How to talk to users — the Michael Seibel playbook every YC founder learns.

The one rule that matters

Ask about past behavior. Never ask about future intent. "Tell me about the last time you tried to solve this problem" gives you a story with real details, real workarounds, and real money spent. "Would you use a product that…" gives you a polite lie.

Rob Fitzpatrick codified this in The Mom Test — required reading in every YC batch. The premise: even your mom will lie to you if you ask her opinion on your startup. She won't lie if you ask her what she did yesterday.

The question templates

• "Walk me through the last time you did X."
• "What's the hardest part about doing X today?"
• "What do you currently use for X? How did you find it?"
• "How much time or money does that cost you per month?"
• "What have you tried that didn't work?"

Notice what's missing: no "would you", no "do you think", no pitch. You're a journalist, not a salesperson.

The 15 / 100 rule

Do 15 interviews before you write a line of code. Do 100 before you claim product-market fit. What matters more than the number is talking to the same customer segment repeatedly until you can predict the next answer before you ask.

When you stop being surprised, you've found the shape of the problem. That's when you start building.

The five mistakes YC partners see in every batch

1. Pitching instead of listening. If you spoke more than 30% of the interview, you did it wrong.
2. Asking hypothetical questions. "Would you pay for this?" is worthless. "How much do you pay today?" is gold.
3. Interviewing friends. They're supportive, not honest. Find strangers who fit your ICP.
4. Wrong ICP. Talking to 50 people who won't ever buy teaches you nothing.
5. Not recording. Memory rewrites conversations to match what you wanted to hear.

What YC actually wants to see in your application

When you write "we talked to 50 users", YC partners want the specifics: who they were, what they said, and what surprised you. See how to apply to YC for the exact application phrasing that signals real research.