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.

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.