A customer problem interview should explain how a person handles a real situation today. It should not be a disguised presentation asking them to approve your idea. Prepare a short discussion guide, recruit people who fit a specific customer group and investigate recent behavior before proposing a solution.
Write the question you need answered
Start with one uncertain assumption. Perhaps independent retailers struggle to keep product information current, or office managers find small maintenance jobs difficult to coordinate. Write what you would expect to observe if that assumption were true. Also write what would contradict it.
Choose participants who have encountered the situation or participate in the buying decision. A user, approver and payer may have different concerns. Do not combine their answers as though they are interchangeable. Friends can help rehearse the interview, but a friendly rehearsal is different from evidence about your target market.
Use a simple conversation sequence
- Set expectations. Explain that you are researching a problem, how long the discussion will take and how you will use notes.
- Establish context. Ask about the person’s role and the work surrounding the problem.
- Find a recent event. Ask them to walk through the last time it happened.
- Explore the response. Ask what they tried, who became involved and what happened next.
- Check the consequence. Ask what delay, inconvenience or cost resulted, without suggesting a number.
- Close with gaps. Ask what you should have asked and whether there is another relevant perspective.
Let the participant correct your understanding. “Have I understood that the approval was easy but finding a suitable appointment was difficult?” is more useful than “So you need a booking app?” The first checks a reported fact; the second jumps to your preferred product.
Replace leading questions
Instead of “Would you pay for a faster service?”, ask “What did you do last time speed mattered?” Instead of “Is this frustrating?”, ask “What happened because the work was delayed?” Instead of “Do you like my idea?”, ask “How do you choose among the options you already have?”
You can discuss an offer later, clearly identifying the shift from research to sales. Keeping those phases distinct makes the earlier answers easier to interpret. A person may politely encourage an idea after hearing it even when the underlying problem is not important to them.
Hypothetical example: an unexpected bottleneck
A founder considering a document-formatting service interviews several small organizations. This example is illustrative. The founder expects buyers to complain about design quality. Recent-event accounts instead describe long waits for internal approval, with the actual formatting taking little time.
That finding does not prove a new approval-management business exists. It shows that better formatting may not address the primary delay. The next research step is to understand who owns approval and whether changing that process is desirable or feasible. Buying design software immediately would not resolve the newly identified uncertainty.
Keep observations and interpretations separate
Use an interview log with participant type, event date, current workaround, consequence, buying role and unresolved questions. Mark direct quotations accurately if you record them, and avoid collecting confidential documents or unnecessary personal details. Obtain permission before recording; rules and expectations vary.
Put your interpretation in a separate column. “Uses two suppliers because one is unavailable on Fridays” is an observation. “Needs a premium subscription” is an inference. Several similar stories can justify another test, but a small convenience sample cannot establish the percentage of a market with a problem.
Review for patterns and disconfirming evidence
After a few conversations, group recurring events and compare the circumstances. Include people who experience no problem or whose current solution works well. Investigate differences rather than averaging everything into a vague satisfaction score. A narrower segment may emerge from the exceptions.
Finish with a decision: continue researching, change the assumed customer, design an offer test or stop. NSF’s I-Corps program provides an institutional example of customer discovery as entrepreneurial learning; its program interview requirements are not universal targets for every small business. The practical objective here is enough reliable understanding to choose the next responsible experiment.
Interview quality checklist
- One uncertain assumption is written down.
- Participants have relevant recent experience.
- Questions ask about events rather than imagined purchases.
- Notes distinguish facts, quotes and your interpretation.
- Conflicting answers remain in the record.
- The next decision is explicit.



