Main idea

"Why didn't we do this before?" is a question asked in good faith, and almost always badly asked. Whoever asks it has today's context — a cost that has become low, a quality that has become acceptable, a skill now present in the team — and applies it retroactively to a situation that did not have it.

The answer generally exists, and it is mundane: at the time, the cost, the quality, the available resources and the commercial context did not make it possible to do it properly. But that answer is available only if someone wrote it down at the moment of choice. Unwritten, it cannot be reconstituted: nobody faithfully remembers what they did not know two years earlier, and memory spontaneously rebuilds the past with the obvious facts of the present.

The resulting trial therefore does not accuse a decision: it accuses an imaginary decision, the one that would have been taken had we known what we could not know.

Why it matters

This gives decision records an unexpected function: they protect the decision less than they protect those who took it, by making contestable a reproach formulated from another state of the world.

It also changes how retrospectives and post-mortems are run: the useful question is not "why this choice", but "what was known, and by whom, at the moment it was made".

Nuances and limits

The context of the time can serve as a universal excuse: some decisions were bad with the information available at the time, and invoking the context then prevents learning anything at all.

And a written record stays partial: it records the reasons its authors judged presentable, not necessarily the ones that really weighed in the room.

Open questions

• On an old decision, how do you tell an error of judgement at the time from a correct decision that the context has made obsolete?