Choose the question you need answered
Decide whether you are investigating the problem, the buyer or the buying process. Interview people who encounter the relevant situation, not only friends who want to encourage you. Tell participants why you are asking and get permission before recording. Store notes responsibly and avoid collecting personal details unrelated to the research.
Ask about the last time, not an imagined future
'Would you use an easier tool?' invites agreement. 'Tell me about the last time this happened' reveals a sequence of actions. Ask what triggered the problem, which workaround they used, how long it took and what they did afterward. Follow up on contradictions gently; people may describe an ideal process before remembering what actually happened.
- What happened the last time you had to do this?
- What did you try first, and why?
- Who else became involved?
- What did the workaround cost in money, time or missed work?
- How was the final decision made?
Worked example: appointment reminders
A hypothetical founder interviews salon owners about missed appointments. Instead of pitching a reminder product, they ask how the last missed booking was handled. One owner describes manually calling clients; another says deposits already address the problem. Those differences reveal that the same complaint may have different causes and different willingness to change.
Separate evidence from interpretation
Keep a participant's direct statement separate from your conclusion. 'Used a spreadsheet last Tuesday' is evidence; 'needs an app' is an interpretation. Record unknowns explicitly. Look for repeated patterns across interviews but do not describe a handful of conversations as a representative market survey.
Decide what to test next
Turn findings into a focused experiment: a clearer service offer, a manual pilot or a different customer segment. Define what would count as a meaningful commitment, such as sharing necessary inputs or agreeing to a paid pilot, rather than treating praise as validation. Continue interviewing when evidence conflicts instead of selecting only the answers that support the idea.