Main idea

The real triggers are known: a big customer complains, a demo breaks, support gets overwhelmed, a regression hits a critical flow, an executive stumbles onto one. None of them measures what the defect has cost since it appeared.

An expensive, discreet defect — three minutes lost each day by a hundred users who no longer report it — waits. A minor defect encountered by a member of the executive committee goes first. The order of treatment follows the defect's social exposure, not its cumulative impact.

Why it matters

This explains why the fix always arrives at the worst moment. It is triggered at peak pressure, when a customer is already unhappy and the subject has become political, whereas the same decision could have been taken calmly six months earlier.

It also shifts the diagnosis: a team that handles "the right subjects" under emergency doesn't have a prioritization problem, it has a trigger problem.

Nuances and limits

Visibility isn't a bad indicator, it is a partial one. What comes back often often hurts; what never comes back is sometimes invisible because users have given up.

Open questions

• Which early signal can replace the important customer's complaint as a trigger?