Main idea

Most documents gain by being complete. A product decision register does the opposite: its ability to be of use depends on what can be held in mind or found in a few seconds when a question comes back. Fifteen entries fit in the head of a Product Manager, a support lead and a sales director. Fifty a quarter fit nowhere, and the register joins the ranks of the documentation an organization produces without reading.

Rarity is therefore not a discipline of writing, it is the condition of use. A few decisions a year — five, ten, twenty depending on the size of the organization — are enough to cover the genuinely cross-cutting questions.

The admission criterion that produces this rarity is single and operational: the decision must prevent future re-discussions. Anything local, one-off, tied to a single feature, already covered by an existing rule, without lasting consequence or too obvious to be contested fails that test. Documenting every local arbitration does not merely dilute the register, it turns it into paperwork — and that is where the exercise's reputation for heaviness comes from.

Why it matters

This overturns the reflex of exhaustiveness that accompanies the setting up of any documentary system, and gives a non-moral reason to refuse an entry: it would cost the other entries.

It also supplies an explanation for the usual verdict — "we had put that in place, nobody used it" — that blames neither the tool nor the teams' willingness.

Nuances and limits

The threshold depends on the organization: a multi-product group can sustain several dozen cross-cutting rules without each losing its legibility, provided they are partitioned by scope.

And dilution is not the only mode of failure: a register of three entries can be useless if the questions that really come back are not in it.

Open questions

• What should be done with a decision that has stopped being of use: take it out of the register, at the risk of losing the record, or keep it, at the risk of diluting what remains?