Run your first five user interviews in one week. Talk to people who used the product in the last month, plus a couple who left. Ask about the last time the problem happened, never whether they like your idea. Count what repeats across the calls, then change one thing before the next five.
That's the whole method. The rest of this post is how to do each part without fooling yourself.
Why most first interviews go nowhere
A product lead I spoke to this month tried it the classic way. Their team emailed users who had hit a problem and offered a gift card for a 10 to 15 minute call. About one in five said yes. By the time the calls happened, most of those users had forgotten what went wrong.
The other common way to waste a call is to ask for opinions. Ask "would you use this?" and people say yes to be nice. They aren't lying on purpose. They just can't predict what they'll do, and they don't want to hurt your feelings. Rob Fitzpatrick wrote a whole book on this trap, The Mom Test, and it's worth an afternoon.
The fix is to ask about what already happened, not what might happen. Last Tuesday is a fact. Next month is a guess. Keep the call on their week and off your idea, and the answers get honest on their own.
Who to talk to
Pick five people who used the product in the last 30 days and two or three who signed up and left. The ones who left are the hardest to book and teach you the most.
If you can see usage data, use it to choose. Casey Winters treats retention as the real measure of fit, so look at who came back in their second week and who didn't, and talk to both groups. Rahul Vohra did a version of this at Superhuman. He asked users how they would feel if they could no longer use the product, then studied the ones who said "very disappointed" to learn what they loved and who else was like them.
Skip friends, investors and anyone who already loves you. They will be kind, not useful.
Book the calls as close to the moment as you can. A user who churned yesterday remembers why. A user who churned last month will give you a reasonable guess.
How to open the call
Read this out, or something close to it:
"Thanks for making time. I'm not selling anything today. I'm trying to understand how you handle [the problem] right now. Can you walk me through the last time it came up?"
That last question moves the call from opinions to a real story. From there, your job is mostly to keep the story going.
Questions to ask and questions to drop
This is the table to keep. The left column is what most founders ask on their first calls. The right column is what to ask instead.
| Instead of this | Ask this |
|---|---|
| Would you use a product that does X? | How do you deal with this today? |
| Do you think this is a good idea? | Tell me about the last time this happened. |
| How much would you pay for this? | What have you already spent fixing it, in time or money? |
| Would this feature be useful? | Why do you want that? What would it let you do? |
| Is this a big problem for you? | Why do you bother? What else have you tried? |
| Anything else you'd want? | What happened next? |
Two habits make these work. Follow every answer with "what happened next?" until they run out of story. And talk for less than a third of the call. When they go quiet, wait. Silence is where people add the detail they skipped the first time.
When someone asks for a feature, don't write it down and move on. Treat it as a clue, not an instruction, and dig under it. "Why do you want that?" usually turns a feature request into the real problem, and the real problem often has a cheaper fix than the feature.
How to tell if a call went well
Compliments feel good and tell you nothing. "This is great, I'd definitely use it" costs the person nothing to say, and neither does "I would" or "I always".
What counts is what they give up for you. At the end of the call, ask for something small that costs them a little. A second call with their cofounder, an intro to someone else with the same problem, a slot to try a prototype, or a preorder all work. If they say yes, the problem matters to them. If they dodge, you learned something too.
After the fifth call, count the repeats. A problem you heard three times is real. A problem you heard once is a lead for the next call. Change one thing based on what you heard, ship it, then book the next five.
The forgetting problem from the top of this post is why we're building Produck. It talks to users inside your app at the moment they get stuck, so you hear the story while they still remember it. If you want help choosing who to call and what to ask, book a call with our team. Helping teams learn from their users is the whole job.
Further reading
- The Mom Test by Rob Fitzpatrick. Short, practical and the best book on this.
- How Superhuman Built an Engine to Find Product Market Fit by Rahul Vohra. The survey question to pair with your interviews.
- Casey's Guide to Finding Product/Market Fit by Casey Winters. Why retention tells you who to call.
- A Playbook for Achieving Product-Market Fit the Lean Way by Dan Olsen. His pyramid helps you check you've picked the right customer and need before you build. Conrad walks through it in What Is PMF.
- Our PMF resources page has the rest of the reading list we share with founders.