Name the decision before the method
“We need research” is too broad to guide useful work. Write the decision: should we change the first task, simplify a workflow, introduce a reminder, or stop building a feature? Identify who will make that decision and what they currently believe.
Then ask what evidence could change their mind. If no plausible finding could affect the decision, be honest about whether the work is discovery, validation of a narrow detail, or simply a request for supporting material.
Separate what you know from what you assume
Create a short ledger of verified observations, reported claims, working hypotheses, and missing information. Include where each item came from and when it was collected. A support theme, an event log, and an interview answer reveal different aspects of the situation.
Do not let one source impersonate another. A clickstream can show a sequence of recorded events. It cannot, on its own, tell you what someone believed. An interview can explore a person’s account, but it does not establish how often the same behavior occurs across every customer.
Better research does not just produce answers. It changes a decision.
Choose a method that can address the uncertainty
When the uncertainty concerns whether people can complete a specific task, a task-based usability exercise may be appropriate. When it concerns a past experience and its context, an interview may help. When it concerns the prevalence of a recorded behavior, inspect the relevant data and its coverage.
This is a planning guide, not a substitute for a research specialist’s judgment. Consider participant selection, accessibility, consent, privacy, and the stakes of making an incorrect decision. Write down the limitations before reporting findings.
Ask about a real episode
Instead of “Would you use this feature?”, explore a recent occasion when the need arose. What happened first? What did the person try? What information did they need? What made them stop? Ask neutral follow-ups and avoid explaining the answer you hoped to hear.
Record behavior and interpretation separately. Preserve contradictory examples. When participants struggle, do not immediately teach them the interface and then count the task as independently completed.
End with a decision and an evidence trail
Your summary should connect the original decision, the findings relevant to it, the limitations, and the recommended next step. Give the decision owner a way to inspect the supporting material. Include unresolved questions instead of hiding them behind a polished presentation.
A useful outcome may be a small design change, a better-defined experiment, a request for additional evidence, or a decision not to proceed. Research does not have to recommend more features to be valuable.
Use the Research-to-Action Brief to make the handoff explicit: this is what we observed, this is what we think it means, and this is the next action we can justify. Keep each statement in its proper category.
