A research report lands in the shared drive. It is thorough, well-structured and full of quotes. It has a persona, a journey map and seven themes. Three months later, nobody can remember what it said, and the roadmap looks exactly as it did before.

This is not a research failure. It is a decision failure, visible through research.

The decision test

The most useful question to ask before any research is: what decision will this inform?

Not “what will we learn?” Learning is good, but it is not a plan. The decision might be whether to build a particular workflow, which segment to target, how to price something, whether a problem is frequent enough to justify investment. If no decision depends on the answer, the research is likely to become content rather than evidence.

A good brief names the decision, the options, and what answer would change the team’s choice. It also names who will act on it and when. Without that, the research is a broadcast. With it, the research is a tool.

Symptoms of decorative research

The findings confirm the roadmap. Every study seems to validate what the team already intended. This is a sign that the questions were shaped to protect the conclusion.

The report is the deliverable. The team measures success by whether the report was completed, not by whether anything changed.

“We need more research.” Used as a way to delay a decision that is actually blocked by disagreement, not by missing information.

The handoff. Research is done by a specialist, summarised, and delivered to people who were not involved. By the time it arrives, the questions have moved.

The persona. A composite character with a name and a stock photo, standing in for a set of behaviours nobody has observed directly.

Match the method to the decision

Research methods are not a hierarchy. They are tools with different costs and different strengths. The right method is the cheapest one that produces a decision-grade answer.

If the decision is whether people understand a label, five quick tests may be enough. If the decision is whether a workflow exists in a market, you may need a broader scan and a different kind of evidence. If the decision is whether to enter a segment, desk research and a handful of expert conversations may be more useful than another round of usability sessions.

The common mistake is to default to a familiar method, usually interviews, regardless of the question. Interviews are excellent for understanding context and language. They are poor at estimating frequency, measuring behaviour or predicting adoption.

Synthesis is an argument

A good synthesis does not just summarise; it makes a claim. It identifies the tension, names the pattern, and explains what it means for the decision.

The weakest synthesis is a list of themes with quotes attached. The strongest is a short argument: we found this pattern, it contradicts our assumption, and here is what we should do differently.

That argument should be falsifiable. If no conceivable finding would have changed the conclusion, the research was not testing anything.

Research is useful when it makes a decision cheaper to make and harder to fake.

Keep it close to the team

The further research travels from the people who will act on it, the more context it loses. A report that crosses three handoffs is a summary of a summary.

The alternative is not to eliminate specialists. It is to keep the team in contact with the raw material. A short clip from an interview. A support thread with the customer’s own words. A session recording watched together. A visiting customer in a team meeting.

Those moments are not a substitute for rigorous research. They are what makes the rigorous research land.

Know the limits

Research describes the present. It can tell you how people behave now, what language they use, where the friction is, and what they have tried. It cannot tell you whether a market will exist in three years, whether a new behaviour will become a habit, or whether a competitor will change the terms.

Those are judgement calls, and they should be labelled as such. Pretending research can answer them produces false confidence; refusing to decide without it produces paralysis.

The useful posture is to be explicit about which questions are empirical and which are bets, and to fund them differently.

The habit

The teams that get the most from research treat it as a habit rather than an event: small, frequent, close to the decision, with a clear owner and a visible consequence.

They also accept that some research will be inconclusive. That is not a failure. The failure is research that was never connected to anything, produced on schedule, and quietly ignored.