You can do customer research before you have customers by studying people who recently experienced the problem, not people who already bought your solution. Review existing evidence, interview suitable participants about real events, observe current alternatives and run small tests that require meaningful behaviour. Keep problem evidence, solution feedback and demand evidence separate; none alone proves a market.

Define the decision and risky assumption

Start with the choice you need to make: which problem to pursue, which audience to prioritise, which workflow to prototype or whether to continue investing. Then list the assumptions that could reverse that choice.

Rank assumptions by uncertainty and consequence. “Finance leads at multi-site charities encounter reconciliation delays monthly” is more testable than “people dislike the process”.

Do not ask strangers whether they like an undeveloped idea. That invites speculation.

Use existing evidence before recruiting

Review credible sources that describe the environment: public guidance, industry records, procurement documents, community discussions, support questions, job adverts and competitor documentation. Record provenance and publication date.

This desk research helps you learn terminology, constraints and possible alternatives. It does not reveal private priorities or prove willingness to buy. Treat online comments as situated accounts, not a representative sample.

Build a simple evidence log with the claim, source, context, confidence and unanswered question.

Recruit for experience, not enthusiasm

Define participants by a recent event they can describe. If you are exploring appointment reminders, recruit people who recently managed missed appointments or evaluated a reminder process—not a broad group who say they like productivity tools.

Possible routes include professional communities, associations, events, referrals and direct outreach. Explain why you are contacting the person and that the conversation is research, not disguised sales. The recruitment guide covers screeners, incentives and bias.

Friends and personal contacts can help pilot a guide, but convenience should not become the whole sample.

Interview the problem and current alternatives

Ask participants to reconstruct the last relevant event:

  • What triggered it?
  • How did they handle it?
  • Who else became involved?
  • What was difficult or costly?
  • Which alternatives did they try or reject?
  • What happened afterwards?
  • What would make them change the current approach?

Follow evidence, not a pitch script. Avoid “Would you use a tool that…?” and “How much would you pay?” without a concrete offer and context. Hypothetical approval is easy; changing behaviour is harder.

The competitive alternatives method helps reveal whether the real rival is software, a spreadsheet, a service, an internal workaround or inaction.

Observe the work where possible

What people do can differ from what they remember. With informed permission, observe a relevant task or ask a participant to walk through a redacted artefact. Note sequence, hand-offs, tools, delays and workarounds.

Protect confidential and personal information. Agree what may be shown, recorded or retained, and collect only what the research requires. If the context involves vulnerable people, regulated data or material risk, obtain appropriate specialist guidance.

Observation gives depth in a specific context. It does not establish how common the behaviour is.

Test propositions through meaningful behaviour

Once problem evidence is credible, present a clear proposition or prototype. Ask participants to complete a realistic task, compare alternatives or explain the offer in their own words. A landing page can test comprehension and acquisition mechanics; it does not by itself prove product value.

Choose a commitment proportionate to the stage: joining a relevant waitlist, sharing required information, attending a structured pilot discussion or paying for a genuine offer. Explain what exists now, what is experimental and what the participant will receive.

Use the guide to an evidence-based value proposition without customer proof to keep early claims honest.

Separate evidence levels

Maintain four columns:

  1. Problem evidence: real events and consequences.
  2. Solution evidence: whether people understand and can use the concept.
  3. Demand evidence: meaningful steps towards adoption or purchase.
  4. Delivery evidence: whether the team can create the result reliably.

Do not move an encouraging quotation from the first column into the third. Look for contradictory evidence and alternative explanations. Decide in advance what would cause you to change audience, problem or approach.

Use a pre-customer research checklist

Before drawing a conclusion, confirm that you have:

  • named the decision and riskiest assumptions;
  • logged relevant existing evidence and its limits;
  • recruited people with recent, direct experience;
  • included more than one route or perspective;
  • asked about events and alternatives before pitching;
  • handled consent and data responsibly;
  • tested understanding or behaviour with an honest artefact;
  • separated problem, solution, demand and delivery evidence; and
  • recorded counterexamples and the next uncertainty.

Illustrative example

Imagine a founder exploring a service that prepares small charities for grant reporting. She reviews public funder requirements, then interviews operations and finance staff who completed a recent report. Several use spreadsheets successfully; the most acute difficulty appears when different programmes code expenditure inconsistently.

She tests a paid diagnostic focused on that hand-off rather than building the original dashboard idea. This is illustrative, not evidence that the market or offer will succeed.

Recognise the limitations

Pre-customer samples often overrepresent accessible, interested people. Interviews reveal remembered experience, prototypes omit real operating conditions and waitlists may reflect curiosity. Payment is stronger evidence of commitment but still does not prove retention or scalable delivery.

Research reduces uncertainty in stages. It cannot remove founder judgement or guarantee a commercial outcome.

Frequently asked questions

Can competitor customers be research participants?

Yes, when recruited ethically and asked about their own experience rather than confidential competitor information. Do not misrepresent your identity or encourage breaches of agreement.

Should I use a survey before interviews?

Usually, exploratory interviews or observation help define answer choices and language first. A survey is useful when you have a clear measurement question and a sample capable of answering it.

How many people should I interview?

There is no magic number. Plan by decision and comparison group, then continue until the important stages have concrete evidence and additional sessions stop materially changing the model.

Your useful next step

Write one risky assumption as a testable sentence. Define the recent experience a person must have to provide evidence, then draft a neutral invitation for three participants outside your immediate circle. Do not mention the proposed feature in the opening request.

Publication note: This guide is published by To-Do Growth and reviewed under our Editorial Policy. Illustrative examples are labelled and are not presented as customer case studies.