To conduct customer discovery interviews, start with one learning decision. Talk to people whose recent experience matches the problem and ask them to walk through what happened, what they did, and what happened next. Use those details to refine a problem hypothesis, then choose another method to test a solution. The interview is a way to learn from behavior and context, not a vote on your idea.
A polite response to a pitch is not evidence that someone will switch, pay, or change a routine. Before the conversation, write down the uncertainty you need to reduce. During it, stay with the person's actual situation. After it, separate what you heard from what you think it means.
Decide what the interview should teach you
Write one question that could change your next action. "What do small business owners want?" is too wide. A more useful learning question names a situation, a role, and a problem: "How do studio managers handle last-minute class cancellations today?" You are not testing a feature in that first question. You are checking whether the problem exists in a real workflow and how people respond to it.
Before recruiting anyone, record the assumption you are carrying. It might be that a certain kind of person encounters the issue, that it happens often enough to matter, or that the current workaround creates a real cost. Add the evidence that would make you revise that assumption. This gives the conversation direction without turning it into a script for a preferred answer.
Talk to people with direct experience
Recruit for the situation, not for encouragement. Describe the behavior or context you need to understand, then speak with people who have recently experienced it. Friends and colleagues can be useful practice partners, but their familiarity with you can make it harder to tell whether an answer reflects the problem or a desire to be supportive.
When the user, buyer, and approver are different people, be clear about which perspective you are learning from. A person who works around a problem may understand the steps. The person who pays for a solution may understand a different constraint. Do not treat one interview as a proxy for every role. If every conversation is with an enthusiastic early adopter, include a less invested user or buyer in a later round; strong interest can make the problem look more universal than it is.
Introduce the conversation plainly: "I am learning how teams handle this situation. I am not selling anything today. Could you walk me through the last time it happened?" Ask permission before recording. If someone does not want to share a detail, move on.
Customer discovery interview questions that stay grounded
Start with a recent event and follow its sequence. Open questions give the person room to describe what matters in their own terms.
| Part of the story | A useful prompt | Listen for |
|---|---|---|
| The moment | "Tell me about the last time this happened." | When it occurred and what set it off. |
| The steps | "What happened next?" | People, tools, decisions, and handoffs. |
| The workaround | "How did you handle it at the time?" | What they actually used, changed, or gave up. |
| The consequence | "What did that make harder?" | Time, money, risk, delay, or extra effort as they describe it. |
| The close | "What have I missed? Who else sees this part of the process?" | Missing context and another relevant perspective. |
Ask follow-up questions that keep the story specific: "What did you try first?" "How did you choose that option?" "What happened the next time?" If an answer turns abstract, return to the latest example. "Would you use this?" and "Would you pay for it?" ask for a prediction. Those questions may be useful in a later offer test, but they do not replace a clear account of current behavior.
Keep your solution out of the opening. A quick demo can change the conversation from discovery to feedback on your idea. If you need reactions to a prototype, label that as a separate test and note what the person actually did with it.
Write down evidence and interpretation separately
Right after the conversation, capture the event while the steps are still clear. Keep five fields:
- Context: the person's role and the situation discussed.
- Observed or reported behavior: the steps they described, in order.
- Current workaround: what they use today and where it breaks down.
- Your interpretation: the hypothesis you see, marked as your reading.
- Next question or test: what could confirm, weaken, or redirect that reading.
Do not turn a memorable comment into a conclusion about a whole market. Keep details that complicate your idea. If one person relies on a manual workaround while another accepts the inconvenience, that difference may point to distinct roles or conditions worth examining. Look for patterns across accounts, and keep the counterexamples beside them.
Harvard Business School's Customer Discovery Basics describes interviews as one of several discovery methods and cautions teams to learn about the problem rather than pitch or validate their own view. It also outlines low-fidelity testing as a way to learn more about behavior. That boundary is useful: an interview can explain context; another test may be needed to see what someone does.
Choose a next test that could change your decision
At the end of several conversations, ask what remains uncertain. If you still do not know when the problem occurs, observe or map the workflow. If you do not know whether a proposed tool is understandable, let a person try a simple mock-up on a real task. If the question is whether someone will change a process, offer a concrete next step, such as a limited pilot with a clear start date and a defined commitment.
Choose the test for the uncertainty in front of you. A count of interviews alone cannot tell you that a problem is large enough, that a solution will be adopted, or that a business can support it. A useful learning loop ends with a decision: keep the assumption, revise it, or test a different part of the problem.
A short guide for your next conversation
- Learning question: What uncertainty could change the next decision?
- Participant: Which role has direct experience with the situation?
- Opening: What recent event can they walk you through?
- Follow-up: What happened before, during, and after it?
- Evidence: What did they do, use, spend, delay, or change?
- Next test: What action could reveal more than another opinion?
If you need to make the conversation clearer, see How to Ask Questions That Move Work Forward. To carry the details into a decision, use the note structure in How to Take Meeting Notes That Turn Into Work.
Make the next move from what people do, with respect for what they tell you. For a day that moves between desk, meetings, and the floor, review the current details of the Standard Issue Zip Hoodie. Explore more practical pieces in The Self Made Journal.
0 comments