What makes a good customer interview question?
A good question gets the participant to tell you what happened, not what they think might happen. People are poor at predicting their own behavior and generous with compliments, so ask about the last time, the real situation and the actual decision.
- Open: begins with what, how, when or tell me about.
- Specific: points to a recent event, not a general habit.
- Neutral: does not signal the answer you hope to hear.
- Single: asks one thing at a time.
- Actionable: you know what you would do with each answer.
When you are done choosing, put the questions into a plan using the discussion guide template.
What are good discovery questions?
Use these when you are learning about a problem before building or changing anything.
- Tell me about the last time you [did the task]. Walk me through it.
- What were you trying to get done?
- What did you use, and why that?
- Where did it go wrong or take longer than you wanted?
- What did you do about it?
- What have you tried to fix it?
- How much does it matter to you, and how do you know?
- Who else is involved, and what do they need?
The question about what they tried is valuable. If nobody has tried anything, the problem may not hurt enough to solve.
For research teams working this way, the customer research tools page shows how interviews fit a product cycle.
What should I ask new customers?
Ask within the first weeks, when the memory is fresh.
- What made you sign up now rather than last month?
- What were you hoping it would do for you?
- What did you do in your first session?
- Was there a moment you felt stuck? What happened?
- What did you expect to find that you could not find?
- What is the first result you want to see?
- What almost stopped you from getting started?
These questions show whether the promise on your site matches the first experience. Gaps between what customers hoped for and what they found are your clearest onboarding fixes.
What are good churn interview questions?
Ask people who canceled or stopped using the product. Keep the tone friendly and curious, because they owe you nothing.
- What were you using it for?
- When did you first think about stopping?
- What changed around that time?
- What did you switch to, if anything?
- What did the alternative do better?
- Was there anything we could have done that would have kept you?
- What did you like, even if it was not enough?
Do not defend the product during the interview. If you argue, you teach customers to stop being candid. Hundreds of past customers are hard to schedule, which is where interviews run through a link help. See the page on AI moderated interviews.
What should I ask after a sale is won or lost?
Ask the person who made or influenced the decision.
- What problem were you trying to solve?
- How did you start looking for options?
- What did you compare, and how?
- What mattered most in the end?
- What almost changed your mind?
- Who else had a say?
- What would have made the other choice better?
Win interviews are often skipped, but they tell you which promises really land. Ask what customers would say to a colleague, in their words, and use it in your messaging.
How do I ask about a new idea without leading?
Describe the problem first, then show the idea, then ask for reactions.
- Before I show anything: how do you handle this today?
- In your own words, what do you think this does?
- Who would use this at your company?
- What would stop you from using it?
- What would you do instead if this did not exist?
- What is missing?
- When would this be useful, and when not?
Avoid "Do you like it?" and "Would you buy this?" Both invite politeness. Better are questions that reveal context, such as "When did you last need something like this?" For a wider view on interview formats, read about moderated vs unmoderated testing.
What follow-ups work in almost any interview?
A short probe often matters more than the main question. These keep the person talking about what actually happened.
- Can you say more about that?
- What do you mean by that?
- Can you give me an example?
- What happened next?
- Why was that important?
- How did you feel at that point?
- Is that typical, or was this time different?
Silence helps too. Wait a few seconds after an answer. People often add the real reason when nobody rushes them. An AI moderator is built to probe short answers in the same way, asking for an example when the response is vague.
Which questions should I avoid?
- Leading: "Don't you think the new design is easier?" Ask how they found it.
- Hypothetical: "Would you pay for this?" Ask what they pay for today.
- Double: "How do you find and share reports?" Split it into two.
- Jargon: use the customer's words.
- Opinion collecting: "What do you think of our brand?" rarely leads to a decision.
- Pitching: if the conversation turns into a demo, you have stopped learning.
Review each question against the decision it supports. If an answer would not change your plan, remove it.
How many people should I ask?
For a clear question, a handful of interviews usually shows the main themes, and more interviews show how common each one is. The right number depends on how varied your customers are. Our guide on how many user interviews walks through how to choose.