The most expensive mistake in product development is building something nobody wants. Customer interviews are the cheapest way to avoid that mistake. But most interviews fail to reveal real needs because the interviewer asks leading questions, accepts surface-level answers, or turns the conversation into a pitch for their existing solution. This guide covers how to run interviews that actually teach you something.

The five-phase interview structure builds depth gradually from rapport to reflection

When to Interview (and When Not To)

Interviews work best during the Empathize stage when you are trying to understand the problem space. They are also valuable during the Test stage when you are evaluating prototypes. They are less useful for validating ideas you have already committed to building (that is what usability testing and analytics are for) or for measuring satisfaction across large populations (that is what surveys are for).

If you need breadth (understanding patterns across many users), use surveys. If you need depth (understanding why one person does what they do), interview. Most design projects need both, but at different stages.

Recruiting the Right People

Interviewing the wrong people is worse than not interviewing at all because it gives you false confidence. You need people who match your target user profile, not just people who are easy to reach.

Avoid friendlies. Your coworkers, your friends, and your existing power users will tell you what you want to hear. Seek out people who have no relationship with you or your product.
Include non-users. People who tried your product and stopped, or who chose a competitor, often provide the most valuable insights. Their perspective is unfiltered by familiarity.
Recruit for the behavior, not the demographic. You do not need "women aged 25 to 34." You need "people who have searched for a design thinking tool in the last 3 months." The behavior is what matters.

Five to eight interviews per persona segment is usually enough to identify major patterns. After about the sixth interview, you will notice the same themes repeating. That is when you know you have sufficient coverage.

Preparing Your Interview Guide

An interview guide is not a script. It is a list of topics you want to cover, with example questions under each topic. You should know the guide well enough that you rarely look at it during the conversation. The conversation should feel natural, not like a questionnaire.

Structure your guide in three phases:

Phase 1: Context (5 minutes)

Start with questions about the person's world, not about your product. "Tell me about your role. What does a typical week look like?" This builds rapport and gives you context for interpreting everything that follows.

Phase 2: Deep dive (20 to 25 minutes)

Focus on the specific problem area. Use storytelling prompts: "Walk me through the last time you tried to [do the thing your product addresses]. Start from the very beginning." Then follow up on whatever is most interesting or surprising.

Phase 3: Reflection (5 minutes)

Ask them to step back and reflect. "If you could wave a magic wand and change one thing about how you [do this task], what would it be?" Then: "Is there anything I should have asked but did not?"

The Art of Asking Good Questions

Good interview questions share three properties: they are open-ended, they are about the past (not the hypothetical future), and they invite stories rather than opinions.

Ask about behavior, not preferences

Bad: "Would you use a feature that does X?"
Good: "Tell me about the last time you needed to do X. What did you do?"

Ask about the past, not the future

Bad: "How often would you use this?"
Good: "How many times did you do this in the past month?"

Ask for specifics, not generalities

Bad: "What problems do you have with your current tool?"
Good: "Tell me about the last time your current tool frustrated you. What happened?"

Listening Techniques

Most interviewers talk too much. The ideal ratio is 20% you, 80% them. Here are techniques to stay in listening mode:

The five-second rule: After they finish a statement, wait five seconds before responding. People often fill the silence with the most interesting thing they have said all day.
Echo the last three words: "...and it was really frustrating." You respond: "Really frustrating?" This encourages them to elaborate without you introducing any new direction.
Ask "tell me more about that": This is the single most useful follow-up in any interview. It signals genuine interest and gives them permission to go deeper.
Watch for contradictions: If someone says "I love my current tool" but earlier described spending 20 minutes working around a limitation, gently explore the gap. "You mentioned spending time on [workaround]. Can you walk me through that?"

What People Say vs. What They Do

A fundamental rule of user research: never trust what people say they will do. Trust what they have done. People are not lying. They are bad at predicting their own behavior. They overestimate how much they will use a new feature, underestimate how attached they are to their current workflow, and forget about the workarounds they have built.

This is why empathy maps separate "says" from "does." The discrepancy between these two quadrants is where the real insights live.

When possible, combine interviews with observation. Ask them to show you how they do the task while they talk about it. You will see steps they forgot to mention, tools they did not think to bring up, and frustrations they have normalized.

Analyzing Interview Data

After your interviews, you need to turn hours of conversation into actionable insights. Here is a practical process:

Debrief within 24 hours. Your memory of the conversation fades fast. Write a summary of the three most important things you learned from each interview while the details are still fresh.
Use affinity diagrams to cluster observations across interviews. Each observation goes on a separate note, then you group them by theme.
Look for patterns, not unanimity. If 5 out of 7 people mention the same pain point, that is a pattern worth acting on. You do not need 100% agreement to have a finding.
Identify surprises. The most valuable interview findings are the ones that surprised you. They indicate blind spots in your understanding.

Remote Interview Tips

Remote interviews work well with a few adjustments. Turn your camera on and ask them to do the same (facial expressions carry important emotional information). Use screen sharing when possible so they can show you their workflow. Record the session (with permission) so you can review moments you might have missed. And give the conversation an extra five minutes at the start for the awkward video-call settling-in period that does not happen in person.

Mistakes That Ruin Interviews

Pitching your solution. If you start explaining your product during a research interview, you have turned a learning session into a sales call. Save the demo for later.
Asking leading questions. "Don't you think it would be better if...?" will always get a yes. Ask neutral questions and let them lead you to their truth.
Interviewing only happy customers. People who love your product will confirm your beliefs. People who left will challenge them. You need both.
Not following up on emotions. When someone sighs, pauses, or uses strong language ("I hate when..."), that is a signal. Explore it.