> ## Content Index
> Fetch the complete content index at: https://agustin-gutierrez.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Answer Can Improve While the Question Stays the Same
- URL: https://agustin-gutierrez.ghost.io/the-answer-can-improve-while-the-question-stays-the-same/
- Published: 2026-09-07T17:32:08.000Z
- Updated: 2026-09-07T17:32:08.000Z
- Description: A decision can get better evidence every year and still be organized around the same untested question. Better measurement can sharpen the answer without improving the frame.
- Author: Agustin Gutierrez Bazaco
- Tags: Decision Intelligence

*Better evidence sharpens an answer. It does not, on its own, test the question behind it.*

**Working proposition**

> A decision frame determines what the organization is trying to optimize before the evidence is evaluated. Better evidence can sharpen the answer inside that frame without testing whether the frame itself is still adequate for the decision.

A question, once framed inside an organization, tends to stay framed. Someone decides what the decision is optimizing, which alternatives are on the table, what unit is being decided about — and from that point on, almost everything that happens next happens inside that frame. The measurement gets better. The dashboard gets sharper. The people in the room get more experienced at answering exactly this question. None of that requires the frame to be re-examined, because none of it touches the frame. It only touches the answer.

That is a different failure from the other two this chapter has already covered. The evidence can arrive on time. The person looking at it can be the person with the authority to act. And the decision can still be organized around a unit nobody has re-tested in years, because nobody in that sequence — not the analyst, not the dashboard, not the decision-maker — was ever asked to test the frame itself.

Take a category review built around SKU velocity: which products earn their shelf space, ranked by how fast they sell. Assume the measurement is excellent — full transaction history, no gaps, real time. Assume the representation is excellent too — every SKU's performance is visible to whoever is making the call. Over a few cycles, the retailer adds more to the scorecard: basket contribution, halo effect on adjacent categories, destination role, margin. The picture gets richer every quarter.

The decision does not necessarily get better because of that. It gets more confident. The question being answered — which individual SKUs are worth the space? — has not moved. Even a much richer scorecard still evaluates each SKU as an individual unit to be ranked. It can tell the retailer more about each SKU without ever asking whether individual SKU productivity should be the organizing logic of the shelf in the first place.

Reframing does not look like a new metric. It looks like a different question: should this shelf be organized around individual SKU productivity at all, or around the shopping missions it is meant to serve? That is not an improvement to the existing decision. It is a different decision, with a different unit, that the richer scorecard was never built to ask.

This is easy to miss precisely because nothing looks broken. No one misread the data. The organization got better at answering its question, and getting better at answering a question feels like progress on the decision itself. It is progress — inside the frame. Whether the frame is still adequate for the decision is a separate question that better evidence does not, on its own, go anywhere near. And sometimes the better the evidence gets, the less reason anyone feels to ask it: a sharper ranking is a more convincing ranking, and a more convincing ranking is easier to mistake for a correct decision. Measurement improving does not necessarily make the framing problem smaller. It can make it quieter.

This sits underneath the other two ways a decision's inputs can fail. Whether the evidence is any good is a measurement problem. Whether what reaches the decision-maker still resembles the reality it claims to represent is a representation problem. Whether the decision being optimized is the decision worth making is a framing problem — and it is untouched by how well either of the other two gets solved.

This essay is part of the In-Store Retail Media Framework under **Decision Intelligence**. New readers can begin with [Start Here](https://agustin-gutierrez.ghost.io/start-here/) or explore the [Principles](https://agustin-gutierrez.ghost.io/principles/) that guide the Journal.