Prepare permitted notes with participant aliases and source positions, extract observations, and keep interpretations and unanswered questions separate. Review findings against complete passages before using them in product or content decisions.
In this guide
What evidence supports the export finding?
The participant labels and observations are authored examples. Keep the source-note positions and approved research permissions with real interviews. The contrast row changes the interpretation: an export-format option may be worth testing before a performance feature. Describe that as a hypothesis until the actual exported fields and receiving system have been inspected.
| Observation | Context | Provisional finding / next check |
|---|---|---|
| P03 asks for a faster export | Reformats dates after each download for destination A | Check whether destination A needs another date format |
| P04 reports no export problem | Uses destination B | Do not generalize P03's friction to all users |
| Feature idea: faster download | No download timing supplied | Measure transfer time separately before prioritizing |
Preserve the circumstances of the observation
Record the research question, task shown or discussed, interview date, participant alias, and relevant context you have permission to retain. Keep personal identifiers out of an AI working copy when they are unnecessary. A participant recalling a past experience and a participant demonstrating a task provide different kinds of evidence.
Use source positions to find the original recording or note. If a speaker's identity or a number is uncertain, resolve that before assigning a firm finding to them. The recording-to-minutes workflow can help prepare checked text, while interview analysis needs its own research question and interpretation review.
- Assign an ID to each relevant observation and preserve its source passage or recording position.
- Describe the observed behavior or reported experience in plain language, with the task context beside it.
- Add a separate interpretation field explaining what the observation might mean and what alternative explanation remains possible.
- Group observations under tentative findings and inspect evidence that conflicts with each one.
- Review the full surrounding passages, then write a short finding with its supporting IDs, limits, and proposed follow-up.
Check that the conclusion has not outgrown the sample
A small purposeful interview sample can reveal useful problems without estimating how common they are across a market. State who was included and why. Keep a count of participants distinct from a count of supporting passages, since one talkative participant can contribute many passages.
Review the finding with someone who can inspect the evidence. Ask whether another explanation fits and which additional observation would change the conclusion. This is a stronger basis for a next task than adding a numerical confidence score that has no defined method.
Request a traceable findings draft
Organize these permitted interview notes for a research review. RESEARCH QUESTION: [state] PARTICIPANT ALIASES AND RELEVANT CONTEXT: [state] NOTES WITH SOURCE IDS/POSITIONS: [paste] Separate observation, interpretation, feature suggestion, and open question. Propose findings supported by specific observation IDs. Preserve counterexamples and different task contexts. Identify passages that need replay or clarification. Do not invent quotes, emotions, participant traits, prevalence, or causal explanations. Return a findings table and follow-up research questions.
Carry the research into a bounded next step
Hand over a specific investigation or improvement question with the evidence IDs attached. If the problem involves instructions, give the writer the relevant passages and the desired reader outcome rather than an unqualified demand to rewrite everything.