from the price cut and the new OEM and replacement markets, arrives through the vectors opened for it, which this vector later absorbs. The rupee column and the pay column show the domestic result only; the lending fees Pravaha earns, a share of the orders won for the borrower and not the borrower’s margin, are on top, and the pair’s share of them is paid separately; after that first year Pravaha’s own market stays roughly flat, and what grows is the share of it this one vector carries; the growth beyond that in the last rows is the lending outside. The Pravaha figures are illustrative and follow chapter 13’s: a design target of Rs 18 crore of margin a year per winning vector (an eighth of Rs 480 crore of turnover at 30 per cent margin), a result growing a little faster than it would by doubling every two years and merging in the neighbouring vectors as it goes, and Rs 461 crore of margin at year ten, which is chapter 13’s run rate of Rs 1,536 crore at the same 30 per cent margin: the illustration assumes that unit cost falls by the same proportion as price (from about Rs 14 lakh to about Rs 11.2 lakh a pump), preserving the 30 per cent margin. Margin here is after materials, payroll at target, agents and shared infrastructure, maintenance and depreciation of the cells, and before interest and tax; the Rs 52 crore of cells is cash in the year it is spent. For pay following output, Edward P. Lazear, “Performance Pay and Productivity,” *American Economic Review* 90, no. 5 (2000): 1346–1361, as cited in chapter 11.

================================================================
# Chapter 15: The Last Six
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-15-the-last-six

## 1. Why a company has edges at all

In 1937 a twenty-six-year-old economist asked a question nobody had thought needed asking: if the market is so good at coordinating work, why do firms exist? Why is anything done inside a company by instruction rather than outside it by contract? Ronald Coase’s answer was that using the market costs something. Finding the other party, agreeing the price, writing the contract, checking it was kept: each is a transaction cost, and where those costs exceed the cost of simply telling someone what to do, a firm forms around the work. The edge of the firm sits where the two costs meet.[^52] Oliver Williamson spent a career on the same line, and his finding was that the edge moves: whenever the cost of a handoff across a boundary rises, the boundary moves to swallow it, and whenever it falls, the boundary retreats.[^53]

Both were describing the edge between a company and the world. This chapter is about the edges inside a company, between its Outcomes, and it argues that they obey the same law and reach the same end. Every boundary between two Outcomes is a handoff. Every handoff costs. When the cost of a handoff is a leak, and a vector can hold both sides of it, the boundary moves, and it moves in one direction only: the downstream Outcome absorbs the upstream one. It keeps moving until there is one Outcome, and then the vectors inside it merge, as chapter 14 showed, until there is one vector. The company that began chapter 13 with ten departments and became five Outcomes does not stop at five. Everything tends to collapse into one, and this chapter follows the tendency to its end, while saying plainly where a real company may stop short of it.

## 2. The leak is the reason

Chapter 3 said where outcomes leak: between two desks, at the moment a result passes from the hands that made it to the hands that need it. A department was a workaround for the fact that no one person held every skill, and the handoff was the price of the workaround. Chapter 13 collapsed the departments into Outcomes, and the reader may have assumed that the leaks went with them. They did not. They moved. Where Pravaha once leaked between sales and the order desk, it now leaks between the winning Outcome and the fulfilling one: an order won on a promise the plant cannot keep, a delivery date set by a vector that does not see the queue in front of the machines, a credit term agreed by one chain and chased by another. The Keeper of chapter 8 settled exactly such a dispute, between Winning and Delivery, over a findability report. That dispute was a leak with two masters standing on either side of it.

A boundary between Outcomes is a cleaner leak than a boundary between departments, because the record shows it and the Keeper can rule on it in a day. But it is a leak, and the doctrine does not tolerate a leak it can close. It closes it the way it closes every leak: by putting both sides of the handoff inside one chain, where an agent carries the result across without a person on either bank.

## 3. The customer is the test

The reason tells you that boundaries will go. The test tells you which boundary goes next, and it is the test the whole book has used: what does the customer feel?

A customer of Pravaha does not feel an enquiry answered and then, separately, an order fulfilled. He feels one thing: he asked, he was promised, the pump arrived, he paid. Two Outcomes serving one feeling are a department by another name. The customer test says that wherever one customer’s experience crosses an Outcome boundary, the boundary is in the wrong place, and the Outcome nearest the customer’s last touch, the cash, should hold the whole of it. The test also says which boundaries hold for longest. The lender who funds the second assembly cell is a different customer from the plant owner who buys the pump, and feels a different result, so the capital Outcome has a reason to stand apart that the design Outcome never had. Boundaries between different customers’ results dissolve last. Boundaries inside one customer’s result dissolve first.

## 4. The vector is the mechanism

Reason and test say what should happen. The vector is how it happens, and it happens without anyone ordering it.

Chapter 14 showed a vector deepening: the same vector producing more each month as its agents compound, absorbing its neighbours because a vector that has learned one market learns the next faster than a fresh vector can. The same force works across an Outcome boundary. The deep winning vector’s pricing agent already reads the plant’s live capacity to set a delivery date; its quartermaster already knows which order is late and why. The day the fulfilling Outcome’s master looks at her floor and finds that the winning vector’s agents hold more of her chain’s state than her own vectors do is the day the boundary has already moved, and all that remains is for the Keeper to write it down. Nobody reorganises the company. The vectors grow through the boundary, and the boundary is found, afterwards, not to be there.

This is the difference between a merger and an absorption. A merger is a decision. An absorption is a reading. The Keeper does not decide that Winning and Delivery will become one Outcome; he reads, on a Monday, that one vector now produces the result both Outcomes were built for, and the Outcome that no longer produces anything of its own is, in the language of chapter 16, written on the wall.

## 5. The order at Pravaha

Pravaha ended chapter 13 with five Outcomes. Here is the order in which they became one, given by the test, with no dates, for a reason section 7 will give.

First, order-to-cash absorbed enquiry-to-order. The customer feels enquiry to cash as one result; the handoff between the two was the firm’s largest leak, a promise made by one chain and kept by another; and the deep winning vector of chapter 14 already held the fulfilment state. One Outcome, enquiry to cash, under one master. The other master became her backup, which is the standby rule of chapter 7 applied to a company that is running out of masters.

Second, the plant absorbed design. Design-to-first-unit had one customer, the plant, and produced one result for it, a first unit. When the plant’s own vectors, with their drawing and costing agents, produced the first unit of a new design faster than the design Outcome could hand one over, the design Outcome had no customer left. Written on the wall.

Third, enquiry-to-cash absorbed the plant. This is the one the reader will doubt, because the plant is where the hands are, and it is the one the customer test insists on most, because a pump on the customer’s floor is the promise kept, and lead time, which is the plant’s number, is the thing the customer feels most. By then the plant’s vectors held its machines, its cells and its robots as agents, and the chain from a quote to a unit in a crate was one chain with two masters standing in the middle of it for no reason the record could show.

Fourth, the Outcome that sold the plant’s spare capacity to other manufacturers merged into the one chain when the chain began to be lent, as chapter 14 describes, because a lent vector is the way spare capacity is sold.

Last, capital. Its customer, the lender, was a different customer until the record was complete enough that the lender read Pravaha’s floor directly and lent against what he saw. From that day a raise was not a result produced by an Outcome; it was a gate on the one chain, signed by the Keeper. The last boundary went where the first one had gone, into the record.

One Outcome. Inside it, over whatever time it took, the vectors merged as chapter 14 showed, each deep one absorbing its neighbour, until one vector carried enquiry to cash for every pump Pravaha made and every pump it made for the companies that borrowed its chain.

## 6. The six

What remains when one vector carries the whole company?

One aligner and one tuner, in the cockpit, holding the gates of the one chain. That much chapter 14 promised.

![Picture 13: The last six](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/13-the-last-six-1600.webp)

*Picture 13. The last six: five Outcomes collapsing into one, then one Outcome into one vector; the six who remain, in two rooms.*

A master, still. The reader expects her to go, since a floor of one tile is not a floor. She stays, because the floor was never the point of her. The master owns the playbook, the written design every agent in the chain is built from; she is the one court for every dispute an inspector raises; she holds the Monday test for the whole Outcome and reads the gap between the design rate and the actual one. A one-vector Outcome has a playbook, a court and a Monday test exactly as a hundred-vector Outcome does. What changes is that her control room streamlines to one tile holding thousands of agents, and the rules of chapter 7 all still run.

A Keeper, still, because there are gating events. An Outcome is born only by his decision. The outside, the board, the law and the lender, has one face. The treasury is his. And every irreversible act the playbook assigns above the master’s limit carries his signature. None of that is produced by the chain, so none of it disappears when the chain becomes one.

And two backups, one for the master and one for the Keeper. Chapter 7 gave the master a standby in another master, and chapter 8 gave the Keeper a standby on the board. When there is one Outcome there is no other master, and a board member who steps into a one-vector company at a day’s notice has not read the floor. So each seat has a backup of its own, selected on intention like every seat, and the two seats that stood beside the Keeper in chapter 8 dissolve into one of them: the Head of Outcomes has nothing to queue when there is one Outcome, and the Keeper’s outside is his own again once there is nobody between him and the lender.

Six people. Not four: the master and the Keeper are not the same craft, and a company whose Keeper also holds the playbook has nobody above the designer of the gates to hear an appeal against them, and chapter 17 needs one. Not eight: there is nothing for an eighth to hold. Six is the smallest core the doctrine knows how to design, and rule 12 of this chapter asks for a written reason for a seventh. It is a possible end, not an inevitable one. A company whose customers feel several different results, or whose hands no machine can yet replace, or whose law requires a person where the doctrine would put an agent, may hold at two or three Outcomes for many years, and the doctrine counts that as a company still collapsing, not a company that has failed. Where the customers feel results that are truly separate, a buyer of pumps and a lender to the plant, say, the boundary between them is not a leak and the right end is more than one Outcome.

## 7. Two rooms, shared

The backups do not wait. A backup who shadows is a person outside the structure, and the tile would count her; a backup who runs the edge is away from the room when the room needs her. So the backups work alongside.

The master and her backup share the control room. Both read the one tile every morning; both see the inspector’s holds, the agent maker’s proposals, the design rate against the actual; either can take the appeal, and the playbook says which of them signs which day. The Keeper and his backup share the command centre the same way: four columns, one Outcome, the doctrine tile, and the queue of gating events, read by two people who could each hold the table alone by lunch. Nothing is handed over when one is away, because nothing was ever held by one of them alone. That is the whole of the company’s insurance, and it costs two salaries.

## 8. Hands, and why there is no timeline

The reader who runs a plant has been waiting for the catch, and here it is. A company of six that makes pumps has a plant with nobody in it, and a plant with nobody in it is a plant whose hands are machines.

The doctrine’s answer is the same answer it gives for every skill: a hand is an agent. A humanoid, a drone, a wheeled robot, a loader, an arm at a bench, any non-human automation, is an agent inside the vector that needs it, built and upgraded by the vector’s agent maker, checked by the vector’s inspector, gated where its act is irreversible, and shown on the tile as one more colour. A plant vector may hold thousands of them. There is no robotics department, no automation head beside the master, no separate rule for a machine that moves. An agent with hands is an agent.

But an agent with hands is bought, installed, commissioned and trusted at the pace at which such agents exist, and that pace is not the doctrine’s to set. The collapse of the office, the chain from enquiry to invoice, needs agents that already exist and a decision to use them, and chapter 13 gave it a year. The collapse of the hands needs machines that can do what the hands did, in that plant, for that product, at that tolerance, and whether that takes a year or a decade depends on what is being made and what the company can afford to buy. A software company collapses to six faster than a pump maker. A pump maker collapses to six faster than a shipyard. So this chapter gives an order and no date. Everything tends towards one, in the order the customer test gives, at the pace the hands allow. A company that announces the date has mistaken the doctrine for a plan.

## 9. Rules for the collapse into one

1. Every boundary between two Outcomes is a handoff, and every handoff is a leak the doctrine works to close.
2. The downstream Outcome absorbs the upstream one. Absorption runs towards the cash.
3. The customer’s felt result sets the order: boundaries inside one result dissolve first; boundaries between different customers’ results dissolve last.
4. Absorption is a reading, not a decision. The Keeper records that one vector now produces both results; he does not order it.
5. The master of the absorbed Outcome becomes the backup of the absorbing one, while there are masters to spare.
6. Inside the one Outcome, vectors merge as chapter 14 describes, until one remains.
7. The smallest possible core is six: aligner, tuner, master, Keeper, and a backup for each of the last two. A company may stop short of it; none goes below it.
8. The master stays for the playbook, the court and the Monday test. The Keeper stays for the gating events.
9. Backups work alongside, sharing the room, never shadowing and never away.
10. Any non-human automation is an agent inside a vector, under the vector’s rules. A plant vector may hold thousands.
11. No timeline. The order is fixed; the pace is set by the hands.
12. Six is the floor, not a target date. A seventh needs a written reason.

## 10. The last merger

The last Outcome to dissolve at Pravaha was capital, and the day it went was a Tuesday, because the lender’s reading of the floor had been agreed on the Monday and the Keeper would not write a death into the record on the same day as the test.

The capital Outcome had one vector, one master and one pair. Its master had held the enquiry-to-cash floor for two years before the raise became her Outcome, and she had been the one who argued, at the Monday test, that the lender now read the floor as well as she did. The Keeper wrote one line in the record: that the capital-to-balance-sheet result was now produced by the one chain and gated at his signature, and that the Outcome was closed. No meeting was called, because there was nobody to call to it.

Its master became the backup to the master of the one chain, and walked that morning into the control room she had once held. The pair of its vector were paid out as chapter 13 pays leavers, and the agents that had kept the lender’s covenants and the bank’s reporting went where the work went, into the chain’s accounts agent, which already knew the covenants because it had been producing the numbers that satisfied them. By noon the control room held two people and one tile, and the command centre held two people, four columns and a doctrine tile that read zero, zero, zero.

Six. A company that had been a thousand people in ten departments, and then ninety-eight people in five Outcomes, and then one Outcome with about forty-five vectors, and then one vector lent to two other pump makers, was now six people and a chain, with Meera Kulkarni still at its gates, and it made more pumps than it had ever made, for more customers than it had ever had, and nothing in it waited on anyone.

The reader may find the number frightening. It is meant to be read the other way. Six is not what is left of a company after the doctrine has finished with it. Six is what a company can be, once everything that was a workaround has gone; how close a given company comes to it is set by its customers, its hands and its law.

[^52]: R. H. Coase, “The Nature of the Firm,” *Economica* 4, no. 16 (November 1937): 386–405.

[^53]: Oliver E. Williamson, *Markets and Hierarchies: Analysis and Antitrust Implications* (Free Press, 1975), and *The Economic Institutions of Capitalism* (Free Press, 1985).

================================================================
# Chapter 16: The Birth and Death of an Outcome
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-16-the-birth-and-death-of-an-outcome

## 1. One rule for both ends

The old company opened a division when a senior man wanted one and closed it when the board lost patience, and between those two events the division lived for years producing nothing anyone would have missed, because nobody had written down what it was for. The collapsed company has one rule for the birth and the death of an Outcome, and the rule is the test the book has used since chapter 5.

An Outcome exists exactly as long as a result has no other owner.

It is born the moment a result exists that nobody produces: a customer asks for something, or would pay for something, that no chain in the company is built to deliver, and no agent inside any vector can deliver as a step. It dies the moment that result has another owner: another Outcome’s vectors now produce it, as chapter 15 showed, or the customer who felt it is gone. Birth and death follow from the same question, which is the Keeper’s question in chapter 8: is this a result in its own right, and who produces it? When the answer is “a result, and nobody”, an Outcome is due. When the answer becomes “a result, and someone else”, the Outcome is done.

That is the whole law. Clayton Christensen spent a book showing that established firms fail at the new thing not because they cannot see it but because, inside the structure they have, it is nobody’s result;[^54] the rule above is the answer to him, applied every Monday.

The law has one exception, and it is a matter of capacity, not of result. When an Outcome’s floor grows past what one master can read, so that the reds reaching her week after week are more than her review can clear, the Keeper splits it in two, each whole, each with its own master and a copy of the playbook, as chapter 7 provides. Nothing new is produced and no result is unowned; the split exists so that no master is ever given a staff. The Keeper records a split as a new Outcome opened from the parent’s playbook, and the rest of this chapter applies to it. The rest of this chapter is procedure.

## 2. Who flags, who decides

Birth begins with a flag, and anyone with a name on a tile may raise it. The aligner who notices that a third of her enquiries ask for a service the company does not sell, the tuner whose agent maker keeps proposing a step the playbook has no place for, the master who sees the same unowned request on three vectors’ tiles: each may flag a result that nobody owns. The flag is a line in the record, not a presentation, and it says three things: what the result is, who would receive it, and why no existing chain can produce it as a step.

Only the Keeper decides a birth. That is not caution; it is the rule of chapter 8, that only he opens an Outcome, funds its first budget, names its first master and gives it its one result. A master cannot open an Outcome, because an Outcome is not inside hers. A pair cannot, because a pair holds a vector. The Keeper decides on the Monday test, in the record, within the week, and the answer is a birth or a reason.

Death needs no decision, because death is written on the wall. An Outcome whose result another Outcome now produces, or whose customer has gone, shows it on the Keeper’s table before anyone says so: its tile goes amber as its result thins, and the tile of the Outcome that has absorbed its work carries the result instead. The Keeper does not decide to close it. He records that it is closed, which is a different act, and section 7 shows it.

## 3. Born whole

An Outcome is born whole, or not at all.

This is the rule most likely to be broken by a well-meaning Keeper, so the chapter states it bluntly. A new Outcome has at least one master and at least one vector from its first day. It has a result of its own, a customer of its own, a playbook of its own and a Monday test of its own. It is not a vector that grew; it is not a trial that was promoted; it is not a project with a sponsor. A vector never becomes an Outcome, because a vector is one running copy of a discipline that already exists, and a new Outcome is a new discipline: a completely new kind of vector, built on a completely new thought about what the result is and how a chain produces it.

The distinction matters because the company already has a way of trying things, and it is not this. Chapter 8’s Keeper approved a trial vector for an overseas market, opened from the winning playbook, with a quarter to prove itself. That trial was a vector of the winning Outcome, aimed at a new market, and if it passed it became one more vector on that Outcome’s floor. Trials make vectors. They never make Outcomes, because a trial runs an existing playbook in a new place, and an Outcome needs a playbook that does not exist. When a flag arrives, the Keeper’s first question is therefore which of the two he is looking at: a known result in an unserved market, which is a trial, or an unowned result, which is a birth. The flag is often wrong about which it is, and the Keeper is the only person placed to tell.

## 4. Where the first master comes from

The first master of a new Outcome comes from inside the company, and the chapter rules out the alternative in one sentence: nobody outside the structure can hold a seat in it on the day it opens, because the seat is a craft learned on a floor, and an outsider has not stood on one.

So the first master is a backup master. Chapter 15 gave every master a backup who shares her control room and could hold it alone any morning; that is the person who has read a floor, settled an inspector’s appeal, kept a playbook and passed the four layers, and who is, by design, not indispensable where she sits. She takes the new Outcome. Her old seat takes a new backup, selected on intention, from the aligners, as chapter 14 describes.

The first vector’s pair are drawn the same way: an aligner or a tuner from another vector, who has held a chain and knows what a gate is, becomes the aligner of the new kind of vector, with a tuner selected to join her. The vector they leave takes a new pair through the four layers. Nothing in this is a promotion. A backup master who takes a birth is paid by the same formula on a result that begins at zero, and the doctrine expects her to earn less in the new Outcome’s first year than she did as a backup, which is one more reason the seat goes to the person who wants the craft.

## 5. What a death leaves

When an Outcome dies, two things happen to what it held, and the chapter will not soften either.

Its agents follow the work. The result the Outcome produced is now produced somewhere else, and the agents that produced it, the rubrics, the tools, the record of every decision they took, go into the vectors that now hold the result, where the agent maker fits them into a chain that already exists. Nothing an agent learned is lost, because an agent’s learning is in the playbook and the record, and both survive the Outcome.

Its people are offered any genuine vacancy they are suited to, and the rest are paid out. The pairs of a dead Outcome do not move to the absorbing Outcome, because the absorbing Outcome’s vectors already have their pairs and one human per seat is the doctrine. They do not move to the edge, because chapter 14’s rule about the edge is for pairs freed when vectors merge inside a living Outcome, where the company’s result is growing and new vectors are being opened; an Outcome’s death is the company’s result moving to fewer hands, not more. They are paid out as chapter 13 pays leavers, above what the law requires, on the day the record closes, and the record says why. The master of a dead Outcome becomes a backup if there is a master to back, as the master of Pravaha’s capital Outcome did in chapter 15, and is paid out if there is not.

A reader who finds this hard should notice that it is exactly what the old company did, after two years of pretending otherwise, to a division everyone knew was finished. The collapsed company does it on the day the wall says so, in writing, and pays better.

## 6. A birth and a death at Pravaha

The two happened in the same quarter, and the Keeper’s record shows them twelve lines apart.

**The birth.** The plant’s spare capacity had been a line in the fulfilment tiles for a year: machines cutting sixty-four per cent of the week, as chapter 13 left them, and the rest standing. The aligner of the OEM winning vector flagged it first, because three of her enquiries in a month had asked whether Pravaha would machine castings for other firms, and no chain in the company could say yes. Her flag said what the result was, machined components delivered to another manufacturer against its drawing; who received it, manufacturers who were not pump makers and would never be Pravaha’s customers for pumps; and why no existing chain could produce it as a step, because the fulfilment vectors’ agents were built to make Pravaha’s own units against Pravaha’s own order book and had no step for someone else’s drawing, someone else’s tolerance or someone else’s price.

The Keeper read it on the Monday. His first question was the one of section 3: known result, new market, or unowned result? A pump for a new customer is a trial. A machined part to another firm’s drawing is a different result, with a different customer who feels a different thing, and the chain that produces it would price, schedule and inspect against rules Pravaha did not have. A birth. He wrote the Outcome’s one result, spare-capacity-to-invoice, and its first budget, a quarter’s worth, in the record, and named its first master: the backup to the fulfilment master, who had shared the plant’s control room for eighteen months and had been the one, as it happened, who kept pointing at the idle thirty-six per cent. The OEM vector’s tuner became the new vector’s aligner, because she knew the castings and the machine shop; a tuner was selected through the four layers in the same week; the plant’s control room took a new backup from its aligners. The new Outcome had a master, a vector, a playbook written in the first fortnight from the fulfilment playbook with every rule about drawings, tolerances and pricing rewritten, and a Monday test of its own. It was born whole, and it invoiced its first casting in week seven.

**The death.** In the same quarter the design Outcome went. Design-to-first-unit had produced one result, the first unit of a new or changed design, for one customer, the plant. For two years the plant’s vectors had been drawing their own changes: their agent makers had absorbed the drawing and costing agents as a step inside the chain, because a change to a seal or a flange was a change to a unit the vector was already making, and the vector could draw it, cost it and cut it faster than it could hand the request across a boundary and wait. By the quarter in question, the design Outcome’s six vectors had produced one first unit in three months, and the plant’s vectors had produced eleven. The design tile had been amber for a year. The result had another owner. Nobody decided this; the wall said it, and the Keeper read the wall on a Monday and recorded it on the Tuesday.

The design Outcome’s agents went into the plant’s vectors, which already held most of them. Its master became the backup to the master of the new spare-capacity Outcome, which was a week old and needed one. Its twelve people, six pairs, were paid out on the Tuesday, each with a line in the record saying why, and two of them came through the four layers for the plant’s vectors inside the year, because selection for intention does not remember whose Outcome you were in.

## 7. The entry

The Keeper’s entry closing the design Outcome was one line, and it is reproduced here because the chapter’s argument is that it needs nothing more:

*Design-to-first-unit closed this day. Its result, a first unit of a new or changed design shipped against an order, is produced by the raw-material-to-finished-stock vectors, which have produced eleven to its one in the last quarter. Agents to those vectors. Master to backup, spare-capacity-to-invoice. Twelve seats paid out. No customer is affected.*

![Picture 14: Birth and death](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/14-birth-and-death-1600.webp)

*Picture 14. Birth and death: the Keeper’s table with one tile going dark and one lighting, and the two entries twelve lines apart in the record.*

There was no announcement, because the people affected had read the same tile for a year and knew what amber meant. There was no farewell, because an Outcome is not a family; it is a chain built to produce one result, and the result was still being produced, better, by another chain. There was no inquiry into whether it should have been closed sooner, because the record shows the Monday on which the wall first said so, and the Keeper’s speed on reds is his own colour in the board’s view of him. The old company would have held a meeting. The collapsed company wrote a line.

## 8. Rules for the birth and death of an Outcome

1. An Outcome exists exactly as long as a result has no other owner.
2. Any human with a name on a tile may flag an unowned result. The flag states the result, the receiver, and why no existing chain can produce it as a step.
3. Only the Keeper decides a birth, on the Monday test, in the record, within the week.
4. Trials make vectors. Births make Outcomes. A known result in a new market is a trial; an unowned result is a birth. The Keeper tells them apart.
5. An Outcome is born whole: at least one master and one vector, a result, a customer, a playbook and a Monday test of its own, from day one.
6. A vector never becomes an Outcome. A new Outcome is a new kind of vector, built on a new thought.
7. The first master is a backup master from inside. Nobody outside the structure opens an Outcome.
8. The first pair are drawn from another vector; the seats they leave are refilled through the four layers.
9. A new master is paid by the same formula on a result that starts at zero.
10. Death is written on the wall: another Outcome produces the result, or the customer is gone. The Keeper records it; he does not decide it.
11. A dead Outcome’s agents follow the work into the vectors that now hold the result.
12. A dead Outcome’s people are laid off and paid out as chapter 13 pays leavers, on the day the record closes, with the reason written.
13. The master of a dead Outcome becomes a backup if there is a seat to back, and is paid out if there is not.
14. The closing is one line in the record. No meeting, no inquiry, no ceremony.
15. One birth is not caused by a new result: when an Outcome’s floor outgrows what one master can read, the Keeper splits it in two, each born whole, each with a master.

## 9. The birth worksheet

One page, filled by the person who flags and signed by the Keeper.

1. The result: one line, stated as what the receiver gets.
2. The receiver: who feels it, and that they are not already a customer of another Outcome for the same thing.
3. Why no chain produces it: the step that does not exist, in the nearest existing playbook.
4. Trial or birth: known result in a new market, or unowned result. The Keeper’s answer.
5. The first master: which backup, and who backs the seat she leaves.
6. The first pair: which aligner or tuner, from which vector, and who refills it.
7. The first budget: one quarter.
8. The playbook: which existing one it starts from, and the rules that must be rewritten before the first gate.
9. The Monday test: the one question the Outcome answers each week from week one.
10. The first invoice: the week by which the Outcome is expected to produce its result once.
11. The Keeper’s signature and the date.

[^54]: Clayton M. Christensen, *The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail* (Harvard Business School Press, 1997). Christensen’s finding was that established firms fail at the new thing not for want of seeing it but because it is nobody’s result inside the existing structure, and his remedy was a separate organisation with its own customers and its own measures. The collapsed company reaches the same remedy by rule rather than by exception: an unowned result is an Outcome, born whole, with its own master, customer and Monday test, on the day it is flagged.

================================================================
# Chapter 17: Accountability, Authority and the Signature
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-17-accountability-authority-and-the-signature

## 1. Where authority went

In the old company, authority was a property of the seat. The regional manager could approve a discount because he was the regional manager; the head of finance could refuse it because she was the head of finance; and when the discount turned out to have lost the company money, a committee met to decide whose fault it was, and decided, usually, that it was the process. Accountability was a property of the organisation chart too, which is to say it was diffuse, and a thing that is diffuse enough belongs to nobody. Two economists gave this a name in 1976: the agency cost, the loss a firm suffers whenever the person deciding is not the person who bears the result of the decision, and they showed it was not an occasional failure but a structural one, built into every arrangement in which one person acts for another.[^55] Forty years later Nassim Taleb put the same point in four words and made it a moral rule: no skin, no game. A person who decides and does not bear the consequence will, over time, decide badly, and should not be allowed to decide at all.[^56]

The collapsed company takes both seriously and arrives at one rule. Authority is not a property of a seat. It is a property of a signature. Whoever signs a gate owns the result that follows, and nobody else does. That is the whole of accountability in a company built on vectors, and this chapter is about what the rule binds, who may sign, what happens when a signed result fails, and why nothing in it can be overruled.

## 2. The signature is the only authority

A gate, in chapter 6’s sense, is the point where one agent’s work is released to the next, and most gates are reversible and open on their own after the period the playbook sets. The gates this chapter is about are the other kind: the points in a chain where an act becomes irreversible, where money leaves, a name goes out, a promise is made. Every irreversible gate carries a signature, and the signature is the only act of authority in the company. There is no approval that is not a signature, no decision that does not sit at a gate, and no authority that exists between gates. Inside a chain, between one gate and the next, the agents act and the record keeps what they did; nobody is in charge of that stretch, because nothing in it can yet hurt anyone. Authority appears exactly where harm can, and nowhere else.

This is a narrower idea of authority than the old company had, and it is meant to be. A master in chapter 7 has three controls, question, pause, stop, and a budget; she does not have authority over the aligner’s gates, because they are not hers. A Keeper in chapter 8 signs anything inside the board’s budget and the doctrine, and nothing outside them. The aligner at her cockpit has more authority at her own gate than the Keeper has, because the Keeper’s name is not on her tile. The company has become a set of gates, each with one name on it, and the names do not stack.

## 3. Who may sign

Anyone whose name is on the tile.

Every vector’s tile carries names: the aligner’s, the tuner’s, and the name of every agent in the chain, the quartermaster, the inspector, the accounts agent, each line agent. The playbook assigns each gate to a name, and a name on the tile is a licence to sign at the gates the playbook gives it, no more. An inspector signs a hold. A quartermaster signs a release inside the chain. An accounts agent signs a payment inside its limit. The aligner signs the candidate going to the client, the quote going out, the offer. The master signs an appeal, a merge, a payment above the vector’s limit. The Keeper signs a birth, a change to the pay formula, anything the board’s budget allows and the doctrine permits.

The reader who has followed this far will ask which gates carry human names and which carry agents’, and the answer is that the playbook decides, vector by vector, on the record, and that the line moves. Today the irreversible gates carry human names, because the record that would earn an agent the right to sign them is still short. What happens as it lengthens is my forecast: I expect that by 2030 the automated system will be both more accurate and more trusted than the person at most gates, and a name moves from a human to an agent when the record shows the agent has never signed wrong at that gate. That is how the gates disappear, one name at a time. Nothing about the rule changes when they do. An agent that signs owns the result that follows exactly as a person does; its record is its accountability, and an agent whose signed results fail is rebuilt by the agent maker and its name taken off the tile, which is what happens to a person’s name too, by a longer route.

Three things follow. There is no signing on someone’s behalf: a backup master who signs does so in her own name, at a gate the playbook gives her that day. There is no delegated authority: a name that is not on the tile cannot be lent a signature by one that is. And there is no committee: a gate has one name, because a result with two owners has none. The one exception is chapter 9’s threshold of judgement: until a new aligner or tuner has passed it, her irreversible gates carry a second name, her pair’s or her master’s, and both names answer for the result.

## 4. What a signature binds

A signature binds the signer to the result that follows it, up to the next gate.

Not to the whole chain’s result; that belongs to the vector and is paid to the pair. Not only to the correctness of the one check; a check that was correct and a result that failed anyway is still the signer’s result to explain, though the review may place the cause elsewhere. The signer owns the stretch of the chain between her gate and the next one, whatever happens in it. The aligner who signs a candidate to the client owns that candidate’s progress until the next gate, the offer, whether the candidate fails the interview or the client changes its mind. The quartermaster that signs a release owns the batch until the inspector’s hold or the dispatch gate. The Keeper who signs a birth owns the new Outcome until its commissioning gate, the first Monday test, and sponsors it through its first quarter.

This is what gives the signature its weight. A signer who owns only the correctness of her check can sign quickly and often; a signer who owns the stretch that follows reads the stretch before she signs. The record makes the stretch visible: between any two gates, every act of every agent is written and replayable, so the signer can see what her signature will set in motion before she sets it. The old company asked its managers to approve things they could not see. The collapsed company shows the signer the whole stretch, and then holds her to it.

## 5. When a signed result fails

The record shows it. Pay follows the result. The signer explains her decision and leads the response, and a review proportionate to the failure separates her decision from the system’s.

There is no tribunal. The review starts from the record: the gate, the signature, the stretch, the point at which the result failed and the agent’s reasoning at that point are all written, in order, with times, and anyone with a name on the tile can replay them. The record makes the review fast; it does not make the answer obvious. The review separates the signer’s decision from the model, the design, the data, later events and causes outside the chain, and a lower share is not a finding of personal fault. There is no reprimand, because the formula of chapter 11 has already done what a reprimand pretends to do: the share moves with the vector’s result, the failed result is in this month’s number, and the signer’s pay beside it. There is no improvement plan, because chapter 11 forbade them; a signer whose signed results fail more often than the design rate allows shows it as a gap, the gap is attention, and the mirror, the rest and the replacement of chapter 11 follow in that order, read from the work.

The reader who has run a company will find this cold. It is colder than a committee and kinder than one. A committee decides whose fault it was by argument, after the fact, among people who were not there, and the person it blames is the one with the fewest friends in the room. The record decides nothing; it shows what happened, the formula pays for it, and the signer knows on the first of the month exactly what her signature cost her, which is the only lesson a failed signature has to teach.

## 6. A signature is never overruled

A master cannot overrule an aligner’s signature. A Keeper cannot overrule a master’s. Nobody can overrule anybody’s, because a signature that can be overruled is not a signature; it is a recommendation, and a recommendation has no owner.

What a master or a Keeper can do is sign at a gate of their own. Chapter 7 showed it: an inspector holds a candidate a tuner signed out; the tuner appeals; the master decides. The inspector’s hold stands in the record as the inspector’s signature, and the tuner’s release stands as hers; the master’s decision is a third signature at a third gate, the appeal, which is the master’s gate and nobody else’s. Nothing is erased. The record shows three names, three gates, three results, and each signer owns her own stretch. If the master’s decision turns out wrong, it is the master’s result, in the master’s number, and the inspector’s hold is still right in the record.

The same is true upward. A Keeper who settles a dispute between two masters, as chapter 8’s did, signs at the dispute gate, within twenty-four hours, on the record, and the cost is written into the losing Outcome’s result, which is to say into that master’s number. The master’s original signature is not overruled; it is answered by a further signature at a further gate. Authority in the collapsed company does not flow down through a chain of command that can reverse what is below it. It sits at gates, one name each, and the only way to change a result is to add a gate and a name.

## 7. What no tile covers

Some things arrive that no gate was built for: a kind of client the playbook never saw, a legal notice, a machine that fails in a way the inspector’s rubric has no line for. No name on any tile covers them, so nobody may sign, and a thing nobody may sign waits.

The rule is that it rises, and that each rise adds a gate on the record. The aligner whose tile does not cover the thing sends it up to the master, with the record of why no gate fits. The master either signs at a gate of her own, if the thing is inside her Outcome and her playbook and her budget, and writes the new gate into the playbook so that the next one has a name; or she sends it up to the Keeper, with her own line added. The Keeper signs at his, inside the board’s budget and the doctrine, or takes it to the board, which answers within seven days. At every step the thing is written, the step is written, and the name that finally signs owns the result. Nothing is decided by a phone call, because a phone call is not a gate. And every unmatched thing leaves a new gate behind it, so that the playbook grows by exactly the number of things it failed to foresee and no more.

## 8. One order, every signature

Here is one order at Pravaha, in the year after the collapse of chapter 13, followed from enquiry to cash with every signature on it, and the one that failed.

The enquiry arrived at 9:04 on a Tuesday, from a process-industry client in the west, for six pumps against a specification. The vector’s conversation agent answered it at 9:04, and the record shows no signature, because answering an enquiry is reversible.

![Picture 15: The signature](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/15-the-signature-one-order-1600.webp)

*Picture 15. The signature: one order drawn as a chain with eight seals on it, the fifth, the appeal release, red.*

**Signature one.** At 11:40 the pricing agent priced the six pumps against the plant’s live capacity and the client’s credit record, and the quote was ready. The quote is a promise, so it is a gate, and the gate carries the aligner’s name. She read the stretch: the capacity the pricing agent had reserved, the delivery week it implied, the margin. She signed at 11:52. From that signature until the order gate, the quote was hers.

**Signature two.** The client accepted on the Thursday. Acceptance creates an order, and an order commits the plant, so it is a gate, and the gate carries the quartermaster’s name, because an order within the quoted terms commits nothing the aligner did not already sign for at the quote; the irreversible act was her signature on the quote. The quartermaster signed at 15:10 on Thursday and raised the order to the plant vector. From that signature to the first article, the order was the quartermaster’s.

**Signature three.** The plant vector’s first article, the first of the six pumps, came off the test rig on the second Tuesday. First article is a gate, and it carries the plant aligner’s name. She read the inspector’s measurements and signed.

**Signature four.** Batch release, the sixth pump, the following Monday. The inspector’s name: it signs that every unit is inside the drawing. It signed five. On the sixth, a seal face measured at the edge of tolerance, and the inspector signed a hold instead. The hold is the inspector’s signature and its result, and the record shows it.

**Signature five.** The plant aligner appealed the hold to the master, because the client’s delivery week was that week and the seal was inside tolerance by the drawing, if only just. The master read the stretch, the measurement, the drawing, the rubric’s margin and the client’s history, and signed at the appeal gate: release. Her signature, her gate, her result.

**Signature six.** Dispatch, the Tuesday. The dispatch agent’s name, inside the chain; it signed the crate out and the invoice agent raised the invoice the same hour. The invoice is money, so it is a gate, and it carries the aligner’s name: signature seven.

**Signature eight.** Cash, thirty-one days later, one day past terms. The collections agent had called on the first day past due, as its rubric says, and signed the receipt into the record. The order was closed.

The one that failed was signature five. The sixth pump’s seal wept at the client’s site in the ninth week. The client’s claim came back through the conversation agent, which is where everything comes back, and the record laid out the stretch: the inspector’s hold, the appeal, the master’s signature, the release. Custody of the pump had passed down the chain, but a later signature does not erase an earlier decision. The failure was not the inspector’s, whose hold was right and stands in the record as right. It was not the aligner’s, who signed the quote and the invoice and owned neither the release nor the pump. It was the master’s, whose gate the release had passed, and whose stretch, from appeal to dispatch, held the failure. The pump was replaced under warranty in the tenth week, at the vector’s cost, and the cost went into the vector’s result and, by the rule of chapter 8’s disputes, into the master’s number for the month, because the signature was hers.

## 9. The entry

The master’s entry in the record, on the day the claim was replayed, was one line:

*Appeal release signed by me on the sixth unit, order 2214, against the inspector’s hold; the unit failed at site in week nine; replaced at the vector’s cost. The hold was right. Rubric margin on seal faces raised to the inspector’s proposal from the same day.*

Beside it, the following month, the record shows her pay: the base unchanged, the share down by the cost of one pump against the Outcome’s result, which was a little under two per cent of her month. Nobody wrote to her about it. Nobody held a meeting. The inspector’s rubric had already been changed by the agent maker from the proposal the inspector had attached to its hold, which the master had accepted in the same entry. The next sixth pump with a seal face at the edge of tolerance did not leave the plant.

That is the whole of accountability in a collapsed company, in two lines of a record and one number. The person who signed owned the result; the record showed it; pay followed; the gate improved; nobody else was blamed, because nobody else had signed.

## 10. Rules for the signature

1. Authority is a signature at a gate. There is no other authority in the company.
2. A gate is a point where the chain waits for a signature. Irreversible gates (money, a name, a promise) default closed; reversible gates default open after the period the playbook sets.
3. Anyone whose name is on the tile may sign, agent or human, at the gates the playbook gives that name.
4. A gate has one decision owner. A new holder below the threshold of judgement is co-signed (chapter 9), and a gate the playbook marks as higher-risk may require an independent second approval; the decision still has one owner. There is no committee, no delegation, no signing on another’s behalf.
5. A name moves from a human to an agent when the record shows the agent has never signed wrong at that gate. That is how gates disappear.
6. A signature binds the signer to the result that follows, up to the next gate.
7. Read the stretch before signing. The record shows it.
8. When a signed result fails, the record shows it and pay follows the result. No inquiry, no reprimand, no improvement plan.
9. A signature is never overruled. A master or Keeper answers it with a signature at a gate of their own, and the record keeps both.
10. A hold is a signature. An appeal is a gate. The master who decides an appeal owns the decision, not the hold.
11. What no tile covers rises, with the record of why, and each rise adds a gate. A phone call is not a gate.
12. Every unmatched thing leaves a new gate in the playbook behind it.
13. The firm is accountable to the outside through one face, the Keeper’s. Inside, the signer is accountable, and the designer of a gate is responsible for its design, which is why the playbook carries the master’s signature too.
14. A failed signature teaches once, in the record, by the number beside it.

## 11. The signature worksheet

One page per gate, kept in the playbook.

1. The gate: what act becomes irreversible here.
2. The name: who signs, human or agent, and when the name last changed.
3. The stretch: what the signer owns, from this gate to which next gate.
4. What the signer must read before signing: the record lines the playbook requires.
5. The hold: which agent may hold here, and on what rule.
6. The appeal: whose gate the appeal is.
7. The cost of a failed result at this gate: where it is written, and whose number it enters.
8. What rises from here when no name fits: to whom, with what record.
9. The record of this gate’s failures, by signer, by month.
10. The designer’s signature: the master’s name on the gate’s design, and the date.

[^55]: Michael C. Jensen and William H. Meckling, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure,” *Journal of Financial Economics* 3, no. 4 (1976): 305–360.

[^56]: Nassim Nicholas Taleb, *Skin in the Game: Hidden Asymmetries in Daily Life* (Random House, 2018).

================================================================
# Chapter 18: The Edge of the Vector
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-18-the-edge-of-the-vector

## 1. One order, three edges

At 8:50 on a Monday an order for forty pumps crosses into Pravaha’s one chain, and the record marks three places on it where the chain stops being Pravaha’s.

The order comes from a compressor maker, the OEM client the deep winning vector took on in chapter 14, which buys pumps as components and still runs on departments. Its purchasing department works, for everything that concerns Pravaha’s pumps, inside Pravaha’s vector: the buyer sees her drawings, her schedule and her approvals in a cockpit that is Pravaha’s, and the vector’s agents hold her side of the order as if it were theirs. That is the vector lent at its edge, the first boundary, and the gate at which the buyer’s department head and Pravaha’s aligner both sign is the edge of the vector.

Beyond that gate is the second boundary: the client’s own company. The forty pumps need an engineering release the client’s engineering department gives, and a payment the client’s finance department makes, and neither department is inside the vector or ever will be unless the client borrows more of the chain. The record shows the order waiting there, and the chapter will show for how long. On the other side of the chain, the castings for the forty pumps come from a foundry whose order book Pravaha’s quartermaster writes into directly, and whose dispatch department ships on Thursdays, by department, as it has for thirty years.

And at the end of the order, past the cash, the third boundary: the pumps carry a conformity certificate that the law requires a named, qualified person to sign, and the client’s bank, which is financing the forty compressors the pumps go into, wants a director’s signature on a document that the record could have produced in a second. That is where the doctrine stops today, and the chapter says so first, because a book that showed only where its doctrine runs would be selling something.

Three boundaries. The vector’s edge, which has a rule. The uncollapsed world, which has a measure. The doctrine’s limit, which has a rank, by how soon each part of it will fall.

## 2. Why a vector has an edge at all

Coase’s question from chapter 15 applies here with the sign reversed. He asked why a firm has an inside; the question at the edge of a vector is why the inside stops. The answer is his answer: the inside stops where coordinating by instruction costs more than coordinating by contract.[^57] For two centuries that line sat at the company’s wall. The collapsed company moves it, because a vector that holds a result from a customer’s first word to cash will, as chapter 14 showed, be lent across the wall when it has nothing left to deepen inside it, and the line between instruction and contract then runs through another company’s building.

The line can be moved two distances. A vector lent whole, as the export-market pump maker borrowed Pravaha’s chain, runs the other company’s result from end to end, with the other company’s pair at the gates once they are selected, and the edge of the vector is the other company’s edge. A vector lent at its edge, as the OEM’s purchasing borrowed it, runs only the stretch of the other company’s work that touches the vector’s result, and the edge sits inside the other company, at one gate. The second is the common case and the harder one, because on one side of the gate is a chain that never waits and on the other is a department that does nothing else.

Williamson, writing about why two firms that depend on each other so often end up as one, found the cause in what he called asset specificity: once each side has built something that is worth more inside the relationship than outside it, each can hold the other to ransom, and the cheapest cure is a boundary both sides can see and neither can move alone.[^58] That is the rule at the vector’s edge, and the rest of the chapter gives it its form.

## 3. The edge: a double signature

At the gate where Pravaha’s vector ends and the client’s department begins, two names sign: Pravaha’s aligner and the client’s department head. Neither signature alone opens the gate. The order does not pass into the chain until both are on the record, and nothing passes back out, a changed drawing, a revised date, a price, until both are on it again.

The reader who has followed chapter 17 will object that a gate has one name, and the objection is right everywhere but here and below the threshold of judgement. A gate inside a company has one name because the result that follows it has one owner. A gate between two companies has two owners, because the result on each side of it is felt by a different customer, Pravaha’s and the client’s, and neither company’s record governs the other’s. A double signature is not a committee. It is two single signatures, each binding its signer to the result on her own side, written on one line because the line is where the two results touch. Pravaha’s aligner owns the stretch from the gate to the next gate in her chain; the client’s department head owns whatever the client’s company does with the order after it leaves the vector. If the forty pumps are late, the record shows on which side of the line the days were lost, and whose signature the lost days sit under.

The double signature is also why the vector can be lent at all. A department head who signs beside an aligner has given nothing up: her company’s authority still passes through her name. What she has gained is that everything up to her signature now arrives from a chain that is faster than any department she has ever worked beside, and the record shows her exactly what she is signing. Chapter 17’s rule holds for her as for anyone: she should read the stretch, and the cockpit shows it to her. Most department heads, the first time they sign beside an aligner, read the stretch and ask why the order took nine days inside their own building to reach the gate. That question is the beginning of the second boundary moving.

## 4. The uncollapsed world, measured

The collapse of Pravaha did not collapse its customers, its suppliers or its lender. It moved the leak.

Chapter 3 said where outcomes leak: between two desks. Inside the one chain there are no desks, and inside it an order does not wait. But the order came from a company with desks and goes back to one, and the castings came from a third. The leaks that used to sit between Pravaha’s departments now sit in three places outside it: inside the client’s departments, between their handovers, where engineering’s release waits for purchasing’s attention and purchasing’s release waits for finance’s; at the boundary itself, in the wait for the second signature, the days between the aligner’s name going on the line and the department head’s; and in the supplier, who makes castings in a week and ships them on the Thursday that a dispatch department works to.

The doctrine’s answer to a leak it cannot close is to measure it, in the same record and in the same unit it uses for everything: days lost per order, by place. Eliyahu Goldratt made the point about plants that the chapter now makes about chains: a system’s throughput is set by its constraint, and a constraint that has moved outside your walls is still your constraint, and still the first thing to find.[^59] The vector’s quartermaster finds it. For every order the record shows the days the order spent inside the chain and the days it spent outside it, and where outside, and the three outside numbers appear in the tile as a colour of their own: the edge colour. A vector whose own stretch runs in a day and whose orders take a month is not a vector with a problem. It is a vector with its constraint on the far side of a gate, and the record names the gate.

What the vector does with the constraint it cannot own is what Goldratt’s plants did: exploit it and subordinate to it. It books the client’s engineering release before the castings are poured, because the release is the slower of the two. It raises the foundry’s purchase order on a Monday so that the castings make the Thursday truck. It puts the second signature in front of the department head at the hour the record shows she signs, and nothing else in front of her. None of this removes a day from the client’s departments. All of it stops the vector adding days of its own to theirs, and the record of what remains is the argument for the next section.

## 5. What the vector does with the client’s leak

A vector that can see, in its own record, that a client’s finance department loses nine days an order between three of its desks is looking at a result nobody owns: the client’s purchase-to-pay cycle, run without the leak. The temptation is to do something about it, and the doctrine forbids every form of doing something except one.

The vector does not offer to collapse the client’s department. A vector serves one result and holds one chain; a client’s purchasing is not its result and never becomes it, because chapter 16 settled that a vector never becomes an Outcome, and a vector that starts selling the client a reorganisation has become a consultancy with a cockpit. The vector does not show the client the record and let the numbers argue, either, however tempting the numbers are; the client’s department head reads her own side of the record already, and what she does with it in her own company is hers.

What the vector does, every time, is report the leak to Pravaha’s Keeper, as a flag of the kind chapter 16 describes: a result that exists, the client’s purchase-to-pay cycle without the nine days; a receiver, the client’s owner; and why no chain produces it, because no Outcome at Pravaha is built to run a stretch of another company’s work as its own result. Every time, because the Keeper decides births on patterns and the record only shows a pattern if every instance is in it. The Keeper reads the flags on a Monday, and for a year he reads them as what they are, a known leak in companies Pravaha does not own, and declines, which is the correct reading of a single flag. In the quarter the record shows the same nine days at six clients, he reads them differently: an unowned result with six receivers who would pay for it. He opens an Outcome, born whole as chapter 16 requires, with a backup master and a new kind of vector whose result is a client’s purchase-to-pay run on Pravaha’s chain, lent at the edge, with a double signature at the client’s wall and a share of the days recovered as its price. The order this chapter follows crossed the client’s wall before that Outcome existed; its record became part of the evidence that opened it.

That is the only legitimate path from a leak outside the company to a result inside it: the flag, the Keeper, the birth. A vector that takes any shorter path has grown a department beside itself, and the tile would count it.

## 6. Where the doctrine stops

Three things the doctrine cannot yet reach, named in the order in which they will stop being true.

**The client’s departments.** This is the nearest limit and the first to fall, because every lending moves it. The doctrine cannot reach into a company that has not collapsed; it can only run the stretch that company lends it, behind a double signature, and measure the rest. But each stretch lent is a department the client no longer needs on that stretch, and a department head who has signed beside an aligner for a year has seen the record. The client’s departments fall one lent stretch at a time, at the pace the client’s owner can bear, and the doctrine’s answer to them is patience and a column of days.

**The hands.** Chapter 15 gave the rule, that an agent with hands is an agent, and the reason there is no date, that the hands fall at the pace at which such agents exist for that plant and that product. The doctrine cannot reach a bench that no machine can yet stand at, and in a pump maker there are still a few. This limit falls at the speed of the machines, which is faster every year and still not the doctrine’s to set.

**The statutory signature.** This is the farthest limit and the last to fall. The law requires that certain acts carry the name of a natural person with a qualification the state recognises: a conformity certificate, a director’s signature on the accounts, a customs declaration, a licence. The doctrine stops at every one of them, exactly, and does not pretend otherwise: at Pravaha the statutory gates carry the name the law requires: the Keeper’s for what a director must sign, and an outside qualified and independent professional, paid for by the company, where the law requires independence, as it does for the auditor, and the chain produces the document, stops, and waits for the signature the law wants. Chapter 17 forecast that by 2030 the automated system will be more accurate and more trusted than the person at most gates. The statutory gate will be the last to move, because it moves by the consent of legislators who have not read this book, and a doctrine that waited for them would never have started. So the doctrine runs up to the statute, signs there in the one name the law will accept, and keeps the record ready for the day the law reads it.

That is the whole of the limit. Everything short of it the doctrine reaches today, and the rank is the order in which the rest will go: the client’s departments, then the hands, then the statute.

## 7. The order, both sides

Order 3127 is the forty pumps of section 1, followed from the OEM’s purchasing desk to Pravaha’s cash, with every day accounted for.

**Day 0, Monday, 8:50.** The OEM’s buyer, working inside Pravaha’s vector, raises the order against the quote the aligner signed the previous week. The vector’s quartermaster prices the castings and the delivery week from the plant’s live capacity. **Day 0, 9:10.** The aligner signs the boundary gate. **Day 2, 16:40.** The OEM’s purchasing head signs beside her. Two and a half days at the boundary, in the wait for the second signature: the record shows the purchasing head signs on Wednesday afternoons, and the vector learns to put Monday orders in front of her by the Wednesday.

![Picture 16: The edge of the vector](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/16-the-edge-of-the-vector-1600.webp)

*Picture 16. The edge of the vector: order 3127 across three companies, two seals at each wall, and the days lost written along the stretches outside the vector.*

**Day 2, 16:41.** The quartermaster raises the casting order into the foundry’s order book, which the vector writes into directly under a boundary gate of its own: the aligner’s name and the foundry’s dispatch head’s. **Day 3, morning.** The foundry’s dispatch head signs; half a day at that gate. **Day 3 to day 7.** Castings poured and inspected, four days of the foundry’s own work. **Day 10, Thursday.** Shipped on the Thursday truck, because the foundry ships on Thursdays. Three days lost in the supplier’s dispatch, measured from the day the castings were ready, and one day on the road.

**Day 11 to day 19.** Pravaha’s stretch: castings in on Friday, forty pumps machined, assembled and tested by the Saturday of the following week, the inspector’s signatures on every one, the plant aligner’s on the first article and the batch. Eight days inside the chain, against a design rate of eight. No days lost.

**Day 19.** The pumps are ready, and the order now waits for the OEM’s engineering release, which the OEM’s engineering department issues after reviewing the test certificates, which it does in a weekly meeting on Wednesdays. **Day 23.** Release. Four days lost inside the client’s engineering department, between the certificates arriving and the meeting that reads them. **Day 23 to 24.** Dispatch, and the conformity certificate signed in the Keeper’s name at the statutory gate, which the record shows took fifty minutes, because the Keeper’s backup was in the room.

**Day 24.** The invoice goes out under the aligner’s signature. Terms are thirty days. **Day 54, then day 63.** The OEM’s finance department pays on day 63, nine days past terms, because an invoice at the OEM passes from purchasing to accounts payable to a signatory, and each handover waits for the next desk’s turn. Nine days lost inside the client’s finance department. The collections agent called on day 55, as its rubric says, and the record shows the call, and the nine days still lost, because a call cannot make a department move faster than its desks.

Sixty-three days from order to cash. Pravaha’s own stretch, eight, and one day for its dispatch and the certificate. The boundary, two and a half days at one gate and half a day at the other. The supplier, four days of work and three of waiting for a truck, and a day on the road. The client’s departments, four in engineering and nine in finance. Terms, thirty. Nineteen of the sixty-three days were lost outside the vector, waiting on a desk, and the record says exactly where, and under whose signature each stretch sat.

That record went to the Keeper as a flag on day 63, the eleventh flag of its kind that quarter. It was one of the six that opened the Outcome of section 5.

## 8. Rules for the edge of the vector

1. A vector’s edge is wherever its result stops being the next customer’s result. It may sit at the company’s wall or inside another company’s.
2. A vector may be lent whole, running another company’s result end to end, or lent at its edge, running one stretch of another company’s work. The same rules govern both.
3. The gate at the edge carries two names: the vector’s aligner and the other company’s department head. Two single signatures on one line, each binding its signer to the result on her own side.
4. Nothing passes the edge, in either direction, without both names.
5. The leak has not gone; it has moved outside. The record measures it as days lost per order, in three places: inside the other company’s departments, at the edge in the wait for the second signature, and in suppliers who ship by department.
6. The edge colour is a colour of the tile. A vector whose own stretch is at design rate and whose orders are slow has its constraint across a gate, and the record names the gate.
7. Exploit and subordinate: book the slower release first, raise the order to make the supplier’s truck, put the second signature in front of the signer at the hour she signs. Never add a day of the vector’s own to the days outside.
8. A vector never offers to collapse the other company’s department, and never argues the record at it. It serves the stretch it was lent.
9. Every time the vector sees a leak in the other company, it flags it to the Keeper, as an unowned result with a receiver. Every time, because births are read from patterns.
10. Only the Keeper decides whether the pattern is a birth. If it is, the Outcome is born whole, with a master and a new kind of vector, as chapter 16 requires.
11. The doctrine stops at three things today: the other company’s departments, which fall one lent stretch at a time; the hands, which fall at the pace of the machines; and the statutory signature, which falls last and by consent of the law.
12. At a statutory gate the chain produces the document, stops, and waits for the one name the law accepts. It does not pretend the gate is not there.
13. The company says where its doctrine stops, in the record, as plainly as it says where it runs.

## 9. The record of days lost

Order 3127, forty pumps, OEM client, as the record keeps it.

| Where the days went | Days | Under whose signature |
|---|---|---|
| Boundary gate, client side: wait for the second signature | 2.5, lost | Client’s purchasing head |
| Boundary gate, supplier side: wait for the second signature | 0.5, lost | Foundry’s dispatch head |
| Foundry: castings poured and inspected | 4, work | Foundry’s dispatch head |
| Foundry’s dispatch department: castings ready to castings shipped | 3, lost | Foundry’s dispatch head |
| Road, foundry to Pravaha | 1 | — |
| Pravaha’s chain: castings in to pumps tested | 8, design rate 8 | Plant aligner; inspector |
| Client’s engineering department: certificates received to release | 4, lost | Client’s engineering head |
| Pravaha’s dispatch, and the conformity certificate at the statutory gate (50 minutes) | 1 | Aligner; Keeper |
| Terms | 30 | — |
| Client’s finance department: invoice received to payment, past terms | 9, lost | Client’s finance head |
| **Order to cash** | **63** | |
| **Of which lost outside the vector, waiting on a desk** | **19** | |

Across the quarter, for this client, the record showed the same four places on every order, with the same signatures under them, and the nine days in finance on all but one. That is a pattern, and a pattern is what the Keeper reads.

## 10. The boundary worksheet

One page per edge, kept in the playbook of the vector that is lent.

1. The other company, and whether the vector is lent whole or at its edge.
2. The stretch lent: from which of the other company’s desks to which.
3. The gate at the edge: what act it covers, and the two names on it.
4. The other company’s signer: who she is, and the hour the record shows she signs.
5. What passes the gate in each direction, and what never does.
6. The other company’s departments on either side of the stretch, named, with the handovers the record will measure.
7. The supplier edges the order crosses, each with its gate and its two names.
8. The edge colour: the three places days are lost, and the threshold at which the tile turns amber.
9. Exploit and subordinate: the three things the vector does to add no days of its own.
10. The flag: the unowned result the leak implies, its receiver, and the Keeper’s last reading of it.
11. The statutory gates the order crosses, each with the one name the law accepts.
12. Where the doctrine stops on this edge, in one line.

[^57]: R. H. Coase, “The Nature of the Firm,” *Economica* 4, no. 16 (November 1937): 386–405, as cited in chapter 15.

[^58]: Oliver E. Williamson, *The Economic Institutions of Capitalism* (Free Press, 1985), on asset specificity and the governance of relationships in which each side has built something worth more inside the relationship than outside it.

[^59]: Eliyahu M. Goldratt and Jeff Cox, *The Goal: A Process of Ongoing Improvement* (North River Press, 1984). Goldratt’s five steps, identify the constraint, exploit it, subordinate everything else to it, elevate it, and begin again, are the steps the vector takes at an edge it cannot own, with the last step replaced by the flag to the Keeper.

================================================================
# Chapter 19: The Long Road to the Vector
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-19-the-long-road-to-the-vector

## 1. The road

This book has argued that the department was a workaround for a constraint, that the constraint has gone, and that a company can collapse to a chain of agents held by one pair of humans by default at the gates. A reader who has come this far is entitled to ask whether anyone has ever done anything like it, and the honest answer is that everyone has, for two thousand years, a step at a time, and stopped.

The road to the vector is long and it is well marked. At each milestone someone took a skill that lived in a person and moved it into something that was not a person: a drill, a yard, a sequence, a line, a machine that stopped itself, an interface, a model. Each time, the result needed fewer people than before, often spectacularly fewer, and each time the people who were left were reorganised around the new machine in exactly the old shape, with layers above them and handoffs between them, and the collapse stalled a step short of the result. Eight of those milestones are told here, properly, as the stories they are. They are not chosen because they failed. They are chosen because each one succeeded at the thing it attempted and then stopped, and the place where each stopped is a lesson the doctrine has already absorbed.

![Picture 17: The long road](https://vinaypasricha.com/assets/images/collapse-or-collapse/pictures/17-the-long-road-1600.webp)

*Picture 17. The long road: eight stations, each labelled with what moved out of people, ending at the vector.*

The thread through all eight is one sentence. Every station moved skill out of people and into a machine or a drill, so every station needed fewer people for the same result. That is also the sentence this book is built on, and the vector is where the sentence finally runs to its end: a chain that holds every skill a result needs, and one pair of humans at its gates. The eight stations show the road; the next chapter hands the reader the keys and a month to drive it.

## 2. The legion

Late on an afternoon in Gaul, in the first century before Christ, five thousand men stop walking.

They have covered twenty miles since dawn with their armour on their backs, and before they eat they will build a town. Surveyors have gone ahead and planted a white flag where the commander’s tent will stand; from it, with measuring rods and coloured flags, they have laid out the two main streets, the lines of each cohort’s tents, the gaps between them, and the rectangle of the rampart, always the same, always in the same proportion to the same places. Each man arrives and knows where his tent goes without being told, because it went there last night and the night before and will go there tomorrow; Polybius, who watched them do it two centuries earlier, said the whole army took its place like a city’s people coming home, each to his own street and his own house.[^60] Some dig the ditch, some cut the turf for the rampart, some set the palisade stakes they have carried all day for this purpose, and in the time it takes the sun to go down the five thousand are behind a ditch and a wall, in streets, with a gate on each side and a guard on each gate.

Nobody in that camp is a hero. That is the point of it. The Gauls on the other side of the valley have men who have fought since boyhood and can do things with a long sword that no legionary can do with a short one. The legion does not need them to. Its men are farmers’ sons with two years of drill, and the drill has taken the skill of fighting out of each of them and put it into the unit: eighty men to a century under a centurion who has done this for twenty years, six centuries to a cohort, ten cohorts to a legion, eight men to a tent who march, dig, cook and fight as one thing. Josephus, a Jewish general who fought them and lost and then lived among them, wrote that their drills were bloodless battles and their battles bloody drills, and that there was no disorder, no rout, no surprise that could reach them, because nothing they met in battle was anything they had not already done a hundred times in the yard.[^61]

The legion is the first station because it is the first time a civilisation deliberately took the most prized skill it had, the skill of the warrior, and decided it would rather have it in a drill than in a man. The drilled unit beat the skilled individual every time it met him, for four hundred years, with ordinary men.

**What collapsed.** The individual skill of the fighter, into the drill of the unit; the knowledge of how to camp, march, build and fight, into a routine every man carried and no man owned.

**What stayed.** Every skill still lived in a body, and had to be put into each new body by two years of the same drill. Above the eight men in a tent stood the century, above the century the cohort, above the cohort the legion, above the legion the army and the Roman state, a pyramid of rank as steep as any that came after it, with the drill as its base.

**The number.** Five thousand men who built a fortified town every evening of a campaign, with a gate on each side, before they ate; and a legion of farmers that was expected to beat a host of warriors twice its size, and usually did.

**Why it stalled.** Because the drill lived in men, it died with them. Four centuries later a late Roman writer named Vegetius compiled a book pleading for the restoration of a drill the army had let lapse, gathering it from older writers and imperial regulations, because the drill itself was no longer practised.[^62] A drill that must be re-taught to every recruit is a machine that is rebuilt from scratch every generation. And a legion’s result scaled only one way: more result meant more legions, each of five thousand men, until the empire could not pay for them. The skill had left the man, but the man was still the unit of output.

## 3. The Arsenal

In the spring of 1438 a Castilian traveller named Pero Tafur stood in Venice at the gate of the Arsenal, the Republic’s shipyard, and watched ten galleys come out armed between the hours of three and nine.

He wrote down what he saw, which is why we have it. Inside the walls was a long canal, and on each side of the canal stood the warehouses. A galley hull was towed along the water, and as it passed each warehouse the men inside handed out onto the deck whatever that warehouse held: from one, the cordage; from another, the bread; from a third, the arms; from a fourth, the ballistas and mortars; and so down the line, so that by the time the hull reached the end of the canal it carried its oars, its stores, its weapons and its men, and went out of the gate complete. In this manner, he wrote, ten galleys fully armed came out between the hours of three and nine.[^63]

Tafur had seen an assembly line, five hundred years before the phrase, and he had seen something the legion did not have: a machine that was not a man. The Arsenal was a yard, a canal and a set of warehouses laid out so that the work came to the parts rather than the parts to the work, and it ran on standard pieces. Venetian war galleys were built to a pattern so that a mast, a rudder, an oar or a bench from one would fit another, which meant the Republic could keep finished hulls under cover in reserve and fit them out in days. In the summer of 1570, with the Turks landing on Cyprus, the Arsenal fitted out and sent to sea some hundred galleys in about two months, a feat no other state in Europe could have attempted.[^64] The Arsenal in those years employed perhaps two or three thousand craftsmen in ordinary times and several times that in emergencies, and the city thought of it, rightly, as the heart of the state.

The craft of the shipwright had been collapsed into a pattern, a stock of parts and a sequence of warehouses. A hull could be armed by men who had never built a ship and would never sail one.

**What collapsed.** The shipwright’s craft, into standard parts and a fitting-out line; the knowledge of what a galley needed, into the order of the warehouses along the canal.

**What stayed.** The men who built the hulls were the arsenalotti, organised in guilds, with positions that passed from father to son and pay that came by the day for attendance. Above them stood the Patroni and the Provveditori, the Senate’s committees of patricians who ran the yard in rotation, and above them the Senate. The design of a galley lived in the head of its master shipwright, the proto, and the Republic lived in fear of losing one.

**The number.** Ten armed galleys in six hours in 1438; a hundred in two months in 1570.

**Why it stalled.** The Arsenal had one customer, the state, and it could not go elsewhere. The state bought whatever it made at whatever pace it made it, so its number was never the one a customer would have asked for, and when the Turkish threat receded the yard kept its men and lost its pace. By the seventeenth century the arsenalotti were a privileged corporation paid to be present, the daily roll was larger than the work, and the historian of the yard records that the proportion of the day spent working fell with every decade.[^65] And the design still lived in a man. When a great proto died, so did the galley he had carried in his head, and the Republic’s ships got worse. A machine whose pattern lives in one person’s memory is a department with a canal.

## 4. The pin factory

In 1776 Adam Smith opened the most influential book on wealth ever written with the smallest manufacture he could think of.

A pin. A man not educated to the business, he wrote, nor acquainted with the machinery it used, could scarce with his utmost industry make one pin in a day, and certainly could not make twenty. But in the way the trade was now carried on, one man drew out the wire, another straightened it, a third cut it, a fourth pointed it, a fifth ground the top to receive the head; making the head took two or three distinct operations; putting it on was a trade of its own; whitening the pins another; putting them into paper another still. Eighteen operations, in some factories eighteen men; in the small one he had seen, ten men, some doing two or three operations each, and those ten men made upward of forty-eight thousand pins in a day. Four thousand eight hundred pins a man, where a man alone could not make twenty.[^66]

Smith knew exactly what he had found. The division of labour, he said, was the cause of the greatest improvement in the productive powers of labour, and it worked for three reasons: each man grew dexterous at his one motion; no time was lost passing from one kind of work to another; and a man who did one thing all day was the man most likely to invent a machine to do it for him. Fifty years later Charles Babbage, who built calculating engines and liked to count, went back to the pin factory with a stopwatch and a wage book and found a fourth reason, which has his name: once the work was split, each operation could be bought at exactly the skill it needed, so the master paid for a child’s hands where a child’s hands would do and for a craftsman’s only where they would not.[^67] The pin factory, between the two of them, is the birth certificate of the department. Every function that chapter 2 described, sales and finance and the rest, is a pin-maker’s motion scaled up to a building.

**What collapsed.** The pin-maker’s whole craft, into eighteen motions that no single man any longer needed to know.

**What stayed.** A man at every motion. And something new, which the legion and the Arsenal had both kept small and the pin factory made the organising fact of industry: between every pair of men, a handoff. The pin passed from hand to hand wherever one man’s operations ended and the next man’s began, and each pass was a place it could wait, be dropped, or be counted twice.

**The number.** One pin a day, or at most twenty, for a man alone; four thousand eight hundred for a man in the line of ten.

**Why it stalled.** Smith saw it himself and put it in Book Five, where fewer people read: the man whose whole life is spent performing a few simple operations has no occasion to exert his understanding and becomes as stupid and ignorant as it is possible for a human creature to become.[^68] The division of labour bought dexterity with attention. And the handoff it created needed a man to watch it, and then a man to watch him, and the pyramid of chapter 2 grew out of the pin factory’s eighteen steps as surely as the pin grew out of the wire. The skill had left the craftsman and gone into the sequence, but the sequence was made of men, and men between men need managing.

## 5. The line

In the spring of 1913, in a four-storey factory at Highland Park outside Detroit, a group of Henry Ford’s engineers tried something with the flywheel magneto.

Until then one man had assembled a whole magneto at a bench, in about twenty minutes. The engineers broke the assembly into twenty-nine operations, put twenty-nine men in a row, and slid the magneto from one to the next along a waist-high rail. The time fell to thirteen minutes, then, when the rail was raised eight inches so that nobody stooped, to seven, then, when a chain pulled the work at a fixed pace, to five.[^69] They tried the same thing with the engine, and then, in October, with the thing itself. A chassis that a team of men had assembled in one place in twelve hours and twenty-eight minutes was pulled along the floor by a rope and windlass past stations of men who each did one thing, and by the following spring the chassis took one hour and thirty-three minutes.[^70] The price of a Model T, which had been over eight hundred dollars in 1908, would fall, for the runabout, below three hundred by 1925, and Ford would build more cars than every other maker in the world combined.

Then the men left. In 1913 Ford had to hire more than fifty-two thousand workers to keep a workforce that at any moment numbered fourteen thousand, a turnover of several hundred per cent, because a man who had been a mechanic would not stand in one place for nine hours doing a seven-second motion for a wage anyone would pay him.[^71] Ford’s answer, in January 1914, was the five-dollar day, more than double the going rate, and it stopped the bleeding overnight. But it came with a condition that is the most instructive part of the whole story. Nearly half the five dollars was paid only to men who met the company’s standards for clean living, and to find out who did, Ford created a Sociological Department whose investigators visited the workers’ homes, asked about their drinking, their marriages, their savings and their boarders, and recommended who was fit to be paid.[^72] Ford had moved the skill of making a car into the line. He then tried to move the attention of the men who stood on it into a department that inspected their lives.

**What collapsed.** The mechanic’s craft, into a sequence of motions and a moving line; the skill of assembly, into the machine tools and the pace of the chain.

**What stayed.** The man as a part of the machine, with his hands at a station and his attention, which the line did not need and could not use, unmeasured except by whether he came back on Monday. The foreman above him and the pyramid above the foreman, which grew at Ford faster than at any company before it. And a company that, having found no way to read a man’s attention from his work, read it from his home instead.

**The number.** Twelve hours and twenty-eight minutes to one hour and thirty-three. Fifty-two thousand hires for fourteen thousand jobs.

**Why it stalled.** The line was a machine for hands, not for attention, and the attention of the men on it went where unmeasured attention goes. Ford’s own company became, within a generation, the most layered and least nimble of the great American firms, so clotted with departments that after the Second World War a team of young outsiders had to be brought in to reorganise it, in departments.[^73] And the surveillance was the doctrine’s warning two generations early: a company that cannot read attention from the result will read it from the person, and the person will leave.

## 6. The loom that stopped

In 1924 Sakichi Toyoda, a carpenter’s son who had spent his life on weaving machines, finished a loom that stopped itself.

The Type G automatic loom shut down the instant a thread broke or ran out, by itself, without a hand on it. That sounds like a small thing and was the beginning of everything that followed. Before it, a weaver stood at a loom to catch the break; after it, one worker could watch many looms, because the looms did their own watching and called for help only when they needed it.[^74] Sakichi called the idea jidoka, automation with a human touch, and his family carried it from looms into cars. The inspector had been moved into the machine.

Thirty years later Taiichi Ohno, running a Toyota plant that could not afford the inventories of the Americans it was trying to catch, took the loom’s lesson to the whole floor. A cord hung above every station, and any worker who saw a defect pulled it and stopped the line, so that a fault was found where it was made and not three departments later. Parts moved only when the next station asked for them, by a card, a kanban, that Ohno said he took from watching an American supermarket restock a shelf; the schedule left the planner’s office and became a signal on the floor. Every task was written down as standard work by the team that did it and improved by them week by week, so that the drill, for the first time since the legion, was written and revised rather than remembered.[^75]

The result was measured in 1986 by a team from the Massachusetts Institute of Technology who compared General Motors’ plant at Framingham with Toyota’s at Takaoka. Framingham needed about thirty-one hours of assembly labour per car and shipped a hundred and thirty assembly defects per hundred cars; Takaoka needed about sixteen hours and shipped forty-five. Framingham kept two weeks of parts beside the line; Takaoka kept two hours.[^76] Half the assembly hours, a third of the defects, and a sliver of the line-side inventory.

**What collapsed.** Inspection, into the machine and the cord; scheduling, into the card; the manager’s knowledge of the work, into standard work written by the team itself; a great deal of inventory, into a signal.

**What stayed.** The hands. Toyota’s plants were full of people, and the person remained the unit of production, with a team leader for every five or six, a group leader above the team leaders, and the ordinary Japanese hierarchy above that. Sales, design, purchasing and finance stayed as departments, as they were everywhere. And the drill that had been written down still had to be learned by people and lived in their habits, which the company itself called the Toyota Way, as if it were a culture rather than a method.

**The number.** Thirty-one hours against sixteen; a hundred and thirty defects against forty-five; two weeks against two hours.

**Why it stalled.** Toyota’s system reached well beyond the shop floor, into product development and suppliers, but it kept the office’s hierarchy and its human coordination, and the office is where this book began. More deeply, the method was carried in people, and people transplant slowly: when Toyota taught it to an American plant in California in the 1980s it worked, and it took years, and most of the companies that tried to copy it from books failed, because the drill was in the habits and the habits were in the men.[^77] A loom that stops itself was the first agent. A floor of them, each holding its own standard, each stopping its own line, each improving its own work without a person’s habits to carry it, is the vector.

## 7. The interface

In 2002 Jeff Bezos sent a memo to Amazon’s engineers that one of them later summarised from memory, because it was short enough to remember.

All teams would henceforth expose their data and their functions through service interfaces. Teams would communicate with each other through those interfaces and in no other way: no direct links, no shared databases, no reading another team’s data, no back doors. Every interface would be built, from the start, as if it might be opened to the outside world. Anyone who did not do this would be fired.[^78] At the same time Bezos reorganised the company into teams small enough to be fed by two pizzas, each owning one thing and answerable for it. What he had done, in effect, was to take the handoff that the pin factory had invented, the point where one group’s work passed into another group’s hands, and replace it with a contract that a machine could enforce. One team did not ask another team for anything. It called an interface, and the interface either answered or it did not, and the record showed which.

A decade later Amazon did to its warehouses what it had done to its software. In 2012 it paid seven hundred and seventy-five million dollars for a company called Kiva, whose squat orange robots carried shelves of goods to a picker standing still, instead of the picker walking miles a shift to the shelves.[^79] By 2023 Amazon had more than seven hundred and fifty thousand robots of many kinds, and in 2025 it announced its millionth, in a company that employed about one and a half million people, with nearly as many machines on its floors as people.[^80]

**What collapsed.** The handoff between teams, into an interface; the walking of the picker, into a robot; a thousand small negotiations between departments, into contracts that answered or did not.

**What stayed.** The people, in unprecedented numbers, and the layers above them. Amazon’s teams owned services, not results; a customer’s order still crossed dozens of teams between the click and the doorstep, and no team owned the whole of what the customer felt. Headcount rose with revenue over two decades, from the tens of thousands to a million and a half, which is the opposite of what a collapse does. The company became the largest and best-run department store in the history of the world, with the emphasis on department.

**The number.** Seven hundred and seventy-five million dollars for the first robots; a million of them thirteen years later, beside a million and a half people.

**Why it stalled.** The interface collapsed the message, not the department. Teams that can only talk to each other through contracts are cleaner than teams that talk in meetings, but each still owns a piece of the work rather than the result, and a company organised around pieces grows people as it grows. The doctrine takes Amazon’s interface and gives it to the agents inside one chain, so that the only human boundary left is the gate where the result becomes the customer’s. Amazon built the agents’ language. It kept the departments that spoke it.

## 8. The assistant

On 27 February 2024 a Swedish payments company called Klarna announced that an AI assistant it had switched on a month earlier had handled 2.3 million customer conversations, two-thirds of all its customer-service chats, in thirty-five languages, around the clock. It was doing, the company said, the work of seven hundred full-time agents. Resolution times had fallen from eleven minutes to under two. Repeat enquiries were down a quarter. The company expected it to add forty million dollars to that year’s profit.[^81]

Chapter 2 cited that announcement as proof that the constraint had gone, and it was. A department’s work had moved into an agent in a month. For the first time on this road, the skill that moved was not the skill of hands or of sequence but of conversation and judgement, the skill the office had always believed was its own. Klarna froze hiring and let attrition take its course, and over the following year its headcount fell by about a fifth, to around three and a half thousand.

Fifteen months later the same company’s chief executive told an interviewer that the company was hiring human customer-service staff again. The AI-first approach, he said, had produced lower quality; it had not been the right path; investing in the quality of human support was the way of the future for the company, and it was critical that a customer always knew a human was there if they wanted one.[^82] Klarna later told the American securities regulator that the remark about lower quality had been misreported and concerned a pilot of human service, not its AI, which it said had handled 66 per cent of its chats in the twelve months to March 2025. The reader should weigh both accounts. The public record does not settle what happened inside the company, and what follows is my reading of it, not a finding.

This is the most important station on the road, and it is important precisely because of the reversal. On my reading, nothing had gone wrong with the agent; what had gone wrong was the organisation around it. As far as the public account shows, Klarna had dropped an agent into a department and measured it by the department’s number, handling time and cost, and the agent had optimised the number it was given, as agents do. Nobody owned the result the customer felt. There was no vector with a human at its gate reading the record for the thing the number did not show; there was a customer-service function, with an agent inside it where the people had been, and when the company wanted better service it said it would hire more people for customer service, the old remedy. The agent existed. The collapse had not been designed.

**What collapsed.** Two-thirds of a department’s conversations, into one agent, in a month; the skill of answering a customer, which the office had thought could not leave a person, into a model.

**What stayed.** The department. The agent was placed where the people had been and measured as they had been, by cost and speed, with no human at a gate owning the result and no record read for what the number missed.

**The number.** Seven hundred people’s work; eleven minutes to two; and then, fifteen months later, a hiring notice.

**Why it stalled.** An agent without an aligner optimises what it is measured by. The doctrine’s answer to Klarna is not fewer agents or more people. It is one pair at the gates of a vector whose result is the customer’s, reading the record for quality every morning, paid on the result and not on the handling time, with the agent inside the chain where the skill belongs and the human where the judgement belongs. Klarna proved the agent in a month and disproved the department in a year, and the doctrine thanks it for both.

## 9. The cut

On 26 October 2022 Elon Musk walked into Twitter’s San Francisco headquarters carrying a sink; the next day he completed his purchase of the company. Within a week he had dismissed about half of its roughly seven and a half thousand employees. By April 2023 he told the BBC that the company had about fifteen hundred people left, down from just under eight thousand, a cut of around eighty per cent.[^83]

The service kept running. There were outages, and there were days when features broke and were fixed by whoever was left, but the thing a few hundred million people opened every morning opened every morning, posted, replied and scrolled, with a fifth of the people who had been held to be necessary to it. Musk had a method for this, which he had used in his rocket and car companies and which his biographer wrote down as a five-step algorithm: question every requirement and find out which person asked for it; delete any part or process you can, and if you are not adding back at least ten per cent of what you deleted, you have not deleted enough; only then simplify and optimise what remains; then accelerate the cycle; then, last, automate.[^84] Applied to a company, the algorithm deleted managers, meetings, teams whose work nobody could trace to the product, and whole departments whose results nobody missed when they stopped. Of all eight stations this is the one that most resembles the collapse this book describes, because it was done to the office, not the floor, and it was done in weeks, and the thing still worked.

And it stalled, for a reason the doctrine can name exactly. Eighty per cent of the people went and nothing was built in their place. On the public record, and on my reading of it, there were no vectors, no playbook, no record that anyone but the owner read, and no gate that carried a name other than his. Every decision that the deleted layers had once made rose to one man, who made them at all hours in public, and its advertising revenue, which depended on advertisers who needed to be able to predict what the service would do tomorrow, fell by about half within a year, by Musk’s own account.[^85] A collapse by subtraction leaves a smaller company with one master and no chain. The doctrine’s six are a design. Twitter’s fifteen hundred were a remainder.

**What collapsed.** Eighty per cent of the headcount, with the service still running; the layers of management between the owner and the engineers; the departments whose results nobody missed.

**What stayed.** Every signature in one man’s hands. No vectors, no record anyone else held, no gates with other names, so that the company’s judgement was one person’s and its attention was his.

**The number.** Just under eight thousand to about fifteen hundred, in six months; advertising revenue down by half.

**Why it stalled.** Deletion is half the doctrine. The other half is the chain, the gate, the record and the pair, and without them the company that remains is not collapsed but merely small. Musk proved that four-fifths of a modern office was a workaround, which this book had argued but no one had demonstrated on that scale. He did not prove what to build in its place, because he did not build it.

## 10. The tally

Eight stations, two thousand years, one sentence. Here is the sentence, station by station.

| Station | Skill moved out of people, into | People needed for the result | What stayed, and stalled it |
|---|---|---|---|
| The legion, 1st c. BC | A drill carried by the unit | A legion of farmers beat a host of warriors | The drill lived in bodies and died with them; more result meant more legions |
| The Arsenal, 1438–1570 | A pattern, standard parts, a line of warehouses | Ten galleys armed in six hours | A customer that could not go elsewhere; the design lived in one man’s head; the guild was paid to be present |
| The pin factory, 1776 | A sequence of eighteen motions | 20 pins a man to 4,800 | A man at every motion, and a handoff between each pair of men; the department was born |
| Ford’s line, 1913 | The moving line and the machine tool | 12 h 28 min to 1 h 33 min | The man as a part; attention unread from the work and read from the home; the pyramid |
| Toyota, 1924–1986 | The machine that stops itself; the card; written standard work | Half the hours, a third of the defects | The hands; the office hierarchy kept; the method carried in habits and transplanted slowly |
| Amazon, 2002–2025 | The interface between teams; the robot under the shelf | A million robots beside a million and a half people | Teams owned pieces, not results; headcount rose with revenue |
| Klarna, 2024–2025 | A model that holds a conversation | 700 people’s work in one agent | The department around the agent; no human at a gate owning the result; people hired back |
| Twitter, 2022–2023 | Nothing new; the workaround was deleted | 8,000 to 1,500, the service running | No chain, no record, no gate but one man’s; a remainder, not a design |

Read down the second column and the thread is plain: at every station the skill left the person and went into something that could hold it without him, a drill, a yard, a sequence, a line, a self-stopping machine, an interface, a model. Read down the third and the consequence is plain too: fewer people, every time, often by a factor that would have sounded absurd the day before. Read down the fourth and you have the reason this book exists. Every station kept the shape of the pin factory around the new machine: people at the stations, handoffs between them, layers above them. The skill collapsed; the organisation did not.

And this is where the road has brought us. The figures chapter 2 laid out are the state of the road today, two centuries after the pin factory and one after the line: in the American economy one manager, supervisor or administrator for every 4.7 other workers, and the cost of excess management reckoned at three trillion dollars a year;[^86] knowledge workers spending fifty-seven per cent of their time communicating and forty-three per cent creating;[^87] twenty-one per cent of employees worldwide engaged in their work, by Gallup’s count, down two points in a year;[^88] and consultants at one of the world’s best firms, handed a language model, completing twelve per cent more tasks a quarter faster at forty per cent higher quality, with the bottom half of the firm gaining most, and doing worse on the one task chosen to lie beyond the model’s reach.[^89] The office has had its loom and its line, in the model, and has not yet had its Takaoka. The vector is the office’s line. The eight stations show every piece of it being invented, one at a time, and the doctrine is the first design that uses all of them at once and stops the organisation from growing back around them.

## 11. Rules for reading the road

1. Every great gain in output in history came from moving a skill out of a person and into something that was not one. There is no exception on this road, and the vector is not one either.
2. Each move needed fewer people for the same result, and the factor was always larger than the people of the day believed possible.
3. Each move was organised, afterwards, in the shape of the pin factory: people at stations, handoffs between them, layers above. That shape is what stalled every station, not the machine.
4. A drill that lives in people must be re-taught to every generation and transplants slowly. A drill that lives in the machine does not. Toyota is the proof of the first; the vector is built on the second.
5. A machine with no customer feeling its result loses its pace. The Arsenal is the proof. Every vector has a customer and a Monday test.
6. A machine for hands that cannot read attention from the work will read it from the person. Ford is the proof, and the doctrine forbids it.
7. An interface between teams collapses the message, not the department. Amazon is the proof. The doctrine gives the interface to the agents and keeps one human boundary: the gate.
8. An agent dropped into a department optimises the department’s number. Klarna, on my reading, is the example, and the fourth form of chapter 4. The doctrine puts the agent in a chain and a pair at the gate, paid on the result.
9. Deletion without design leaves a smaller department store. Twitter is the proof. The six are built, not left over.
10. Nobody on this road was wrong. Each stopped where the organisation of his day stopped him. The reader has no such excuse, because the pieces are all on the table, and the next chapter is where he picks them up.

## 12. A note on sources

The eight stations are told from primary sources where they exist, and they exist for all but the legion’s drill, where the sources are the two ancient writers who saw it. Polybius described the Roman camp and the legion’s organisation in the sixth book of his *Histories* in the second century before Christ; Josephus described the drill in the third book of *The Jewish War*, having fought the legions and lost. Pero Tafur’s account of the Arsenal is in his own travel book, written in the 1450s about a visit of 1438; the 1570 mobilisation is from Frederic Lane’s histories of Venetian shipping, which also record the Arsenal’s decline. Smith’s pin factory is in the first chapter of *The Wealth of Nations* and his warning about its cost in the fifth book; Babbage’s extension is in his own book of 1832. Ford’s times are his own, from the autobiography he wrote with Samuel Crowther; the turnover and the Sociological Department are from Stephen Meyer’s history of the five-dollar day. Toyota’s method is from Taiichi Ohno’s own book and the Massachusetts Institute of Technology study published as *The Machine That Changed the World*; the Framingham–Takaoka figures are that study’s. The Amazon memo is known from Steve Yegge’s 2011 recollection of it, which this book attributes to him, and the robot figures are the company’s own announcements. Klarna’s figures are from its press release of 27 February 2024 and its chief executive’s interview with Bloomberg in May 2025. Twitter’s headcount is Musk’s own, to the BBC in April 2023; the algorithm is from Walter Isaacson’s biography; the revenue fall is Musk’s own statement. The chapter 2 figures are re-cited at their original sources. Where a figure is an estimate or a range in the sources, the chapter uses the more conservative end and says “about”.

[^60]: Polybius, *The Histories*, Book VI, 26–42, on the layout of the Roman camp and the organisation of the legion; the comparison to a city whose people return to their own quarters is at VI.41.

[^61]: Josephus, *The Jewish War*, Book III, chapter 5 (III.70–107, Loeb numbering), on the training of the legions: “one would not be wrong in calling their drills bloodless battles and their battles bloody drills.”

[^62]: Vegetius, *Epitoma rei militaris* (De Re Militari), late fourth century AD, Book I, on the lapse of the ancient drill and the need to restore it.

[^63]: Pero Tafur, *Travels and Adventures, 1435–1439*, trans. Malcolm Letts (George Routledge, 1926), the Venice chapter: “and so, from each side, everything that was required, and when the galley had reached the end of the street all the men required were on board… In this manner there came out ten galleys, fully armed, between the hours of three and nine.”

[^64]: Frederic C. Lane, *Venice: A Maritime Republic* (Johns Hopkins University Press, 1973), on the Arsenal’s reserve of hulls and the mobilisation of 1570; and Lane, *Venetian Ships and Shipbuilders of the Renaissance* (Johns Hopkins, 1934).

[^65]: Robert C. Davis, *Shipbuilders of the Venetian Arsenal: Workers and Workplace in the Preindustrial City* (Johns Hopkins University Press, 1991), on the arsenalotti’s privileges, attendance pay and the fall in hours actually worked.

[^66]: Adam Smith, *An Inquiry into the Nature and Causes of the Wealth of Nations* (1776), Book I, chapter 1: the eighteen operations, the ten men, and “upwards of forty-eight thousand pins in a day.”

[^67]: Charles Babbage, *On the Economy of Machinery and Manufactures* (Charles Knight, 1832), chapter XIX, “On the Division of Labour,” with his table of pin-making operations, their times and their wages.

[^68]: Smith, *Wealth of Nations*, Book V, chapter 1, part III, article II: “The man whose whole life is spent in performing a few simple operations… generally becomes as stupid and ignorant as it is possible for a human creature to become.”

[^69]: Henry Ford with Samuel Crowther, *My Life and Work* (Doubleday, Page, 1922), chapter V, on the flywheel magneto line: twenty minutes to thirteen minutes ten seconds, then seven, then five.

[^70]: Ford and Crowther, *My Life and Work*, chapter V: the chassis “assembling line” and the fall from twelve hours and twenty-eight minutes to one hour and thirty-three minutes. The Model T’s price history is in David A. Hounshell, *From the American System to Mass Production, 1800–1932* (Johns Hopkins University Press, 1984).

[^71]: Stephen Meyer III, *The Five Dollar Day: Labor Management and Social Control in the Ford Motor Company, 1908–1921* (SUNY Press, 1981): in 1913 Ford hired more than 52,000 workers to maintain a workforce of about 14,000.

[^72]: Meyer, *The Five Dollar Day*, on the Sociological Department and the conditions attached to the profit-sharing portion of the wage; the five-dollar day was announced on 5 January 1914.

[^73]: Hounshell, *From the American System to Mass Production*, on Ford’s later organisation; the postwar reorganisation by the group known as the Whiz Kids is described in John A. Byrne, *The Whiz Kids* (Doubleday, 1993).

[^74]: Toyota Industries Corporation’s own account of Sakichi Toyoda’s Type G automatic loom of 1924 and the principle of jidoka; and Taiichi Ohno, *Toyota Production System: Beyond Large-Scale Production* (Productivity Press, 1988; Japanese original 1978), chapter 1.

[^75]: Ohno, *Toyota Production System*, on the andon, the kanban and the supermarket, and standard work.

[^76]: James P. Womack, Daniel T. Jones and Daniel Roos, *The Machine That Changed the World* (Rawson Associates, 1990), chapter 4, the comparison of GM Framingham and Toyota Takaoka in 1986: adjusted assembly hours per car 31 against 16; assembly defects per 100 cars 130 against 45; inventories of parts two weeks against two hours.

[^77]: Womack, Jones and Roos, *The Machine That Changed the World*, on the NUMMI plant at Fremont, California, opened in 1984 as a joint venture of GM and Toyota.

[^78]: Steve Yegge, “Stevey’s Google Platforms Rant,” October 2011, recounting Bezos’s 2002 mandate; and Brad Stone, *The Everything Store: Jeff Bezos and the Age of Amazon* (Little, Brown, 2013), on the two-pizza teams and the mandate.

[^79]: Amazon.com, Inc., press release, 19 March 2012, on the acquisition of Kiva Systems for $775 million in cash.

[^80]: Amazon’s own announcements: more than 750,000 mobile robots deployed (2023) and the millionth robot (2025); employee count from Amazon’s annual report for 2024, about 1.55 million.

[^81]: Klarna, press release, 27 February 2024: 2.3 million conversations in the first month, two-thirds of customer-service chats, the work of 700 full-time agents, resolution under two minutes against eleven, repeat enquiries down 25 per cent, an estimated $40 million profit improvement in 2024; as cited in chapter 2.

[^82]: Sebastian Siemiatkowski, interview with Bloomberg, May 2025, as reported on 8–9 May 2025: the AI-first approach had produced “lower quality”; “investing in the quality of human support is the way of the future for us”; headcount down about 22 per cent to roughly 3,500 after a hiring freeze from late 2023. Klarna Group plc, letter to the US Securities and Exchange Commission, 16 July 2025 (Form CORRESP): the company stated that the “lower quality” remarks had been misreported and concerned a pilot using freelance customer-experience staff, not AI-driven service, and that its AI assistant handled 66 per cent of customer-service chats in the twelve months to 31 March 2025.

[^83]: Elon Musk, interview with the BBC, 12 April 2023: about 1,500 employees, down from just under 8,000; the acquisition closed on 27 October 2022.

[^84]: Walter Isaacson, *Elon Musk* (Simon & Schuster, 2023), on “the algorithm”: question every requirement, delete any part or process you can, simplify and optimise, accelerate cycle time, automate; and the rule that if you are not adding back at least ten per cent you have not deleted enough.

[^85]: Musk’s own statement, posted on the platform on 15 July 2023, that the company remained cash-flow negative because of an approximately 50 per cent drop in advertising revenue.

[^86]: Gary Hamel and Michele Zanini, “Excess Management Is Costing the U.S. $3 Trillion Per Year,” *Harvard Business Review*, 5 September 2016; as cited in chapter 2.

[^87]: Microsoft, *Work Trend Index Annual Report*, May 2023; as cited in chapter 2.

[^88]: Gallup, *State of the Global Workplace* 2025; as cited in chapter 2.

[^89]: Fabrizio Dell’Acqua et al., “Navigating the Jagged Technological Frontier,” Harvard Business School Working Paper 24-013, September 2023; as cited in chapter 2, including the task outside the frontier.

================================================================
# Chapter 20: The Reader’s Month
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-20-the-readers-month

Nothing in this book asks the reader to believe it. It asks for one month and one vector. Here is the month, as a calendar, with the Monday as day one. It is the test the doctrine itself was first run on, at the scale of a single result, and it costs almost nothing.[^90]

**Week one: find the vector.**

*Monday.* Pick one result your customer feels, from first word to cash: one order, one hire, one claim, one patient, one case. Not a department’s output; a result someone outside the company receives and would miss. Write it in one line. Then write, on one page, every desk it crosses today between the first word and the cash, in order. Count the handovers. That number is the one you will beat.

*Tuesday.* Take the last ten instances of that result and, from your own records, write down how many days each took, and where each waited. You will find that most of the days were spent between desks, not at them. Mark the three longest waits. Those are the leaks of chapter 3, and they are yours.

*Wednesday.* Mark the gates: the three or four points in the chain where an act becomes irreversible, where money leaves, a name goes out or a promise is made. Everything else in the chain is reversible and needs nobody’s approval. Chapter 17 is the rule; most readers find they have been approving twenty things and only four of them were gates.

*Thursday.* List the skills the chain needs, step by step, and beside each one write whether an agent that exists today can do it: answer the enquiry, draft the document, schedule, check the record, chase the payment. Be honest in both directions. In my experience, for most office results in 2026 the answer is yes for most steps; measure your own.

*Friday.* Name the pair. One aligner, who will hold the gates for the month, and one tuner, who will stand behind her and change the chain when it is wrong. Choose them by chapter 12: not the two most skilled people on the current chain, but the two who most want the result and will tell you the truth about the number. Tell them the month’s rule: they hold the gates, nobody else touches the result, and the record keeps everything.

**Week two: build beside the old chain.**

*Monday to Wednesday.* Build the chain: the agents from Thursday, in order, with the record in the middle so that every act is written, and the four gates with the aligner’s name on them. Use what exists. Nothing in this week should be invented; everything should be assembled. But do not build the chain on the department’s screens. An agent that fills a form designed for a clerk, or waits in an inbox designed for a manager, is the motor belted to the old shaft of chapter 4. Connect it to the record directly, and leave the screens to the old chain.

*Thursday and Friday.* Run the chain beside the department, on live instances, with nobody acting on what it produces. Let it answer the enquiry the department also answers, draft what the department also drafts, chase what the department also chases, into the record only. Compare. This is the fortnight chapter 13 gave the revenue chain; run it for two days at least, and do not switch until the chain handles the exceptions and recoveries you have seen as well as the routine. Fix what broke.

**Week three: switch.**

*Monday.* For this one result, the department stands down. Every new instance goes into the chain. The aligner holds the gates; the tuner reads the record every evening and changes what the agents did wrong; the record keeps every day. Nobody in the company intervenes on this result by any route except the gates, and anyone who tries is told, kindly, that the test is the test.

*Tuesday to Friday.* Hold. There will be a day when something goes wrong that the chain has no step for. It rises to you, as chapter 17 says it must, with the record of why. Add the step, name the gate, and go on. Do not add a person.

**Week four: the Monday test.**

*Monday.* Read the week. For every instance of the result that entered the chain and finished: how many days from first word to cash; for every one that has not finished, its age and where it waits, and keep watching it; against the ten you measured in week one; how many handovers, against the number you counted; how many gates, against the twenty you used to approve. Read the aligner’s week from the record, not from her: gates answered, how fast, what she caught. That is attention, read from the work, as chapter 11 reads it.

*Tuesday to Thursday.* Let it run, and watch the number move each day as the tuner’s changes take hold. The vector grows by design, and in its second week the design shows.

*Friday.* Decide, and write the decision in a record of one line. Either the vector stays open and the department never gets this result back, and you have a vector, a pair, a record and a Monday test, and the collapse of chapter 13 has begun at the one place it should: the result nearest the cash. Or the vector did not beat the department, and you write down in one line why, often, though not always, one of chapter 19’s rules for reading the road, and you fix that and run the month again, or extend it, or stop.

That is the whole test. One result, one pair, one month, one line at the end of it. The eight stations of chapter 19 took two thousand years to invent the pieces. The reader has them all, and a Monday.

[^90]: The month described here is the one the author’s own company first ran on a single recruiting result before any department was dissolved; chapter 6 shows the vector it produced. The test does not depend on it.

================================================================
# Chapter 21: A Letter to the World
URL: https://vinaypasricha.com/books/collapse-or-collapse/chapter-21-a-letter-to-the-world

To whoever is reading this, in whatever year, holding a company, a ministry, a job, or a child’s future:

I am writing to tell you what I believe will happen. Not what might, and not what I hope. The twenty chapters before this one are argument and method and proof, and you may weigh them as you like. This letter is different. It is what I see when I look down the road the book has drawn, and I have written it plainly so that it can be checked against the years as they come, because a prophecy that cannot be checked is only a mood. These dates, and the claim of long-run abundance, are my forecasts, not findings; the chapters before this one are the evidence, and the years will be the test.

Three clocks are running as you read, and none of them is yours to stop.

The first is a competitor. Somewhere in your industry, this year or next, one company will take the chain from first word to cash out of its departments and put it into vectors, with one pair of people at the gates, and from that quarter it will quote in hours, deliver in days and keep ninety-eight rupees of every hundred of growth. You will hear of it as a rumour, then as a lost customer, then as a price you cannot match. The second clock is the machine itself. The capability of the models that hold skill is doubling faster than any reorganisation you could begin today will finish, so that every quarter you wait, the gap between what the chain can carry and what your departments carry grows, and the cost of closing it grows with it. The third clock is your best people. The first collapsed firm in your industry will pay its aligners far more than you pay, for a result they own, and the ones who leave you will be exactly the ones who wanted the result most. Those three clocks are the reason the title of this book has no comma of mercy in it. You will collapse your organisation, or it will be collapsed.

Here is how the years will go.

## The first three years

In the first three years the skill leaves the person and does not come back.

Every skill an office result needs, the answering, the drafting, the searching, the checking, the scheduling, the chasing, the pricing, is already in the machine as you read this, and within three years it will be there completely, at a cost a small firm can pay. The visible sign will be small and you will almost miss it: departments will stop hiring for skill. The advertisement that asked for eight years’ experience will ask for none, and nobody will quite say why. In every industry one company, usually not the largest, will do what chapter 13 describes and collapse its revenue chain on one Monday, and within a year its people will be paid like partners and its competitors’ best people will be writing to it. You will say that your industry is different, that it is regulated, relational, technical, local. Every industry has said this, and in every industry the first firm will be in a part of it nobody thought could move. Governments will notice in these years, commission a report, and do nothing, which is correct, because nothing has yet happened to them.

## Years four to six

In years four to six the second collapse begins, and it is the one the book is named for.

The firms that did not collapse themselves will be collapsed by their markets. A department-run company competing with a vector-run company loses on price, on speed and on people at once, and it does not lose slowly; it loses a customer a week until a board meeting that should have happened three years earlier happens in a hurry, and the collapse is done by force, badly, by people who have not read the record. This is the collapse that costs jobs carelessly, and it is the one I am writing to spare you. In the same years the vectors of the first collapsed firms will cross their own walls, lent whole to companies in markets they do not serve and lent at the edge to the customers and suppliers who still run on departments, as chapter 18 describes, and the first industries will find that most of their results are produced by a handful of chains that a few hundred people hold. Unemployment will rise in these years, visibly, for the first time, and it will be worst where companies protected their departments longest, because a forced collapse pays nobody out well. You will say that customers want to deal with people. They want the result, and they will deal with a person at the one gate where a person matters, which is what they always wanted and were never given.

## Years seven to nine

In years seven to nine prices fall, governments act, and the worst passes.

When an industry runs on a few shared vectors, its costs fall by the fraction the chapters have shown, a tenth of the people for four times the result, and in a competitive market most of that reaches the customer as price. Insurance, logistics, professional services, recruitment, banking, the back half of every manufacturer: each will be a fraction of its former price, and the money that used to pay for departments will stay in households. Governments will act in these years, late and all at once: income floors, because the old link between a job and a living has visibly broken; a shorter statutory working week, because there is less work to share; and taxes on the output of vectors, because that is where the value now is and where it can be counted. You will say the state will stop it. The state will tax it, then copy it, in that order. And in these same years the thing the book has promised will become visible in the figures: new Outcomes will be born faster than old ones die, because a result that costs a tenth of what it did is a result ten times as many people can buy, and every new buyer is a new unowned result for someone to flag. Humans will move, in large numbers, from skill work to intention work, from doing the steps to holding the gates and opening what comes next, and the unemployment that peaked in year six will be falling by year nine.

## Years ten to twelve

In years ten to twelve countries diverge, and the state collapses itself.

The countries that let their companies collapse early, and caught the people who fell with floors rather than with protected departments, will be rich in a way that will embarrass the ones that did not; the ones that protected their departments by law will have kept their jobs for a decade and lost their industries at the end of it. The state itself, which is the largest department store in every country, will begin to run as Outcomes: a result a citizen feels, a chain that produces it, a master who holds it, a Keeper above her, and the ministry dissolved into agents inside the chain. A small state will do this first, and I believe it will be Singapore; New Zealand will be the second, before the span is out, because it is small enough to decide and honest enough to measure. The large states will come last, and the ones with the most citizens waiting on the most desks will come last of all, because the desks are where their power sits. And money will change its shape in these years. Pay as a share of a result, equal to the base at target and rising without a ceiling, will be the ordinary arrangement, and the word salary will begin to sound the way the word wage sounds to us now: a thing paid for time by people who could not yet measure the result.

## Years thirteen to fifteen

In years thirteen to fifteen the job as we knew it is gone, and we are fine.

Most adults will not hold a job in the sense their parents did, a place they went for hours to perform a skill under supervision, because almost nothing that was a skill is still done by a person and almost nothing that is done by a person is supervised. Output across the industries that collapsed will be many times what it was at a fraction of the labour, and the prices of the things those industries make will be near enough to zero that the question of how to afford them will have become a question for historians. You will ask what a human being does when skill is no longer needed, and I will tell you, because the chapters have already answered it. A human holds a gate: decides the irreversible thing with her name on it, which no amount of capability has made unnecessary, because the question at a gate is not what is correct but what we are willing to be held to. A human wants a result and opens a vector to get it, which is the only way anything new has ever begun. A human cares for another human, which was never a job and was always most of the work. And a human chooses what to want, which is the thing the pin factory took away from the man at the eleventh motion and the collapse gives back. The question of meaning was a question the job asked. The job is gone. The question goes with it.

## Unemployment

I have left the hardest objection for its own place, because it deserves one and because it is the one every reader has been holding since chapter 2.

Will this cause mass unemployment? In years four to nine, yes, in the countries and industries that wait to be collapsed rather than collapsing, and I will not pretend otherwise; a forced collapse is a careless one, and carelessness has a cost in people. But it will not cause the permanent idleness of a population, and I can say why with three arguments the book has already made.

History first. Every station on the road of chapter 19 cut the labour a result needed, by factors that terrified the people of the day, and every one of them left more people in work a generation later, not fewer, because a cheaper result is a result more people buy, and more buying is more results, and every result is work. The pin factory did not end the pin-makers; it made pins so cheap that a thousand trades that had never used a pin began to, and the trades needed people. Arithmetic second. A company that produces four times the result with a tenth of the people has not destroyed nine-tenths of the demand for what it makes; it has cut its price, and a lower price finds buyers who were never customers before, and those buyers want things nobody has yet built, which is a new Outcome, which needs a master, a pair and a chain. Chapter 16 shows how one is born; the world of years seven to nine will show thousands being born a month. And the human case last. The work that is going is skill work, and skill is the one thing the machine does better. The work that is coming is intention work: holding gates, opening vectors, caring, choosing, and there is no limit to it, because there is no limit to what people can want for themselves and for each other. The pin factory’s child, who was made as stupid as a human being can be made by one motion repeated for forty years, can learn a cockpit in a week, because the cockpit asks nothing of him but attention and the truth about a number.

So the honest prophecy is this. There will be a hard decade for those who held skill and nothing else, and for the companies that protected departments until the market took them. There will not be a permanently idle humanity. The thing a country can do about the hard decade is to collapse early, and to catch the people who fall with a floor, and to open the cockpits to everyone, selected on intention and nothing else, which is the one selection a nation can afford to make universal. I know this costs something to believe, so I will tell you what it cost me.

In 1989 I collapsed a reorder book: five men, five desks and a month, in a company that was not mine, turned into a program that ran in an hour, with the registers kept beside it until nobody trusted them more than the screen. That was the first collapse in this book and the smallest. Thirty-six years later I did it to my own company. GoodSpace runs today as Outcomes and vectors, with pairs at the gates and a Keeper who is not me, and it is not finished, and the places where it falls short of its own doctrine are named in these chapters where they belong. It cost me the pleasure of being needed, which is the thing an owner gives up last. What it gave back was a company that runs when I am asleep, and people in it who are paid for a result they own. I would do it again tomorrow, and I am asking you to do it before then.

## And after

Beyond year fifteen I will say only three lines and show you one morning.

Every result on earth will be produced by a chain that holds every skill it needs. Every chain will be held at its gates by a few people chosen for what they want. Every person will be free to want.

And here is the morning. Meera Kulkarni, whom the reader left in chapter 14 with her vector lent across its first wall, wakes on a terrace outside Pune, at the hour she chose, and the chain tells her what it did in the night: forty orders quoted, thirty-one won, the first castings poured in a plant where no one has stood since her daughter was small, a client in another country whose department head signed beside her at four in the morning. Two things need her. She reads them, in the order the room has already put them, and signs one and holds the other, with her name, and her name means something because it is the only one on the gate. By nine she is done with the company, which is the size of a nation and runs through her terrace. The rest of the day is hers. She spends part of it on a flag she is writing, because three of her clients have asked for something nobody makes, and she wants to be the one who opens it; part of it with her mother, who is old; and part of it doing nothing in particular, which the pin factory never allowed anyone. Nobody asks her hours. Her share rises with the result, and the result rose again this month, as it has every month since the vector was lent across its first wall. She is not unusual. She is what a working person is, after.

Collapse your organisation now, or your organisation will collapse.

Vinay Pasricha

Noida, October 2026

================================================================
# The Twelve Tenets
URL: https://vinaypasricha.com/books/collapse-or-collapse/twelve-tenets

I. **The department was a workaround.** Functions existed because no one person could hold every skill a result needs. The constraint has gone.

II. **Organise around outcomes, not activities.** Work that produces a result is an Outcome; work that serves a result is an agent. There is no third thing, and no function sits beside the Keeper.

III. **Every outcome runs on a vector.** One result, one market, one chain of agents, one pair of humans by default.

IV. **Machines hold the baseline; humans supply the peaks.** The chain carries the continuous work; the human is spent only where a human changes the result.

V. **Pay for intention, not skill.** Skill is in the chain and cheap. Intention that holds is rare, and is paid very highly, far above the market.

VI. **No silent stalls.** Every agent reports its own flow; a vector that can stall unseen is a department with software.

VII. **Every vector has a standby.** The tuner for the aligner, another master for a master, a board member for the Keeper; reversible gates default open after the period the playbook sets.

VIII. **Count vectors, not employees.** The unit of production is the vector; the company’s size is the number of results it carries at once.

IX. **The aligner’s clock is her own.** The agents work on the customer’s clock; she works on hers, for hours she fixes and keeps.

X. **Hire for intention; measure attention.** Intention is selected once, at the door, on the eight values. Attention is read every day from the work, never from the person.

XI. **Measure, don’t price.** Outcomes serve each other without invoices. Value between them is measured and not priced.

XII. **A vector is a vector, whoever pays it.** Employee, contractor, agency or lent chain: whoever works from a cockpit and is read by a control room is inside the company.

================================================================
# The Vocabulary
URL: https://vinaypasricha.com/books/collapse-or-collapse/vocabulary

**The people**

*Keeper.* The one human who holds the company: allocator and judge, owner of the four core outcomes, designer of the whole, keeper of the doctrine. He runs nothing. Beside him, two humans and no functions; at one Outcome, one backup. *Master.* The one human who holds an Outcome: owns its playbook, reads its floor, decides its appeals and its Monday test. She has no staff and no agents beside her. *Aligner.* The one human who runs a vector and holds its gates. *Tuner.* The one human who builds and changes a vector’s chain, accepts or refuses the agent maker’s proposals, and can run the vector alone. *Pair.* An aligner and a tuner, each the other’s standby. *Backup.* The second person who shares a master’s or a Keeper’s room and can hold it alone any morning. The six: aligner, tuner, master, Keeper, and a backup for each of the last two.

**The units**

*Outcome.* A result someone outside the company receives and would miss, with the vectors that produce it and a master who holds it. It exists exactly as long as the result has no other owner. *Vector.* One running copy of an Outcome’s discipline: one result, aimed at one market, carried by one chain of agents, held by one pair. Vectors deepen, merge, and are lent. *Agent.* A skill, held by the machine, standing inside the vector that needs it, behind a gate. Every former function is either an agent or an Outcome. *Quartermaster.* The agent that manages a vector’s line agents and answers for the whole vector; the one face the aligner sees. *Line agent.* An agent that does one step of the chain. Among them, in every chain: the inspector, which checks every output against the playbook and may hold; the accounts agent, which keeps the vector’s cost and value and pays within its limit. *Agent maker.* The agent that builds, tests and upgrades a vector’s agents and proposes changes to its tuner. *Hands.* Any non-human automation, a humanoid, a drone, a wheeled robot, a machine that loads itself. An agent with hands is an agent.

**The rooms**

*Cockpit.* The pair’s room: one face at a time, nothing shown until asked, the gates, the signature. *Control room.* The master’s room: the floor of tiles, one per vector, coloured by state; the red nearest her feet. Nobody beside her. *Command centre.* The Keeper’s room: four columns, profit, customers, employees and society; the Outcomes as tiles; the doctrine tile. *Tile.* One vector or one Outcome, shown as a colour: pale is flowing, amber is a warning, red needs a human now. The tile takes the colour of the worst thing inside it. *Doctrine tile.* Three counts that must read zero: humans outside the structure, signals read from outside the engine, vectors without a pair or a written exception. *Edge colour.* The colour inside a tile that shows days lost outside the vector, at its edges and in the uncollapsed world.

**The acts**

*Gate.* A point where the chain waits for a signature. Irreversible gates (money, a name, a promise) carry a name and default closed. Reversible gates default open after a period the playbook sets; irreversible gates default closed. *Split.* A new Outcome opened from an existing playbook when a master’s floor holds more reds than she can read. *Co-signature.* The second name on an irreversible gate while its holder is below the threshold of judgement. *Signature.* The only authority in the company. Whoever signs a gate owns the result that follows, up to the next gate. Anyone whose name is on the tile may sign. A signature is never overruled. *Double signature.* The gate at the edge of a lent vector: two single signatures on one line, the vector’s aligner and the other company’s department head. *Hold.* An inspector’s signature that stops an output; either of the pair may appeal; the master decides at her own gate. *Flag.* A line in the record raised by any human with a name on a tile, naming an unowned result, its receiver, and why no chain produces it. Only the Keeper decides a birth. *Written on the wall.* The death of an Outcome: another Outcome produces its result, or its customer is gone. No decision is needed; the Keeper records it. *Monday test.* The one question each part of the company answers each week: the result, with no department in the chain.

**The rules of motion**

*Playbook.* The written design of an Outcome: chain, gates, limits, standards, thresholds. The master owns it; every agent maker reads it; nothing spreads between vectors except through it. *Record.* Everything every agent and human did, written, timed and replayable. A failure is reviewed from the record, in proportion to its size. *Design rate.* The rate at which a vector’s result is expected to grow as its agents compound. The gap between the design rate and the actual rate is the pair’s form. *Deepening.* The same vector producing a larger result each month, until one vector carries the company’s whole result; then it is lent. *Lent vector.* A vector running inside another company, whole or at its edge, behind a double signature, with the pair’s share following the result across the wall. *Collapse.* Dissolving departments into Outcomes and agents by the one test, revenue chain first and whole, then one department at a time. The second collapse is the one the market does to a company that waited. *Trial.* A known result tried in a new market: one vector, from an existing playbook, a quarter to prove it. Trials make vectors; births make Outcomes. *Born whole.* How an Outcome begins: at least one master and one vector of a new kind, from day one. A vector never becomes an Outcome. *Mirror, rest, replacement.* What follows falling attention: shown the record, given paid rest while the tuner runs the vector, then replaced. No warnings, no improvement plans. *The eight values.* What intention is, and what the four layers test: Teamwork, Friendliness, Helpfulness, Honesty, Reliability, Truthfulness, Integrity and Kindness, each tested by a behaviour the record can show. Each has a minimum standard, and a human reviews every verdict. *Setting.* A number the doctrine gives as a starting value and the playbook changes on the evidence, such as how long a reversible gate waits. A tenet is a principle; a setting is not.

================================================================
# The Scorecard
URL: https://vinaypasricha.com/books/collapse-or-collapse/scorecard

Twelve questions, to be asked of your own company on a Monday morning and answered honestly by lunch. Twelve yes: your company has collapsed, whatever it calls itself. Fewer: it has not, whatever it calls itself.

| | Tenet | The Monday test | Yes / No |
|---|---|---|---|
| I | The department was a workaround | Pick any department. What result does it produce that a customer would notice if it stopped? If it only helps the others produce theirs, its work belongs inside their chains. | |
| II | Outcomes, not activities | List everyone who reports to the top. For each: what Outcome does she master? If the answer is a function, the department has come back through the front door. | |
| III | Every outcome runs on a vector | For each result you sell, can you point to the one chain that produces it end to end, and name the two people who hold it? | |
| IV | Machines hold the baseline | For one vector, count last week’s human hours spent on work an agent could carry to standard, and the hours spent where only a human could change the result. Is the first number falling every quarter? | |
| V | Pay for intention | What would you pay the one person whose intention you would trust with a whole result? Do you pay it? | |
| VI | No silent stalls | Find the oldest unfinished piece of work in the company. How long has it waited, and who knew before you asked? | |
| VII | Every vector has a standby | Name the person whose two-week absence would stop a result. Name who runs it in her place, starting tomorrow, without a handover. | |
| VIII | Count vectors, not employees | Describe your company in one sentence without saying how many people work in it. | |
| IX | The aligner’s clock is her own | When in the day is your best people’s attention highest? Is that when they do the work that needs it? | |
| X | Hire for intention; measure attention | The last person who left of her own accord: how many months before her resignation could the record have told you? | |
| XI | Measure, don’t price | How many hours did your managers spend last quarter negotiating budgets, recharges or credit with other parts of your own company? | |
| XII | A vector is a vector, whoever pays it | Every result you depend on that someone you do not employ produces: can you see each on the same screen as your own people, with the same alarms? | |

================================================================
# The Key to the Pictures
URL: https://vinaypasricha.com/books/collapse-or-collapse/key-to-the-pictures

**How to read them.** A person in warm light on the plain floor, with no label, is a human. A figure of cool blue light on a disc, with a name and a role above it, is an agent. White light is data; gold is a result delivered; amber is a warning, a slowing flow or a customer who may leave; red is stopped, or needing a human now. A glowing threshold on the floor is a gate. The aligner is the woman in the white shirt; the master is the older woman in the dark jacket; the Keeper is the man at the table.

| No. | Chapter | Picture |
|---|---|---|
| 01 | 6 | She enters. The quartermaster, a figure of blue light on his disc, answers; the wall forms behind him with only what happened overnight. |
| 02 | 6 | Where are we? The vector’s journey as a path of light on the floor; every mandate in flight a white mark, the gold marks signed; the two of them standing at today. |
| 03 | 6 | A reversible gate: a batch of five hundred prospects waits as a cloud of white points behind a threshold of light; she says the word and it opens. Unanswered past its period, it would have opened on its own. |
| 04 | 6 | The signature: he reads the contract back; the page hangs between them; she signs. The one act only a human performs. |
| 05 | 6 | A meeting: the aligner, solid and warm on the plain floor, with the whole vector around her, every agent labelled. Six agents and one human, and nobody has to be told which is which. |
| 06 | 6 | Her clock is her own: a dawn terrace, tea in hand, the quartermaster beside her on his disc; half a world away the agents work on a customer’s clock. |
| 07 | 6 | The connected boundary: the aligner with the humanoid, present twice, as an agent of light in the room labelled FLOOR and as itself at the machine beyond the glass. The hands have connected. |
| 08 | 7 | The control room: the master before a floor of a thousand vectors, three red at her feet, the nearest red vector’s own quartermaster risen from its tile to explain. The agent belongs to that vector, not to her. |
| 09 | 8 | The command centre at rest: four columns of light for profit, customers, employees and society, each with a gold line; the Outcomes as tiles, one amber and one red; the blue threads showing which feeds which; the Keeper with his two humans; the small tile at the end reading DOCTRINE 0 0 0. |
| 10 | 12 | Selecting for intention: the four layers, declared, record, witnesses, live, as a single vertical chain ending at the tuner’s signature. |
| 11 | 13 | The collapse itself: ten departments above, five Outcomes below with their vector counts, and two arrows, 1,000 to 100 and 1x to 4x. |
| 12 | 14 | The ladder is gone: a ladder lying on its side beside a single vector drawn as a shaft going down, crossing the line of the company’s edge. |
| 13 | 15 | The last six: five Outcomes collapsing into one, then one into one vector, with six figures in two rooms. |
| 14 | 16 | Birth and death: the Keeper’s table with one tile going dark and one lighting, twelve lines apart in the record. |
| 15 | 17 | The signature: one order drawn as a chain with eight seals on it, the fifth one red. |
| 16 | 18 | The edge of the vector: one chain across three companies, two seals at each wall, and the days lost written along the stretches outside. |
| 17 | 19 | The long road: eight stations, each labelled with what moved out of people, ending at the vector. |