How to Run Customer Interviews That Actually Tell You Something

Most customer interviews are useless.

Founders walk in hoping to validate their idea. They ask leading questions. They hear what they want to hear. They leave feeling confident. Then they build something nobody buys.

The problem is not the interview itself. The problem is how founders run them. They treat interviews like pitches instead of research. They talk more than they listen. They ask about the future instead of the past.

A good customer interview is uncomfortable. You are not there to feel validated. You are there to learn whether your assumptions are wrong before you waste months building the wrong thing.

This guide covers how to run interviews that actually tell you something useful.

TL;DR

• Interviews are for learning, not validating. Go in expecting to be wrong
• Talk to people who have the problem today, not people who might have it someday
• Ask about past behavior, not future intentions. "Would you use this?" is worthless
• Shut up and listen. If you are talking more than 20% of the time, you are doing it wrong
• Look for patterns across interviews, not permission from one enthusiastic person
• Ten good interviews will teach you more than a hundred surveys

Why Interviews Beat Surveys

Surveys give you data. Interviews give you understanding.

A survey can tell you that 60% of respondents "experience this problem." It cannot tell you why, how badly, what they have tried, or what would actually make them switch to something new.

Interviews let you follow the thread. When someone says something surprising, you can ask follow-up questions. You can dig into the emotion behind the answer. You can discover problems they did not even know they had.

Surveys are useful later, for validating patterns you have already discovered. Interviews are for discovering those patterns in the first place.

Who to Interview

Interview people who have the problem you are trying to solve. Not people who might have it. Not friends who want to be supportive. Not random people who will talk to you.

Good interview targets:

• People actively experiencing the problem today
• People who have spent money trying to solve it
• People who have tried other solutions and found them lacking
• People in your target market who match your ideal customer profile

Bad interview targets:

• Friends and family (they will tell you what you want to hear)
• People outside your target market (their feedback does not matter)
• People who "might" have the problem someday (speculation is not data)
• Anyone who has not actually experienced the pain

If you do not know who your ideal customer is, start there. Read how to find your ICP before running interviews.

How Many Interviews

For early validation, aim for 15 to 25 interviews.

The first five will feel chaotic. You are still figuring out which questions work and which do not. You will hear surprising things that make you rethink your assumptions.

By interviews ten to fifteen, patterns start to emerge. The same problems come up repeatedly. The same objections. The same workarounds.

By interview twenty, you are mostly hearing confirmation of patterns you have already identified. New interviews add less and less.

If you are still hearing completely new information at interview twenty-five, either your customer segment is too broad or you are talking to the wrong people.

The Questions That Matter

The best interview questions focus on past behavior, not future intentions.

Ask about the past:

• "Tell me about the last time you experienced this problem."
• "Walk me through what happened."
• "What did you do about it?"
• "What solutions have you tried?"
• "How much did you spend trying to solve this?"
• "What did you hate about those solutions?"
• "Why did you stop using them?"

Ask about the present:

• "How are you handling this today?"
• "What is the most frustrating part?"
• "How often does this come up?"
• "What does it cost you when this happens?"

Do not ask about the future:

• "Would you use a product that does X?" (They will say yes to be polite)
• "How much would you pay for this?" (They have no idea)
• "Would you recommend this to a friend?" (Meaningless without context)

People are terrible at predicting their own future behavior. They are reasonably good at describing what they have actually done in the past. Focus on the past.

One exception: asking what they pay today is useful. "How much do you spend on this currently?" or "What budget exists for solving this?" tells you something real. Just do not ask what they would pay for something hypothetical.

The Mom Test

There is a book called The Mom Test by Rob Fitzpatrick. The core idea is simple: ask questions that even your mom could not lie to you about.

Bad question: "Do you think this is a good idea?" Your mom will say yes because she loves you.

Good question: "When was the last time you had this problem, and what did you do about it?" Your mom cannot fake past behavior. Either she experienced it or she did not.

The goal is to ask questions where the answer is factual, not opinion. Facts are useful. Opinions are noise.

How to Structure the Interview

Keep it simple. Thirty minutes is enough.

Opening (2 minutes): Thank them for their time. Explain you are researching a problem and want to understand their experience. Tell them there are no right or wrong answers.

Context (5 minutes): Understand their situation. What is their role? What does their day look like? This helps you understand the rest of their answers.

Problem exploration (15 minutes): This is the core. Ask about their experience with the problem. Go deep. Follow surprising answers with "tell me more about that."

Current solutions (5 minutes): What are they doing today? What have they tried? What worked and what did not?

Wrap up (3 minutes): Ask if there is anything else they think you should know. Ask if they know anyone else you should talk to. Thank them again.

The Skill: Shutting Up

The hardest part of customer interviews is not talking.

Founders want to explain their idea. They want to pitch. They want to see the light bulb go on in the customer's eyes. That impulse will ruin your interviews.

Your job is to listen. Aim to talk less than they do. A lot less. If you find yourself explaining, pitching, or filling silences, you are doing it wrong.

When there is silence, wait. Do not fill it. Often the most valuable things come out after a pause, when the interviewee has time to think.

When they say something interesting, resist the urge to respond with your solution. Instead, say "tell me more about that" or "why do you think that is?"

You are a detective, not a salesperson.

What to Do With the Answers

After each interview, write down:

• Key quotes (their exact words)
• Surprising things you learned
• Patterns that confirm or contradict your assumptions
• Questions you should have asked but did not

After every five interviews, step back and look for patterns:

• What problems come up repeatedly?
• What language do people use to describe the problem?
• What solutions have they tried and why did those fail?
• How much pain is this actually causing?

The patterns matter more than any individual interview. One enthusiastic person does not validate your idea. Ten people describing the same problem in the same way starts to mean something.

Red Flags

They cannot describe a specific instance. If someone says "yeah, that happens sometimes" but cannot tell you about a specific time it happened, the problem is not real enough.

They have never tried to solve it. If the problem exists but they have never spent time, money, or effort trying to fix it, it is not painful enough.

They keep changing the subject. If you ask about problem A and they keep talking about problem B, maybe problem B is what actually matters.

They say yes to everything. Some people are agreeable. They will nod along to whatever you say. Their feedback is useless. Look for people who push back, disagree, or say "that is not really how it works."

Common Mistakes

Pitching instead of listening. You are not there to convince them. You are there to learn.

Asking leading questions. "Do you hate how slow your current tool is?" leads them to agree. "How do you feel about your current tool?" lets them tell you the truth.

Talking to the wrong people. Ten interviews with people outside your target market teach you nothing useful.

Stopping too early. Three interviews is not enough. You need enough to see patterns.

Only hearing what you want to hear. Confirmation bias is real. Actively look for evidence that contradicts your assumptions.

Where to Go From Here

Need to identify who to interview? Read how to find your ICP

Want a structured validation process? Read validate your startup idea in 14 days

Looking for signs your idea is working? Read signs your startup idea has legs

Closing Thought

Customer interviews are not about feeling good. They are about learning fast.

The founders who succeed are the ones who discover their wrong assumptions early, before they have spent six months building the wrong product. Interviews are the cheapest, fastest way to do that.

Go in expecting to be wrong. Ask about the past, not the future. Listen more than you talk. Look for patterns, not permission.

Ten honest interviews will teach you more than any amount of market research, competitor analysis, or brainstorming. They will also be uncomfortable, because the truth often is.

Do them anyway. Your company depends on it.