Practical founder guide
Customer interview worksheet
Separate accounts of real behaviour from polite encouragement.
“That sounds useful” is encouraging, but it does not establish that someone has the problem your idea is meant to solve. A useful early conversation uncovers a specific situation, the person’s current response, and the consequences of leaving the problem unresolved.
This editorial worksheet helps you preserve the difference between what a person said, what happened, and what you inferred. It does not prescribe a sample size or establish market demand on its own. The worked example is hypothetical.
Define the uncertainty first
Complete this sentence:
I am uncertain whether [a specific group] experiences [a specific problem] during [a situation].
For example: “I am uncertain whether freelance designers working with repeat clients lose track of overdue invoices because their records are spread across several tools.” This is narrower than “Do freelancers need a finance app?” It also leaves room to discover that the assumed cause is wrong.
Choose participants because they have recent experience of the situation, not because they are likely to praise your idea. State your role honestly. Ask permission before recording, allow people to skip questions, and avoid collecting names or client details you do not need.
Open without pitching
A short introduction could be:
I’m exploring how people handle [task]. I’d like to understand your experience rather than sell you something. Could we discuss a recent example? You can skip anything you prefer not to share.
If you know the participant, acknowledge that the relationship may affect how direct they feel able to be. Do not pressure them to prove that a problem exists. YC’s user-interview lecture and transcript recommends examining experiences and existing attempts to solve a problem. The prompts below apply that general principle to a task record.
Recover a sequence, not a sales reaction
You need not read these prompts in order. Follow the participant’s example and ask for clarification when a step is unclear.
| Learn about | A prompt to try | Avoid assuming |
|---|---|---|
| Context | What was happening when this task came up? | That it is frequent or urgent |
| Sequence | Could you walk through what you did from beginning to end? | That your workflow is their workflow |
| Tools | What did you use, and where did you keep the information? | That replacing a tool is the desired outcome |
| Friction | Which step took more effort than expected? | That every inconvenience needs a product |
| Consequence | What happened because of the delay or mistake? | A cost without supporting evidence |
| Alternatives | Did you try another way? What happened? | That no existing solution exists |
| Priority | What other work competed for your attention? | That this was the most important problem |
If someone imagines a future feature, return to a recent example before treating the suggestion as evidence. Feature requests can reveal a problem, but they do not automatically determine the solution you should build.
Copy this evidence record
Use a participant code rather than a full name if identity is unnecessary. Distinguish an exact quote from a paraphrase, and obtain permission before publishing identifiable material.
| Field | Notes |
|---|---|
| Participant code and relevant context | ____________________ |
| Recent event discussed | ____________________ |
| Steps in the current workflow | ____________________ |
| Tools or workarounds already used | ____________________ |
| Friction and consequence | ____________________ |
| Exact quote or clearly marked paraphrase | ____________________ |
| Your interpretation | ____________________ |
| Other explanations that could fit | ____________________ |
| What remains unknown | ____________________ |
| A follow-up question or test | ____________________ |
Keep interpretation separate. “Uses three tools” is an observation if described or shown. “Needs an all-in-one app” is an inference. The same observation might support an integration, a process change, or no product at all.
Classify the signal without inventing certainty
These categories are a practical sorting aid, not a scientific score:
- Opinion: a preference or reaction, including a compliment.
- Reported episode: a specific account of an event; useful detail, but still a self-report.
- Observed workflow: an example or demonstration you were allowed to see; it may still be atypical.
- Commitment: an agreed next action, such as a scheduled trial. Track whether it is completed separately.
- Observed outcome: what happened in a test or delivery, with its context and limitations.
One interview or purchase does not establish a market. Different business questions require different evidence, and circumstances can change.
Worked example: a hypothetical invoice conversation
A designer describes checking a spreadsheet on Fridays and searching email to confirm payments. You see a redacted example with permission. The designer reports one delayed reminder last month but cannot estimate its financial effect.
| Observation | Interpretation to test | Still unknown |
|---|---|---|
| The weekly review uses a spreadsheet and email | Moving between records may add work | How often other designers face this |
| One delayed reminder was reported | A checklist might help | Whether the delay caused meaningful cost |
| A further conversation is scheduled | There is willingness to continue talking | Whether the person will attend or use a solution |
This is a fictional example. It does not claim that customer research took place or that a product is wanted.
Turn the conversation into a next decision
- The strongest observation was…
- It challenged or supported this assumption…
- An alternative explanation is…
- Before spending more on a solution, I need to learn…
- My next test is…
Strategyzer’s Learning Card separates observations, interpretation, and subsequent action. Maintain that distinction across interviews. Contradictory accounts can reveal different groups or a question that was too broad.
If you identify a task worth testing, continue to the MVP scope worksheet. If organizing the work is the obstacle, use the founder action plan to choose one habit to adjust.