Customer discovery interviews

Why most interviews produce false positives

People are kind. Ask 'would you use a tool that did X?' and most will say yes, because saying yes costs them nothing and disappointing you feels rude. Those yeses are the single most common cause of a product being built for nobody.

The fix is to stop asking about the future and ask about the past. Behaviour that already happened cannot be invented to please you. What did you do last time this went wrong? What did it cost? What did you try? What did you pay for? Those answers are evidence.

Recruiting the right ten people

Ten focused interviews with people who have the problem beat a hundred survey responses from a general audience. Recruit where the problem is discussed: the professional association, the subreddit, the WhatsApp or Slack group, the LinkedIn search for the exact job title, the review section of the tool they currently tolerate.

Screen before you book. Two questions are enough: how often does this situation happen to you, and what do you currently do about it? Anyone who answers 'rarely' or 'nothing' is not in your segment and will pull your findings towards a comfortable lie.

Never interview friends and family about commercial demand. They are optimising for your feelings, which is the opposite of what you need.

The question set

Opening: 'Walk me through the last time [situation] happened.' Then stay quiet. The story contains the frequency, the workaround, the people involved and the cost, without you leading any of it.

Cost: 'How long did that take, and who else got pulled in?' Convert to money later; do not put a figure in their mouth now.

Workaround: 'What are you doing about it today?' followed by 'What have you already tried and dropped?' Dropped tools reveal both the demand and the reason competitors failed.

Spend: 'Have you ever paid for anything to deal with this?' Existing spend is the strongest predictor that a budget exists.

Priority: 'Of the problems in your work right now, where does this sit?' A real problem that ranks eighth will never get bought.

Close: 'Who else should I talk to?' A referral is also a small signal that the conversation mattered to them.

Questions to never ask

'Would you use this?' — hypothetical, and free to answer yes.

'How much would you pay?' — people cannot price something they have not used; test price with an actual offer instead.

'Do you think this is a good idea?' — invites them to review you rather than describe themselves.

Any question containing your solution before you have finished hearing their problem. Once you pitch, the interview is over and a sales call has begun.

Scoring what you heard

For each interview record four numbers: frequency of the problem, hours or money it currently costs, whether they already pay for a workaround, and where it ranks against their other problems. Anything else is colour.

The pattern you want across ten interviews: the same problem described in the same words, an existing workaround people already spend on, and at least a handful ranking it in their top three. That is a segment worth serving.

The pattern that means stop: enthusiasm with no existing spend, ten different versions of the problem, or a top-three ranking from nobody. It is far cheaper to hear that now than after six months of building.

Then convert findings into a testable offer — a landing page with a price, a pre-order, a paid pilot — because interviews establish the problem and only money establishes the market.