Customer Discovery Interviews: Questions and How-To Guide

Customer discovery interviews validate whether a real, painful problem exists before you build - by asking about the customer's past behavior, not your idea. Pairing them with the jobs to be done lens sharpens what you learn about the progress a customer is trying to make. The best questions are open and backward-looking ("Tell me about the last time you dealt with X"), never leading ("Would you use a tool that does Y?"). Aim for 15-30 conversations per segment, listen for specifics over compliments, and count only costly signals as validation.

This is the field guide for the interviews themselves. They are the primary way to measure product-market fit before you have the volume for a survey, and they pair directly with finding design partners and getting your first customers.

What Is a Customer Discovery Interview?

A customer discovery interview is a structured, one-on-one conversation with a potential customer to understand their problems, workflows, and priorities - before you commit to a solution. Popularized by Steve Blank's customer development method and Rob Fitzpatrick's book The Mom Test, its goal is to gather evidence, not to pitch or sell. You are testing your assumptions about the problem, not validating your ego.

What Questions Should You Ask in a Discovery Interview?

Great questions dig into real, recent behavior and its cost. Bad questions ask people to predict their future behavior or to rate your idea - both produce polite lies. Use this contrast:

What Are the Rules of the Mom Test?

Rob Fitzpatrick's core idea: ask questions so good that even your mom cannot lie to you about your business. Three rules:

• Talk about their life, not your idea. The moment you describe your product, everyone turns polite and the data goes bad.
• Ask about specifics in the past, not opinions about the future. "Last time" beats "would you."
• Talk less, listen more. You already know your idea; you are there to learn theirs.

How Many Interviews Do You Need, and Who With?

There is no magic number, but patterns usually emerge by 15-30 interviews per distinct segment - when new conversations stop surprising you, you have enough for now. Interview people who have the problem, not people who like you; friends and supporters give the most dangerous false positives. Recruit from your accelerator network, relevant communities, LinkedIn, and warm intros, and screen for the acute-pain persona.

How Do You Turn Interviews into a Product-Market Fit Signal?

Discovery interviews are your PMF instrument before you have metrics. Score signals by cost to the interviewee, not by warmth:

• Strong: they already pay for a workaround, built a spreadsheet hack, or ask "when can I use it?" and offer to be a design partner.
• Medium: they describe the problem vividly and unprompted, and can name the cost.
• Weak (ignore): "cool idea," "I would probably use that," "let me know when it launches."

Log every interview the same way - problem, current workaround, cost, and the costliest commitment they made - so you can see the pattern across the batch instead of remembering the most flattering call.

TL;DR

• Interview to learn, not to pitch. The goal is evidence about the problem, not validation of your idea.
• Ask about past behavior, never hypotheticals: "Tell me about the last time..." beats "Would you use...".
• Follow The Mom Test: talk about their life, ask for specifics in the past, listen more than you talk.
• 15-30 per segment until conversations stop surprising you; interview people with the pain, not fans.
• Count only costly signals - paying for a workaround or offering to pilot - as real validation.

How to Recruit Interviewees Without Bias

Recruiting well is half the battle, because the wrong people produce false comfort. Source from the communities where your target persona already talks - LinkedIn, relevant subreddits, accelerator slacks - and screen for the acute-pain profile rather than willingness to chat. Friends, family, and existing fans are the most dangerous recruits because they want you to succeed and will hand you the pleasant lies The Mom Test warns against.

Offer a small, non-cash incentive if needed, but avoid paying so much that people say what they think you want. The best recruit is someone who recently spent money or effort on the problem, since they can describe a real workaround and its cost. Keep a simple tracker of who you contacted, who agreed, and which segment they represent so your 15 to 30 interviews per segment stay balanced instead of clustered in one easy channel.

Running the Interview: Structure and Cadence

Open by setting context: you are researching a problem, not selling, and there are no wrong answers. Spend the first minutes on their world - their role, their workflow, the last time the problem bit them - before anything resembling your solution. Let silences sit; the urge to fill them is what pushes interviewees toward pleasing answers. Take verbatim notes on the costly details: the workaround, what it cost, and what they tried before.

Keep it to thirty minutes so it respects their time and you can fit more conversations. Close by asking who else has this problem and whether they would be a design partner, which surfaces your next recruit and your strongest validation signal at once. The discipline of listening more than talking is what separates evidence from flattery.

Discovery Interviews vs Later-Stage Customer Research

Discovery interviews happen before you have a product or a metric, so they trade statistical power for raw signal about whether a painful problem exists. Once you have usage data, you shift to analytics and surveys, which confirm patterns at scale but miss the why. The two are complementary: interviews find the hypothesis, data tests it.

For a startup, the mistake is skipping interviews and jumping to a survey or a build. A survey of people who have not experienced the problem yields noise, and a build based on an untested assumption wastes the scarcest resource - time. Run the interviews first to earn the right to the louder, quantitative methods later, and you will have a sharper survey and a more focused roadmap.

FAQ

What Is a Customer Discovery Interview?

It is a structured one-on-one conversation with a potential customer to understand their problems and workflows before you build a solution. Drawn from Steve Blank's customer development method, its purpose is to gather evidence about whether a painful problem exists - not to pitch your product or make a sale.

What Questions Should I Ask in Customer Discovery?

Ask open, backward-looking questions about real behavior: "Tell me about the last time you dealt with this," "How do you handle it today," "What did that cost you," and "What have you tried and why did you stop." Avoid leading questions like "Would you use a tool that does this," which invite polite but useless answers.

How Many Customer Discovery Interviews Do I Need?

Usually 15 to 30 per distinct segment. The practical stopping point is saturation - when new interviews stop surprising you and you can predict the answers. If every conversation still reveals something new, keep going; if you are hearing the same problems and workarounds repeatedly, you have enough to act.

What Is the Mom Test?

The Mom Test, from Rob Fitzpatrick, is a set of rules for asking questions so good that even your mom could not lie to you about your business. The rules: talk about the customer's life instead of your idea, ask for specifics about the past rather than opinions about the future, and listen more than you talk.