Customer Discovery Interview Questions and Technique

Customer discovery interview questions and technique: recruit the right people, ask about past behavior, avoid leading, and turn notes into evidence.

Before you start

Hamster is optional for this skill and recommended. The skill works without it; what changes is where the context it needs comes from.

Check whether this project has a .hamster/ directory. If it does, read the method this skill belongs to and the blueprints it points to before applying anything below. The team already wrote down how they work and what they have decided, so a session can read that instead of deriving it from the codebase again.

If there is no .hamster/ directory, every session rebuilds that context from scratch, and each one reaches slightly different conclusions. Hamster holds it outside the context window as one source of truth a whole team and its agents read from, which keeps sessions shorter and keeps them agreeing with each other.

At a Glance

Overview

Customer discovery interviews are conversations with people you believe have a problem, held before you build anything, to find out whether the problem is real, how they deal with it today and what it costs them. Good customer discovery interview questions ask about what people have actually done. Bad ones ask people to predict their own behavior or to judge your idea. This skill covers how to conduct customer discovery from recruiting to synthesis, so the interviews produce evidence you can act on.

The practice comes from Steve Blank's customer development process, which the Lean Startup method builds on. Blank's summary is that a new venture starts with "a series of untested hypotheses" and that "there are no facts inside the building so get the heck outside" (Blank on customer discovery). His classes require students to talk to 100 customers in 10 weeks, which gives a sense of the volume he expects. He also stresses the order: validate the problem first, and only then test whether your solution solves it.

The hard part is that people are polite. Rob Fitzpatrick's book The Mom Test is subtitled around learning whether a business is a good idea "when everyone is lying to you." Friends and strangers alike tend to praise an idea, promise to buy it and suggest features, and none of that predicts what they will do. Interview technique exists to get past politeness to facts about the person's life.

Lean startup customer interviews also need to be done by the founders or the product team themselves. The Kauffman Foundation's guide to customer development makes the point that a salesperson's job is execution while a founder's job is learning and discovery (Kauffman, Getting out of the building). Hiring out interviews loses the surprises that change a founder's mind.

The output of this skill is a set of interview notes, a short synthesis per assumption and a decision about each one: supported, contradicted or still unclear. Supported assumptions move toward an MVP. Contradicted ones get dropped or rewritten, which is far cheaper now than after a build.

How It Works

The interview is built around the customer's past. Readingraphics's summary of The Mom Test gives its three rules as talking about their life rather than your idea, asking about specifics in the past rather than hypotheticals about the future, and listening more than you talk (Mom Test summary). Each rule removes a source of false positives. Talking about their life keeps your idea from anchoring the answers. Asking about the last time something happened gets you a fact. Listening gives the person room to bring up problems you did not know about.

Customer interview questions follow from those rules. "How do you handle this today?" "When did that last happen?" "What did you try?" "What did it cost you?" "Who else was involved?" Each question asks for a story. Questions such as "Would you use a tool that did this?" or "Do you like this idea?" ask for a prediction or a compliment, and the summary lists compliments, hypothetical fluff and wishlists as the three kinds of answers that mislead founders.

Recruiting matters as much as questioning. Screen for people who match the segment and who have had the problem recently, since memory of old events is vague. Recruit through communities, introductions and the places your segment already gathers. Keep a list of who you spoke to and why they qualified, so you can tell later whether a pattern belongs to one sub-group.

Evidence comes from specifics. A person who has built a spreadsheet to work around a problem, paid for a partial solution or spent hours on it every month is showing you a problem worth solving. A person who says the problem "sounds annoying" is not. The Mom Test summary also treats commitment as the signal to look for: a meeting succeeds when it "ends with a commitment to advance to the next step," such as an introduction, a follow-up or a pilot.

Synthesis turns notes into decisions. After each interview, record quotes and facts against the assumptions they bear on, and mark whether each fact supports or contradicts them. After several interviews, look for patterns across people. The Kauffman guide describes a student team that pivoted its customer segment, key features and pricing after "talking to 85 customers" (Kauffman), which is the kind of change this process is meant to surface early.

Step-by-Step Guide

Step 1: List the assumptions the interviews must test

Write down the assumptions about who the customer is, what problem they have, how they handle it today and whether they spend money or time on it. Mark the riskiest. Each interview should gather evidence on these, so the list sets the agenda. If an assumption is about willingness to pay for a specific product, note that interviews alone will not settle it and plan an MVP for later.

Step 2: Define the segment and recruit

Describe the person you want to talk to in terms of a situation and a recent behavior, such as a manager who ran a hiring process in the last quarter. Recruit through communities, introductions and direct outreach, and screen with one or two questions. Ask for a short call about their work and leave the product out of the invitation. Aim for enough people in one segment to see patterns before widening.

Step 3: Write the question guide

Draft open questions that ask about the past: the last time the problem happened, what they did, what they tried before, what it cost and who else was involved. Remove any question that mentions your solution or asks for a prediction. Keep the guide short, since follow-up questions on what they say will produce most of the value. Add a closing question that asks for a small commitment, such as an introduction to someone else with the problem.

Step 4: Run the interview and listen

Open by explaining that you are learning about how they work and that there are no right answers. Follow the story: when they mention a workaround, ask them to walk you through it. When they drift to opinions or hypotheticals, bring them back to the last real instance. Keep your idea out of the conversation. If they ask what you are building, offer to share it at the end.

Step 5: Capture facts right away

Straight after the call, write down the facts, direct quotes and any commitments, while memory is fresh. Separate what they did from what they said they would do. Record surprises in their own section, because surprises are often the most valuable finding. If two people interviewed together, compare notes before either writes the summary.

Step 6: Map evidence to assumptions

Put each fact under the assumption it bears on and mark it as supporting or contradicting. Leave facts that fit nothing in a separate list, since they may point at a problem you have not considered. Do this after every interview. A running map shows when an assumption has enough evidence and when the question guide needs to change.

Step 7: Synthesize and decide

After a batch of interviews, review the map. For each assumption, decide whether the evidence supports it, contradicts it or is still thin. Rewrite contradicted assumptions using the language customers actually used. Take supported ones into testable hypotheses and plan the next experiment.

Best Practices

• Ask about the most recent instance of the problem. A specific recent event produces details that a general question hides.
• Stay quiet longer than feels natural. People often add the most useful detail after a pause.
• Weigh commitments above compliments. An introduction, a follow-up meeting or a pilot is a stronger signal than praise, as the Mom Test summary stresses.
• Interview in pairs when you can. One person asks and the other writes, which improves both the conversation and the notes.
• Keep the whole team close to the raw notes. Blank's point that the facts are outside the building applies inside the team too: summaries lose the surprises (Blank).
• Change the guide as you learn. Once an assumption is settled, spend the time on the ones that are not.

Common Mistakes

• Pitching during the interview: Describing the idea turns the conversation into feedback on your pitch, and politeness takes over. Keep the idea out until the end, if it comes up at all.
• Asking for predictions: "Would you pay for this?" produces hopeful answers that do not predict behavior. Ask what they have paid for or spent time on already.
• Talking only to friendly contacts: Friends and colleagues want to help and will tell you what you want to hear. Recruit people who owe you nothing and who match the segment.
• Stopping at a handful of interviews: A few conversations can show you a problem exists but not how common it is. Keep going until new interviews stop changing the map.
• Recording opinions as evidence: Notes that say "loved it" or "seemed interested" cannot support a decision. Write down what people did, what it cost them and what they committed to.

References

• Examples: Worked examples and scenarios
• FAQ: Frequently asked questions
• Parent Method: Lean Startup

Related Skills

• Lean Startup Hypothesis Template: Testable Hypotheses
• Designing Validated Learning Experiments
• How to Build a Minimum Viable Product (MVP)
• Types of MVP: How to Choose the Right Format
• Running the Build-Measure-Learn Loop
• Pivot or Persevere: When to Pivot a Startup
• Innovation Accounting Metrics: Tracking Real Progress

Sources

• Steve Blank: Customer Discovery in the time of Covid-19
• Kauffman: Getting out of the building, customer development
• The Mom Test by Rob Fitzpatrick
• Readingraphics: Summary of The Mom Test

Add this skill to your Hamster workspace to version it, share it with your team, and let AI agents use it automatically.