Interview Guide
A structured set of questions anchored in past behaviour rather than in hypothetical intention.
Definition
An interview guide keeps sessions comparable without making them a script. It should be a set of areas with opening questions, not a questionnaire read aloud.
Craft convention. True because the industry converged on it. Breaking it costs familiarity, not correctness.
On this plate
Fill it in
Loading your draft…
Saved in this browser only. Nothing is sent anywhere.
Opening
Purpose, duration, recording, right to stop, and that there are no wrong answers.
Establish context and get them talking about their actual work.
Core questions
Anchor every question in something that actually happened. Prefer 'tell me about the last time' to 'do you usually'.
Get a specific story with details, not a generalisation.
Ask what they do outside the tool. Spreadsheets and notes are unmet requirements.
Follow frustration. It marks real problems rather than stated preferences.
For adoption questions, ask what they considered and why they chose what they chose.
Closing
Leave room for what you did not think to ask.
Write down the leading questions you are tempted to ask, so you notice yourself asking them.
Rewriting the question you wanted to ask
Hypothetical questions produce confident answers that predict nothing. Each of these swaps intention for recalled behaviour.
| Instead of | Ask | Why |
|---|---|---|
| Would you use a feature that did this? | Tell me about the last time you needed to do this. | People are poor forecasters of their own behaviour and generous to the person asking. |
| How often do you do this? | How many times did you do it last week? Walk me through the most recent one. | Frequency estimates are reconstructed. A recent instance is remembered. |
| Do you find this confusing? | Show me what you did here. What were you expecting? | The first invites agreement. The second produces an observation. |
| What features would you want? | What did you do when the product could not do it? | The workaround is the requirement, described accurately. |
| Do you like this? | What would you do next? | Preference is cheap. Intended action is informative. |
Running the session
Do
- Treat the guide as areas to cover, not a script to read in order.
- Follow the interesting answer, then return. The unplanned branch is usually where the finding is.
- Ask for the last time, then for the story, then for the parts they skipped.
- Leave silence after an answer. The second thing people say is usually more honest than the first.
Do not
- Do not explain your product mid-interview. The session stops being research the moment you start selling.
- Do not ask two questions in one sentence, which lets the participant answer the easier half.
- Do not correct a participant's understanding until the end.
- Do not schedule so many questions that there is no room to follow anything up.
Local chart
TPL·002 · Region VIII · Grid G4
This entry and the entries it is stated to connect with. Each line is labelled with what the connection means.
Continue from here
Each link says what the connection is, so you can tell a principle from an alternative from a thing people mix this up with.
Often used with
These usually appear in the same screen or the same decision.
- Confirmation BiasBIAS·009 · Cognitive bias
- Contextual InquiryRES·012 · Research method
- Jobs to Be DoneRES·018 · Research method
- User InterviewsRES·003 · Research method
- Research PlanTPL·001 · Template
Nearby in the Atlas
Adjacent territory, found through shared tags rather than a stated link.