Main idea

Filing an interview verbatim under "customer feedback", tagging a support ticket, classifying a sales call report: these gestures make the material findable and leave it in exactly the state it arrived in. Nothing has been understood after that wasn't understood before.

Structuring is another gesture, and it is described as a sequence of transformations: a request becomes a signal, the signal is connected to the situation it appeared in, the situation reveals a stake, and the stake is tied to a segment, a constraint, a piece of evidence or an alternative. At each step, the information changes nature — it stops being what a customer said and becomes something you can reason about.

So classification bears on location, structuring on the status of the information. The two are easily confused because they produce the same impression of order.

Why it matters

This gives a test to an organization that believes it has capitalized on its customer feedback: ask which stake and which segment a specific request is tied to. If the answer is the folder it sits in, what happened was tidying.

It also directs the tooling effort. A better filing plan brings you no closer to any decision; what brings you closer to a decision is the requalification work, which remains a work of interpretation.

Nuances and limits

Tidying isn't useless: without it, the raw material you need to be able to reopen in order to reinterpret becomes impossible to find. It is insufficient, not superfluous.

And the chain request → signal → situation → stake assumes you already know which stakes exist. On a new market, requalification mostly produces hypotheses, which have to be owned as such.

Open questions

• At what point does requalifying a request become premature, for lack of enough cases for the identified stake to be anything other than a generalization from a single customer?