Your first five customer interviews: questions that change what you build.

1. Choose one person and one uncertainty.

Write the assumption that would change your next decision if it proved wrong. For example: first-time founders struggle to turn their mission into a concrete first test. That is a hypothesis to investigate. “People will love my planning app” cannot guide the conversation because it starts with your preferred answer.

Invite people who have recently faced the situation. Ask for twenty minutes to understand their experience, explain that you are learning rather than selling, and give them an easy way to decline. A participant should never feel that access to help, a community, or a personal relationship depends on agreeing with your idea. Ask before recording; plain notes are enough for this first round.

• Name the situation, role, and recent event that make someone relevant.
• Write the assumption you most need to test.
• Recruit beyond the people who already want to encourage you.
• Agree on time, note-taking, and how their words may be used.

2. Ask for a recent story, then follow the details.

Begin with a specific event: “Tell me about the last time you tried to turn an idea into something someone could use.” Let them describe the sequence. When an answer becomes general, ask for an example. When a word such as frustrating, expensive, or confusing appears, ask what happened that made it so.

The following questions are a menu, not a script you must finish. Spend more time on an actual difficulty than on checking every box. Leave room for silence. You are trying to understand their experience in their vocabulary, including the parts that contradict your expectations.

• What were you trying to get done that day?
• What triggered the need to do it?
• Walk me through what you did, from the first step.
• Where did you get stuck, stop, or ask someone for help?
• What did you try instead, and how did that work?
• What did this cost you in time, money, or another tradeoff?
• How did you decide whether the result was good enough?
• Is there something I should have asked that I missed?

3. Keep your pitch out of the evidence.

“Would you use an app that makes this easier?” encourages a prediction that costs the person nothing. “How did you handle this last week?” asks for something that already happened. Both answers may be sincere, but only one gives you a concrete behaviour to examine. Questions about hypothetical purchases belong in a later test with an actual offer and a real choice.

If you show a prototype, mark that as a separate part of the conversation. Give the person a realistic task and observe what they do before explaining how it works. Keep prototype findings separate from the earlier account of their problem. Your explanation can make an experience look clearer than it would be for somebody arriving alone.

• Replace “Do you like this?” with “What did you expect to happen here?”
• Replace “Would this help?” with “What do you currently do when this happens?”
• Replace “Is this easy?” with “Show me how you would start.”
• Write down where you prompted or explained; that assistance affects the result.

4. Write an evidence note after each conversation.

Record the situation while it is fresh. Use an anonymous participant code, the date, a recent example, the current workaround, and the cost of the problem. Keep a direct quote separate from your interpretation. “I abandoned it after choosing a name” is their reported behaviour. “They need a brand generator” is your interpretation, and another solution could fit the same evidence.

The example below is hypothetical: Participant P03 described spending two evenings designing a logo, then stopped because they could not explain what the project would provide. The existing workaround was asking a friend for feedback. The next uncertainty is whether a one-page promise worksheet helps them choose a first delivery. That is specific enough to test without claiming that three other people have the same problem.

• Participant code and relevant context.
• A recent event and the sequence they described.
• Existing workaround and the tradeoff it required.
• Their words, then your separate interpretation.
• Evidence that challenged the original assumption.
• One uncertainty to investigate next.

5. Turn the round into one decision and one test.

Compare the records after the round. Look for a repeated situation with a concrete cost, existing effort to solve it, and a next step you can deliver. Keep the exceptions visible. Five interviews can expose an assumption or suggest a pattern; they cannot establish its prevalence across a market. If the pattern is still vague, narrow the group or improve the questions before adding features.

Choose a small delivery that addresses the strongest problem you heard. Define what completion looks like, invite a person to try it, and observe use and return. For the hypothetical worksheet, the first test might be whether the person can state one promise and schedule one delivery after completing it. That tests an immediate result. It does not establish a durable outcome or an impact count.

Write the decision in this form: “We heard [specific problem] in [relevant context]. We will test [complete promise] with [person or group]. We will observe [behaviour] by [date]. The result will tell us whether to [continue, change, or stop].” Attach what would disconfirm the idea. A round that changes what you build has earned its time.

• List repeated behaviours and meaningful exceptions.
• Name the assumption that survived and the one that changed.
• Choose one complete delivery and one observable result.
• Set a review date and the result that would change your decision.