Correction and Recovery: what to do when a cycle fails.

How do you recover when an improvement effort fails?

Chapter 15 starts from the premise that failures will happen: a constraint misidentified, an improvement that does not work, a cycle that falls short, commitment that wavers, events that disrupt the plan. A doctrine has to govern recovery as well as success.

It sets out five principles. Honesty: name the failure directly, because hiding or minimising it creates a false sense of progress. Rapid re-diagnosis: go back to the Check stage and ask why output fell short and where in the sequence. Failure is often more instructive than success, because it shows what is actually limiting output. Recovery needs the same discipline as improvement, not a lurch back to old methods or on to something entirely new. And a cycle is not wasted if it revealed that the constraint was misidentified.

The chapter adds practical guidance. Do not overreact to one bad cycle; a pattern of them is what signals a fundamental problem. Be willing to admit the diagnosis was wrong and return to the data. Distinguish a flawed sequence from a good plan the team did not follow: the second is a human constraint that needs clarity and accountability, and the sequence should be redesigned only if it was itself flawed. Treat effort spent on the wrong constraint as sunk cost. Recovery, in the book’s view, is normal and expected.

Chapter 14 lists the failure modes this chapter recovers from. For the thinking that should come before a fresh attempt, see The SIV Method.