How to Find Product-Market Fit: Signals, Surveys and Experiments
Product-market fit means a particular market finds enough value in your product to keep choosing it. A promising interview or successful launch is an opening signal. The useful question is what customers repeatedly do, why they do it, and whether that pattern survives beyond the founding team’s personal attention.
Assess a specific customer segment and use case before making a company-wide claim.
Combine repeat behavior, commercial evidence and customer explanations.
Use surveys to investigate value, not to award yourself a pass mark.
Run one focused experiment and define the decision before seeing the result.
Start with one customer and one recurring job
Write a narrow working definition: “Independent accounting practices use our tool to collect client documents before monthly close.” This names a buyer, a task and a natural usage cycle. “Small businesses need automation” leaves too many possible customers and benefits to test. Record exclusions too: a seasonal tax preparer may have a different problem from a practice doing monthly bookkeeping.
Michael Seibel’s product-market-fit essay emphasizes persistent customer demand and warns against treating funding or employee count as proof. Turn that distinction into a practical evidence rule: every claim about fit should identify the customer group, observation window and behavior behind it. Keep a separate hypothesis for adjacent segments until their evidence becomes comparable.
Measure return to value, not just return to the screen
Choose the action that represents useful work completed. For the accounting example, logging in is weaker evidence than a practice completing a client document collection. Define the first useful event, the repeat event and the interval in which repetition matters. A monthly workflow should not be judged by whether customers open it every day.
Amplitude’s retention documentation makes those event definitions explicit. Compare cohorts that started in similar periods, and give each cohort enough time to complete the cycle. Show counts beside percentages: eight returning practices out of ten tells a different story from eight hundred out of a thousand. Also record paid renewals, support effort and cancellations. Strong usage supported by unsustainable manual work needs a different response from strong usage with manageable delivery costs.
Use a product-market-fit survey to explain the pattern
Ask experienced users how they would feel if the product disappeared, what benefit they get, who else has that need, and what prevents fuller use. Include people who stopped using it when possible. Note invitation count, response count, customer role and usage history. A survey sent only to the founder’s favorite accounts will mostly confirm what the founder already believes.
Rahul Vohra’s first-person Superhuman account describes using disappointment responses and customer segments to guide product changes. Treat that approach as a diagnostic tool, not a universal threshold. A small, enthusiastic respondent group cannot establish broad demand. Compare their answers with actual behavior and ask why nonresponders or departed users may differ.
Worked example: test a friction point before adding features
Consider a hypothetical document-collection product with twenty practices in its first mature monthly cohort. Twelve complete a second collection. Interviews with the other eight suggest that inviting clients takes too long. The team suspects onboarding friction rather than missing reporting features. It writes down that hypothesis before changing the product.
For the next cohort, the team provides an import template and clearer invitation preview. It measures time to first completed collection, repeat collection next month and support minutes per practice. Suppose fourteen of twenty new practices repeat. That improvement is encouraging, but it is not proof of causation: the cohorts may differ. Check acquisition source, practice size and founder assistance, then repeat or run a suitable controlled comparison.
If the new workflow reduces support effort but repeat behavior stays flat, keep the operational improvement while revisiting the value hypothesis. Do not quietly change the success criterion to make every release look successful.
Run a review that produces a product decision
Bring one page to a regular product review: segment definition, cohort chart, recent customer explanations, costs of serving the segment and the most important unresolved question. Separate evidence from interpretation. “Six accounts renewed” is an observation; “pricing is right” is an interpretation that also requires information about discounts, alternatives and margins.
Finish by choosing one action: deepen the current use case, remove a specific barrier, change the segment, or pause expansion while you learn. Assign an owner and a date for the next evidence review. When the product has a clearer customer and repeatable benefit, connect that learning to the go-to-market plan. Distribution experiments work better when you know whose problem you are promising to solve.
Keep cohort definitions stable across reporting periods.
Interview both retained and departed customers.
Record discounts and founder assistance alongside usage.
State what result would change the team’s current belief.
Frequently asked questions
Yes. Revenue may come from a few bespoke projects, discounts or founder relationships. Examine whether a defined customer segment repeatedly receives value and whether the delivery model can support the business.
Match the schedule to how quickly customers can experience the core benefit. Avoid surveying people before they have useful experience, and track which cohorts and respondents are included.
Choose a repeat-value measure suited to the product, then interpret it alongside commercial and qualitative evidence. A single metric cannot explain customer motivation, delivery cost and the durability of demand.
- The Real Product Market Fit - Y Combinator
- Build a retention analysis - Amplitude
- How Superhuman Built an Engine to Find Product Market Fit - First Round Review — Rahul Vohra, Superhuman founder