How do you talk to users without getting false validation?

Why most founder interviews produce applause, not evidence

"We talked to users" is one of the most common phrases in failed-startup post-mortems. The conversations happened — they just produced compliments instead of facts. The moment your idea enters the room, every answer that follows is reacting to your pitch, not reporting reality.

Rob Fitzpatrick's The Mom Test names the trap: if you ask the wrong questions, even your mother will lie to you — not maliciously, but politely. Three lies dominate, and all three feel like progress:

• Compliments ("this is really cool, I'd use it") — data about manners, not market.
• Hypotheticals ("would you pay for this?") — a guess about an imagined future people are bad at predicting.
• Feature ideas ("you should add…") — deflection that lets someone engage without admitting the core problem doesn't burn.

The shift: from pitching to mining

The fix is a posture change. You are not in the meeting to convince anyone of anything. You are a miner extracting facts about someone's life as it already is. Past behaviour is evidence; opinions about the future are not.

• What they did the last time the problem came up — not what they would do.
• What the problem already costs them in time, money, or risk.
• What they have already tried — the ugly spreadsheet, the intern, the workaround.
• Who else suffers it, and who pays when it goes wrong.

The test for a good interview

If the conversation ends and the person never learned what you're building, you ran it perfectly. Five interviews that surface real past behaviour teach you more than a launch to silence — silence has no transcript you can study.

Frequently asked questions