Use it for
- The coded data or the themes from your analysis, with the participants behind each theme.
- Quotes and clips for each theme, with participant IDs.
- The research questions and the decisions that the study must support.
Steps
- Start from the decisions
List the decisions the readers must make. Put the findings that affect those decisions first. Move other findings to an appendix, even if they are interesting.
- Separate findings from insights
Write each finding as what people did or said. Write each insight as why it happens. Keep them as two sentences, so readers can challenge the insight without losing the finding.
- Attach evidence to each claim
Each finding needs at least two quotes or clips from different participants. Add the number of participants who fit, and the number who said the opposite.
- Rate confidence
Give each finding a confidence level: high, medium or low. Base it on how many participants support it, how consistent they are and how directly the data shows it.
- Write implications and recommendations
For each insight, write what it means for the product. Then write a recommendation that a team can act on. Keep the recommendation apart, so readers can accept the insight and reject the action.
- State the limits
Write who was not in the sample, what the method cannot show and which questions are still open. A clear limit makes readers trust the rest of the report more.
Codebook
Copy the codebook as a table. It pastes into a spreadsheet with one row for each code.
Research question
The question the study set out to answer, tied to a decision.
- Include
- Write the question as it was set before the study, and name the decision it supports.
- Exclude
- Do not rewrite the question to match what you found. If the study answered a different question, say so.
- Example
- “Why do trial users who connect a bank account still not upgrade? Decision: what to change in the trial.”
Finding
A pattern in what participants did or said, without interpretation.
- Include
- Write a statement that a reader could verify by watching the sessions, with the number of participants.
- Exclude
- Do not explain why the pattern happens. Put the why in Insight.
- Example
- “Most participants who connected a bank account did not open the forecast page during the trial.”
Insight
An explanation of why a finding happens, based on the evidence.
- Include
- Write the cause that the participants' stories point to, and link it to the finding.
- Exclude
- Do not write what the team should do. Put that in Recommendation.
- Example
- “Participants saw the bank connection as the end of setup, so they waited for the product to tell them what to do next.”
Evidence
The quotes, clips and counts that support a finding or an insight.
- Include
- Add 2 or 3 quotes or clips from different participants, with IDs. Add the count of participants who fit and who did not.
- Exclude
- Do not use one quote as the only support. Do not edit a quote to make it fit the claim.
- Example
- “P3: I linked my bank and then I thought, OK, now it will send me something.”
Confidence level
How sure the team is about a finding, based on the strength of the evidence.
- Include
- Use high when many participants show it directly and none contradict it. Use medium or low when support is thin or mixed.
- Exclude
- Do not raise confidence because the finding matches what the team believed before.
- Example
- “Medium: more than half showed it, but two participants explored the forecast without a prompt.”
Implication
What the insight means for the product, the business or the next study.
- Include
- Write the consequence for the decision named in the research question.
- Exclude
- Do not write a specific solution. Put that in Recommendation.
- Example
- “The trial ends before many users see the feature that would make them pay.”
Recommendation
A specific action that a named team can take, based on an implication.
- Include
- Write one action, the team that owns it and how to measure if it worked.
- Exclude
- Do not write a vague direction such as "improve onboarding". Name what to change.
- Example
- “Growth team: open the forecast page right after the bank connection. Measure forecast views in the trial.”
Limitation
A limit of the study that affects how far readers can trust or apply the findings.
- Include
- Write who was missing from the sample, what the method could not show and which questions are open.
- Exclude
- Do not list general limits of qualitative research. Name the limits of this study.
- Example
- “All participants were on the monthly plan. Annual trial users may behave differently.”
Report outline
- Summary: the 3 to 5 most important insights and recommendations on one page.
- Research questions, decisions and method: who, how many, when and how.
- Findings with insights, evidence and a confidence level, in order of impact on the decisions.
- Implications and recommendations, each with an owner.
- Limitations and open questions.
- Appendix: other findings, the codebook and the full participant list.
Mistakes to avoid
- Do not order findings by how often they came up. Order them by how much they affect the decisions readers must make.
- Do not merge findings and insights into one sentence. If the insight is wrong, readers will throw away the finding too.
- Do not hide disagreement. Report the participants who said the opposite. It shows where the finding stops.
- Do not write percentages for small samples. "7 of 12" is honest. "58%" suggests a precision that the data does not have.
Guides that use it