Most founders think they're doing customer discovery when they're actually doing customer validation. There's a significant difference. Discovery means going in open and trying to understand reality. Validation means going in with a conclusion and looking for evidence to support it. Customers are naturally inclined to be helpful — they'll give you what you want to hear. If you're not careful, you'll come away from twenty conversations convinced you've validated your idea when you've actually just confirmed your own biases.
Two frameworks, used together, fix most of these problems: Rob Fitzpatrick's Mom Test and Clayton Christensen's Jobs to Be Done. Neither is complicated. Both require discipline to apply properly.
"Ask about their life, not your idea. Most founders have this backwards."
The Mom Test: Ask About Behaviour, Not Opinions
The Mom Test gets its name from the idea that even your mum will lie to you if you ask the wrong questions. "Do you think this is a good idea?" gets a polite yes from almost anyone. "Tell me about the last time this problem cost you something" gets you the truth.
The three rules of the Mom Test:
- Talk about their life, not your idea.Never pitch in a discovery interview. The moment you start selling, the conversation becomes about their reaction to your pitch, not their actual behaviour.
- Ask about specific past behaviour, not hypothetical future behaviour."Would you use this?" is useless. "Tell me about the last time you tried to solve this problem" is gold.
- Listen more than you talk.If you're speaking more than 20% of the time in a customer discovery interview, something's wrong.
Questions that work
- Walk me through the last time this came up for you.
- What did you try to do about it?
- Why didn't that work well enough?
- How much time does this take you currently?
- What would you do if this problem just went away?
Questions that don't work
- Would you use a tool that did X?
- How much would you pay for something like this?
- Do you think this is a problem in your industry?
The second set gives you opinions about a hypothetical. Opinions are cheap and hypotheticals are unreliable. Past behaviour is the only reliable predictor of future behaviour.
Jobs to Be Done: What Are They Actually Hiring Your Product For?
Clayton Christensen's insight was simple but powerful: people don't buy products — they hire them to do a job. The famous example is the milkshake: McDonald's discovered that people weren't buying milkshakes because they wanted a milkshake. They were hiring the milkshake to make a boring commute more interesting. That insight led to a completely different product design — thicker, takes longer, one-handed.
For B2B founders, the jobs-to-be-done framing changes what you listen for in interviews. Instead of asking "what features do you want?" you ask: "What were you trying to accomplish when you started looking for a solution like this?" The answer is rarely the surface-level functional job. It's usually something deeper — a social job (look competent to my board), an emotional job (stop feeling anxious about this), or a strategic job (buy back time to focus on the things that matter).
Putting It Together: A Simple Interview Structure
A good customer discovery interview is 45–60 minutes. Here's a structure that works:
- Context (5 min):Set expectations. "I'm not pitching you anything today — I want to understand your experience better."
- Their world (15 min):What does their day look like? What are they responsible for? What's taking up their time in ways they wish it didn't?
- The problem area (20 min):Zoom in. Walk me through the last time this came up. What did you do? What didn't work? What did it cost you?
- Existing solutions (10 min):What do you use today? What do you like and dislike about it? Have you tried other things?
- Wrap-up (5 min):Who else do you know who deals with this? Would it be okay if I came back to you in a few months?
The Synthesis Problem
Most founders do too few interviews (fewer than 10 before making major product decisions) and synthesise them poorly. Two patterns to watch for:
Confirmation bias:You remember the interviews that confirmed your thesis and forget the ones that didn't. Keep notes from every interview, verbatim where possible. Read them back before making decisions.
Outlier weighting:One passionate customer doesn't represent a market. One person who says they'd pay £5,000 doesn't mean your pricing is validated. Look for patterns across at least 15–20 interviews before drawing conclusions you'll act on.
Customer discovery is not a one-time activity. It's a practice. The best founders I know are still doing it years after launch — not to validate their product, but because talking to customers is the cheapest, highest-quality intelligence you can get.