How to run user interviews that reveal the truth
Questions, habits and follow ups that get past polite answers.

Zander Whitehurst
7 min read
User interviews are one of the most valuable skills a designer can have, and one of the easiest to get wrong. Done well, they reveal what people actually need. Done badly, they produce a pile of polite, unhelpful answers that confirm whatever you already believed.
My first ever interview was a disaster. I asked a participant whether she would use a new feature we were planning. She said "oh yes, definitely, that sounds great". We built it. She never used it. Neither did anyone else. That was the day I learned that people are lovely, helpful and completely unreliable at predicting their own behaviour.
Ask about the past, not the future
The single most useful rule in interviewing is to ask about what people have done, not what they might do. "Would you use this?" invites guesses. "Tell me about the last time you did this" invites stories, and stories contain truth.
Good questions sound like:
- Tell me about the last time you tried to do this.
- What happened next?
- What was the hardest part?
- What did you do instead?
Each one anchors the conversation in real experience rather than hypotheticals.
Prepare a guide, not a script
Write down the topics you want to cover and a handful of opening questions. Then let the conversation breathe. A rigid script turns an interview into a questionnaire. A loose guide lets you follow the interesting threads, which is where the insights live.
Order your topics from broad to specific. Start with the person and their context, move towards the problem and only at the very end touch on any solution ideas.
Get comfortable with silence
This is the hardest skill and the most powerful. After someone answers, wait. Count to five in your head if you need to. People fill silence, and what they say after the pause is often far more honest than their first answer.
I am naturally a talker, so this took me ages to learn. I used to write "SHUT UP" at the top of my interview notes. Slightly aggressive, but it worked.
Follow up relentlessly
The first answer is rarely the real answer. When something interesting comes up, dig into it:
- Why was that frustrating?
- Can you show me?
- What did you expect to happen?
- How often does that happen?
"Can you show me?" is especially valuable. Watching someone use their actual tools reveals workarounds and habits they would never think to mention.
Do not lead the witness
Leading questions quietly push people towards the answer you want. "Do you find the dashboard confusing?" suggests that it is. "How do you use the dashboard?" lets them tell you.
Watch your reactions too. If you light up when someone says something you hoped to hear, they will notice and start saying more of it.
Take notes that capture quotes
Write down exact words wherever you can. A direct quote is far more persuasive to your team than your summary of what someone meant. Record sessions if you have permission, but still take notes, because notes force you to listen for what matters.
Mark moments that surprised you. Surprises are the most valuable part of any interview, and they are easy to forget once you start synthesising.
Talk to the right people
Five interviews with the right people beat twenty with the wrong ones. Recruit people who have recently experienced the problem you are exploring. Recent memories are far more detailed and accurate than distant ones.
If you can, include a few people who tried your product and gave up. They often teach you more than your happiest customers.
Debrief straight away
Spend ten minutes after every interview writing down your top three takeaways while they are fresh. Patterns start appearing after three or four sessions, and these quick debriefs make synthesis far easier later.
One more thing I tell every cohort: interviews are not about being clever. They are about being curious. Show up genuinely interested in the person in front of you and you will be amazed by what they tell you. We practise interviewing as part of discovery in the AI Product Management Bootcamp.
Start this week
Find two people who have used your product recently and book thirty minutes with each. Ask them about the last time they used it, then listen far more than you talk. Those two conversations will teach you more than a month of guessing.
Memorisely