How to Run Customer Discovery Interviews That Actually Teach You Something

Customer discovery interviews are the cheapest and most underused tool in early-stage business building. Done well, twenty conversations can save a year of building the wrong thing. Done badly, they produce a stack of polite encouragements that quietly confirm what the founder already wanted to believe.

The difference is method. Strangers will not volunteer the truth — they will be helpful, which is not the same thing. Your job is to design the interview so the truth surfaces despite the politeness.

The single most important rule: ask about the past, not the future

Humans are unreliable predictors of their own future behavior but excellent reporters of what they actually did. 'Would you pay for a tool that did X?' produces meaningless yeses. 'Tell me about the last time you ran into this problem. What did you try? What did it cost you?' produces evidence. Every question in your script should anchor to a specific past event, not a hypothetical future product.

A simple interview script

Spend the first five minutes on context: who they are, what they do, how their work or life is structured. Then move into the problem area without naming your solution. The goal is to understand how often the pain occurs, how painful it is on a one-to-ten scale, what they currently do about it, how much they spend or save in time and money, and what would have to be true for them to switch.

• 'Walk me through the last time you experienced this.'
• 'What did you try? In what order?'
• 'What did that cost you in time and money?'
• 'Who else was involved in the decision?'
• 'What stopped you from solving it better?'
• 'If you had a magic wand, what would the ideal outcome look like?'

What not to do

Do not pitch. The moment you describe your solution, the conversation shifts from learning to selling, and the data becomes worthless. Do not ask leading questions ('don't you think it would be great if...'). Do not let the interviewee politely redirect to vague enthusiasm — ground them again with 'tell me about a specific time.' Do not interview your friends, family, or anyone who has a personal interest in being nice to you.

How to know it is working

After ten interviews, you should be able to describe the customer's world back to them in language they would use themselves. You should hear at least three independent descriptions of the same workaround, the same trigger event, and the same frustration. If the interviews still feel scattered, your customer definition is too broad — narrow it and run another ten.

The output is not a list of features the customer requested. It is a clear understanding of the job they are trying to get done, the cost of not doing it, and the alternatives they already use. Features come later, and they come easily once that picture is sharp.

Finding people to interview

The most common reason founders skip interviews is that they do not know where to find willing strangers. The trick is to go where your customer already gathers and to ask for help learning, not for a sales meeting. Niche online communities, industry subreddits, LinkedIn, professional associations, and the comment sections of articles your customer reads are all rich sources. A short, honest message — 'I'm researching how people handle X and would value fifteen minutes to learn from your experience' — converts far better than anything that smells like a pitch.

If reaching your customer is genuinely hard, treat that as a finding rather than an inconvenience. A customer you cannot reach for a free conversation will be even harder to reach with a paid product. Sometimes the most valuable result of customer discovery is discovering that your chosen segment is too scattered or too guarded to build an affordable business around.

Turning interviews into decisions

Interviews only matter if they change what you do. After each batch of ten, sit down and convert the raw conversations into a small number of clear conclusions.

• Summarize the customer's world in their own words — if you cannot, your segment is too broad.
• Note the recurring trigger event that makes the pain urgent; that moment is when they will buy.
• Quantify the cost of the problem in time and money; it sets the ceiling on what you can charge.
• List the alternatives they use today — those, not other startups, are your real competition.
• Decide one thing to test next based on what you heard, then run the next ten interviews.

Turning interviews into decisions

Interviews are only valuable if they change what you build or how you sell. The biggest mistake founders make is treating conversations as a checklist to complete rather than evidence to act on. After each batch of interviews, force yourself to write down the patterns you actually heard, separating what people said they wanted from what their past behavior suggests they will do. The goal is to leave with a short list of conclusions specific enough to act on — a sharper customer segment, a clearer top problem, a feature to cut, a message to test.

Be especially careful with politeness and projection. People are kind in interviews and will often praise an idea to avoid disappointing you, so weight concrete past actions and expressed willingness to pay far more heavily than enthusiastic hypotheticals. Watch for the difference between a problem people merely acknowledge and one they are already spending time or money trying to solve; only the latter signals a market. When the same painful problem and the same language keep surfacing across many interviews, you have found something real. When every conversation points in a different direction, you have not yet found a customer worth building for, and more interviews — not more building — is the right next step.