Talking to potential customers sounds obvious and easy. It's actually a skill most people get subtly wrong in the same way: they ask questions that feel like research but are secretly designed to get reassurance instead of truth. Fix that one flaw and customer conversations become the highest-leverage hour you can spend before building anything (see The Smallest Possible Version).
The trap: leading questions that flatter you
Ask someone "would you use an app that helps you cook healthier meals?" and almost everyone will say yes — it's a hypothetical, it's polite, and agreeing costs them nothing. This is the single most common mistake in early customer conversations: asking about a hypothetical future, which people answer with hope and politeness rather than truth. You walk away encouraged by data that was never real.
People are terrible at predicting what they'd do. They're excellent at accurately reporting what they've already done.
The fix is a shift that changes everything: stop asking about the future, and ask about the past. Instead of "would you," ask "did you" or "when was the last time." Past behavior is real data. Hypothetical future behavior is often just kindness wearing a lab coat.
The Mom Test: a simple filter for good questions
A useful mental test, borrowed from a well-known startup guide: would this question get an honest, useful answer even from your own mother, who loves you and doesn't want to hurt your feelings?
"Do you think my business idea is good?" fails instantly — of course she'll say yes. "How do you currently handle this problem?" passes — it's about her actual life, not your idea, so even someone who loves you unconditionally would answer it honestly, because it isn't a referendum on you at all.
Every question you plan to ask can be run through this filter. If the honest answer requires the other person to risk hurting your feelings, rewrite the question so it doesn't.
Questions that actually work
"Walk me through the last time you dealt with [the problem]." Real, specific, recent behavior — the gold standard of customer research.
"What have you already tried to solve this?" Reveals whether the problem is painful enough that people have already spent effort or money on it — a much stronger signal than any opinion about your idea.
"What's the hardest part about that?" Digs past the surface complaint toward the real friction, which is often not what you assumed going in.
"Tell me about the last time that happened." If someone struggles to recall a recent, specific instance, the problem likely isn't urgent or frequent enough to build a business around — a genuinely useful negative signal, delivered gently.
"What are you using now, and what do you like or dislike about it?" Existing behavior and existing tools reveal real constraints and real competition your idea will actually face.
What to avoid entirely
Don't pitch during discovery. The moment you describe your idea, you contaminate the conversation — people start being nice to your idea instead of honest about their problem. Keep discovery conversations focused entirely on their experience; save the pitch, if you need one at all, for later and separately.
Don't ask "would you pay for X?" — a hypothetical about money is just as unreliable as any other hypothetical. A far stronger question: "would you pay for it right now?" — a real, immediate ask (even a small deposit or pre-order) is the only thing that reliably separates true interest from polite enthusiasm.
Don't only talk to people who already like you. Friends and family are the least reliable source of honest feedback about a business idea, precisely because they don't want to hurt you. Seek out strangers in your target group whenever possible.
How many conversations is enough
You don't need a formal study — somewhere between 5 and 15 real conversations with people genuinely in your target group is usually enough to spot a clear pattern: either the same problem and frustration keeps coming up unprompted, or it doesn't. A consistent pattern across even a handful of honest conversations is a far stronger signal than a hundred hypothetical "yes" answers to a leading survey question.
The takeaway
Customer conversations are only useful if they surface truth instead of politeness. Ask about specific past behavior, not hypothetical future behavior — "when did you last," not "would you." Run every question through the Mom Test: would even your mother answer honestly? Avoid pitching during discovery and avoid hypothetical money questions; a real ask (a deposit, a pre-order) beats any amount of polite enthusiasm. A handful of honest conversations beats a hundred leading ones.
Try this: write down the exact three questions you'd ask a prospective customer this week, then run each one through the Mom Test before you ask anyone anything. Rewrite any question your mother couldn't answer honestly without risking your feelings.
Sources & Further Reading
Sources
- The "Mom Test" framework and the emphasis on past-behavior over hypothetical-future questions comes from Rob Fitzpatrick's widely-used customer-discovery methodology, itself building on Steve Blank and Eric Ries's customer-development and Lean Startup work.
Go deeper
- 📘 The Mom Test — Rob Fitzpatrick (short, sharp, and entirely built around this one idea)
- 📘 The Four Steps to the Epiphany — Steve Blank (deeper customer-development methodology)
- 🌐 Search "customer discovery interview questions" for further real-world question banks
- 🌐 Companion piece: The Smallest Possible Version (in this library)