How to Measure Customer Discovery Progress
Measure customer discovery by the evidence it produces, not by activity: how many conversations were with people who match your target customer, how many confirmed the problem with a recent example, how many made a commitment such as a pilot, a deposit or an introduction, and what decision the team made as a result.
By Javier Alonso, Founder (CEO & CTO). Updated .
Why the number of interviews is not enough
"We spoke to 40 people" sounds like progress, but it says nothing about who those people were or what the team learned. Friends, other founders and people outside the target market are easy to reach and tend to be polite, so the count rises while the evidence stays weak.
Discovery is progressing when each week the team knows more about one specific customer: what problem they have, how they handle it today and whether they would act to solve it. The measures below track that.
Five measures that show real progress
- Qualified conversations: conversations with people who match the ideal customer profile. Agree the criteria before counting.
- Problem confirmed with a recent example: the person described the last time the problem happened, what it cost them and what they did about it. Opinions about the future do not count.
- Existing workarounds: how many already spend time or money on a fix, such as a spreadsheet, an agency or a competitor. A problem nobody works around is rarely urgent.
- Commitments: a follow-up meeting, an introduction to a colleague, a pilot, a pre-order or a deposit. Commitments cost the customer something, which is why they are stronger than compliments.
- Decisions: what the team changed because of what it heard, for example narrowing the customer, dropping a feature or stopping the idea.
A simple weekly view
Keep one row per week with the five numbers and one sentence on what the team now believes. After three or four weeks the pattern is usually clear: either confirmed problems and commitments rise as the customer gets narrower, or they stay flat, which is itself a useful result.
The customer interview questions guide covers how to run each conversation so the answers are usable, and how to validate a startup idea covers what to do with the evidence.
Example
A founder building scheduling software for physiotherapy clinics holds 12 conversations in two weeks. Only 7 are with clinic owners; the rest are therapists who do not choose software. Of the 7, 5 describe a recent double-booking problem and 4 already pay a receptionist to fix schedules by hand. Two agree to a free pilot. The founder narrows the target to owners of clinics with three or more therapists and stops interviewing individual therapists. The weekly view shows 7 qualified conversations, 5 confirmed problems, 4 workarounds, 2 commitments and 1 decision.
Common mistakes
- Counting every conversation, including friends and other founders.
- Treating "I would use that" as confirmation. It is an opinion about the future.
- Pitching the solution during the interview, which turns discovery into a sales call.
- Measuring only at the end, when there is no time left to change direction.
For your role
- Founders
- Report the five numbers to yourself every week. If confirmed problems and commitments are flat after a month, change who you are talking to before you build more.
- Mentors
- Ask for the last week's qualified conversations and the decision they led to. It moves the session from opinions to evidence.
- Accelerators & incubators
- The same five numbers, collected the same way for every team, make discovery comparable across a cohort.
Primary sources
- The Mom Test — Rob Fitzpatrick
- Startup School: How to Talk to Users — Eric Migicovsky
- Principles — The Lean Startup
Common questions
- How many customer interviews are enough?
- There is no fixed number. Many teams see clear patterns after 10 to 20 qualified conversations with one customer segment. Stop when new conversations stop changing what you believe, not when you reach a target.
- What counts as a qualified customer conversation?
- A conversation with someone who matches your target customer criteria and has the problem you are studying, agreed before you start counting. Friends, investors and other founders usually do not qualify.
- What is the strongest signal in customer discovery?
- A commitment that costs the customer something: time for a follow-up, an introduction, a pilot, a pre-order or money. Compliments and "I would use it" are the weakest signals.
Put it into practice
Turn interview notes into a structured validation plan with the evidence in one place.