Sequencing

Two organizations attempt the same change with comparable resources and comparable people. One succeeds and one does not, and the difference is frequently the order in which things were done.

Order matters more than content because it determines who has to act before there is any evidence that the change works, and acting without evidence is the thing people find hardest.

The general shape

Establish the problem before proposing the answer. A solution offered to people who do not believe there is a problem is heard as a preference. Frequently the first several weeks of a change programme should contain no proposal at all, only a shared picture of the situation.

Secure the people whose assent decides it. Not everybody. The small number whose participation others will read as a signal.

Change the structures before asking for the behaviour. If the metric, the form, or the approval path still favours the old way, the request is competing against the organization’s own machinery.

Do the visible thing early. One concrete alteration people can see, so that the change has a physical existence rather than a rhetorical one.

Remove the old way. Later than the launch and earlier than feels comfortable, once the new one is genuinely operable.

Encode it. Into hiring, promotion, budget, and calendar, which is what determines whether it survives the leader’s attention moving on.

The most common inversion

Announce first, then work out the implementation.

The reasoning is that the announcement creates commitment and the details will follow under pressure. What actually happens is that people are told to do something the organization is not yet able to do, they attempt it, it does not work, and the failure occurs in public.

That first failure is expensive out of proportion to its size, because it is the first evidence anybody has about whether the change works, and the evidence is negative. Every subsequent difficulty is then interpreted against it.

The second inversion is asking for behaviour while the structures still reward the opposite, which is covered in the previous subchapter and is the more insidious of the two because nobody has to decide anything for it to defeat the change.

Start where the evidence will be generated

Where a change can be introduced in one area first, the choice of area is a sequencing decision and it is usually made on the wrong criterion.

The instinct is to start where resistance is lowest. That produces a successful pilot in the least representative setting, and the result persuades nobody, because everybody else can name the reasons their situation is different.

Better to start somewhere ordinary, with a supervisor whose judgment other supervisors respect. A change that works in a typical unit, run by somebody credible, produces evidence that transfers. It also risks a visible failure, which is the price of evidence that means anything.

Pace against absorptive capacity

The binding constraint on a change programme is not usually resources. It is how much alteration people can absorb while continuing to do their jobs.

That capacity is finite, it varies by area, and it is smaller than leaders estimate because the leader is not absorbing anything. They are directing.

Two practical consequences. Sequence changes so that no single group is absorbing two at once, which frequently means a slower overall programme and a faster completed one. And check the pace with the people carrying it rather than assuming, because they will not volunteer that it is too much.

Write the sequence down

Order, owners, dates, and the conditions under which the next step begins. One page.

The value is not the plan surviving contact, which it will not. It is that a written sequence makes the dependencies visible, so that when something slips it is apparent what else has to move, rather than the programme quietly proceeding to step four with step two incomplete.

Include the reversal conditions. If this is not producing anything by a stated date, we stop. Written at the beginning, that decision is made by somebody with no stake in the answer, which is the argument made throughout this Library for pre-committing to exit criteria.

The entries that follow deal with the pieces: the early win and its trap, identifying who must be persuaded first, communicating over months rather than at a launch, and encoding a change so it survives.