Customer Discovery Interviews: A Practical Framework

Why Customer Discovery Interviews Fail

Rob Fitzpatrick's "The Mom Test" identifies the core problem: people lie in interviews. Not maliciously -- they are polite, they want to be helpful, and they are terrible at predicting their own future behavior. Ask someone "Would you pay for a product that does X?" and they will almost always say yes, because saying no feels rude. This politeness bias means that interviews designed around hypothetical questions produce systematically misleading data.

Effective discovery interviews avoid hypotheticals entirely. They focus on past behavior, concrete situations, and observable facts. Instead of "Would you use this?" ask "How do you handle this today?" Instead of "Would you pay EUR 500/month for this?" ask "How much do you currently spend solving this problem?" The answers to backward-looking, fact-based questions are far more predictive of future behavior than the answers to forward-looking, opinion-based questions.

The Interview Structure: Past, Present, and Future

Structure each interview in three phases. Phase 1: Past behavior (15 minutes). Ask about the last time they experienced the problem you are investigating. "Tell me about the last time you had to do X. Walk me through what happened." Follow up with specifics: "How long did it take? Who else was involved? What tools did you use? What went wrong?" This reconstructs their actual experience, which is far more informative than their general impression of the problem.

Phase 2: Current alternatives (10 minutes). Understand how they solve the problem today. "What do you currently use for this? How did you choose that solution? What do you like about it? What frustrates you?" If they have not solved the problem at all, that is a data point -- it might mean the problem is not painful enough to justify a solution, or it might mean existing solutions are so bad that they have given up looking. Phase 3: Priorities and constraints (5 minutes). Explore what would make them switch. "If you could change one thing about how you handle this, what would it be? What has prevented you from solving this differently?" End with: "Is there anything I should have asked that I did not?" This open question frequently surfaces the most valuable insights of the entire interview.

Questions That Reveal Willingness to Pay

Willingness to pay is the hardest thing to assess in discovery interviews because people systematically understate it. Never ask "How much would you pay for this?" Instead, use indirect methods that reveal the economic value of the problem. "How many hours per week does your team spend on this?" (Multiply by their labor cost to estimate the problem's economic impact.) "Have you tried to solve this before? How much did you spend?" (Past spending reveals actual price tolerance.) "What would it mean for your business if this problem was completely solved?" (This reveals the upside value, which anchors a higher willingness to pay.)

The strongest willingness-to-pay signal is when the person is already spending money on a partial or inferior solution. If they pay EUR 2,000/month for a tool that solves 40% of their problem, you have strong evidence that they would pay more for a tool that solves 100% of it. If they are not spending anything and not actively searching for a solution, the willingness to pay is likely low regardless of what they say in the interview.

How Many Interviews Are Enough

The answer is: enough to reach saturation, where new interviews stop producing new insights. In practice, this is typically 15-25 interviews for a well-defined customer segment. After 10 interviews, you will have identified the major themes. Interviews 11-20 confirm and refine those themes. Interviews 21-25 should produce diminishing returns -- if they are still generating surprising insights, you have not yet reached saturation and should continue.

Interview multiple segments separately. If you are exploring two potential customer segments, run 15 interviews per segment, not 30 combined. The patterns that emerge from one segment may be completely different from another, and mixing them in analysis produces muddy, non-actionable insights. After each round of interviews, synthesize your findings before deciding whether to continue or move to the next phase of validation.

Synthesizing Interview Data into Actionable Insights

After completing your interviews, synthesis turns raw notes into strategic inputs. Start by identifying patterns across interviews. Create a grid with interviewees as rows and key themes as columns: problem severity, current solution, switching triggers, willingness to pay, and decision-making process. Fill in each cell with specific quotes and data points. The patterns that emerge from this grid -- not from individual interviews -- are your actionable insights.

Look for three outputs. Problem validation: is the problem real, frequent, and painful enough that people will pay to solve it? If fewer than 60% of interviewees describe the problem as a top-3 priority, the problem may not be painful enough to drive purchasing behavior. Segment clarity: which customers experience the problem most acutely and are most likely to buy? Your ICP should emerge from the interview data. Value proposition: what specific outcome do buyers want, expressed in their own language? Use the exact words and phrases your interviewees used -- these become your messaging foundation, and they will resonate more than anything a marketing team writes in a conference room.