How to conduct interviews

This lesson teaches you how to run an interview that produces real insights instead of polite nods. You will learn how to ask open questions that surface honest answers, how to follow up when someone says something interesting, and how to avoid leading the conversation toward the answers you want to hear.

You have interviews scheduled. Now you need to make sure those conversations produce insight you can actually use.

Most customer interviews fail because the interviewer talks too much, asks leading questions, or accepts surface-level answers. The customer says something vague like "it was fine" and the interviewer moves on instead of digging deeper.

In this chapter, you will learn how to structure the conversation, what questions to ask, and how to get past polite answers to the real story. You will also learn why what people say and what they do are often completely different, and how to uncover the truth.

By the end, you will have a framework for running interviews that reveal genuine insight, not just confirmation of what you already believe.

Start with a few minutes of context. Thank them for their time, explain why you are doing this research, and tell them how long it will take. Keep it brief. You want to get them talking, not listen to you explain your methodology.

Set expectations: there are no right or wrong answers. You want their honest experience, even if it is critical. Reassure them that the conversation is confidential.

Ask permission to record. I use Zoom or Google Meet with Fireflies running for the transcript. Recording lets you focus on listening instead of frantic note-taking. Most people say yes if you explain it is just for your reference and will not be shared externally.

Then get out of the way. Your job is to listen, not to pitch or defend. If they criticise something, do not explain why it is that way. Just ask them to tell you more.

The biggest trap in customer interviews is asking people what they think. Opinions are unreliable. People tell you what sounds good or what they believe you want to hear.

Focus on behaviour instead. "What did you actually do?" is more useful than "What do you think about...?" People rationalise opinions but struggle to lie about their actions.

When I was leading the experimentation programme for NU.nl, the largest news website in the Netherlands, we had an eye-tracking lab where we could see exactly what people looked at on the website. We also interviewed users about their reading habits.

Nearly everyone said they read serious news and never touched the gossip section. They were educated professionals. They did not read celebrity drama.

But the data told a different story. Gossip was the most popular part of the site by far. And in our lab tests, every single user was on a gossip page within sixty seconds. Even the ones who had just told us they never read it.

What people say and what they do are completely different. Your questions need to uncover behaviour, not opinions.

Key questions to ask:

• What was going on in your business when you started looking for a solution?
• Walk me through how you found us. What else did you look at?
• What nearly stopped you from buying?
• What has the experience been like since you started?
• If you could change one thing, what would it be?

Adapt based on who you are interviewing. For churned customers, focus on what changed. For lost prospects, focus on why the competitor won.

When you hear something interesting, follow up. "Tell me more about that" and "Can you give me an example?" are the most powerful prompts in customer research. They unlock detail without leading the witness.

Resist the urge to fill silence. When someone finishes answering, pause for a moment. Often they will keep talking and share something more revealing. The real insight is usually in the second thing they say, not the first.

Do not lead the witness. Saying "So you found the onboarding confusing, right?" puts words in their mouth. Instead say "How would you describe the onboarding experience?" Let them choose the words.

For a jewellery company that sold both in retail stores and online, we noticed they sold far more rings offline than online. The data was clear, but we did not understand why.

Customer interviews revealed the problem. The ring sizing guide was a five-page PDF from the factory, designed for B2B wholesale buyers. It was technical and confusing. Most website traffic was mobile, and nobody opens a five-page PDF while shopping on their phone. In the store, staff could measure your finger in seconds.

Once we knew the problem, the fix was obvious. But without talking to customers, we would have kept guessing.

I once interviewed customers for a coffee retailer who were buying beans all year round, then suddenly switching to coffee pads every summer. The pattern looked like a data error. Some kind of tracking bug.

When I called them, I discovered these customers had boats, camper vans, or holiday homes. In summer, they wanted the same brand but needed a different format for their portable espresso machine. That insight became a marketing campaign: "Take your favourite coffee with you, even when you travel."

You cannot get these insights from analytics. You have to ask.

If you are recording with a transcription tool like Fireflies, take minimal notes during the call. Jot down timestamps for moments you want to revisit. Focus on being fully present in the conversation.

If you are not recording, write down exact phrases they use. Their language is gold for your marketing. "I was drowning in spreadsheets" is more valuable than your summary "they had data management issues."

Immediately after the call, spend five to ten minutes writing up your key takeaways while everything is fresh. What surprised you? What patterns are starting to emerge? What quotes stood out?

The recording is your backup. The real-time notes capture your gut reaction while it is still sharp.

Good customer interviews feel like conversations, not interrogations. You ask open questions, listen far more than you talk, and follow interesting threads wherever they lead.

The best insights come from unexpected places. A data anomaly turns out to be camper vans. A high churn rate turns out to be a PDF. You will never predict what you learn. You just have to ask and keep digging until you find it.