Useful customer research tests a decision the founder might change. It does not collect compliments to decorate an endorsement pack. Define the target customer, investigate the real problem and preserve an honest research record, including the findings that challenge the original idea.
Begin with a narrow research question
Choose one segment and one uncertain assumption. For example, ask whether a particular business role experiences a costly workflow problem, how often it occurs and what is currently done about it. Avoid starting with would you use our amazing platform. That phrasing introduces the solution before establishing the problem and encourages a polite answer rather than useful evidence.
The official endorsing-body guidance makes market understanding relevant to a viable and scalable proposition. This article's interview process is an editorial preparation method, not a mandatory Home Office template. Its purpose is to produce genuine evidence and a better decision. No fixed number of interviews guarantees endorsement or proves market demand.
Recruit people who resemble the intended buyer
Distinguish users, buyers, budget holders and gatekeepers. They may be different people with different incentives. A student who likes an app is not necessarily the person who pays for an institutional contract. A relative working in the sector can provide an introduction, but a collection of friends' opinions is not automatically representative research. Record how participants were selected and what limitations that creates.
Use consent-based outreach and explain the research purpose. Do not imply that the interview is a sales meeting if it is not, or collect personal information that is unnecessary. Keep contact details privately and separate them from an anonymised analysis. If recording a conversation, obtain appropriate permission and explain storage and use. An endorsement pack does not need to expose every participant's private data publicly.
Ask about behaviour before opinions
Ask the person to describe the last time the problem occurred, the steps taken, the people involved and the cost or consequence. Probe what they tried, what failed and why the current workaround persists. Ask who controls a purchase and what would make a change difficult. Specific examples are more informative than a hypothetical declaration that a product sounds useful.
Then explore the proposed solution carefully. Show a prototype or description without claiming results that have not been achieved. Ask what is confusing, what would prevent use and what evidence a buyer would require. Do not treat a request to keep in touch as a committed customer. Record interest, trial participation, a non-binding letter and a paid contract as different levels of evidence.
Preserve the research chain
Keep the interview date, participant category, questions, notes and consent status. Separate direct observations from your interpretation. If a participant estimates a cost, label it as their estimate rather than an independently measured statistic. If you later summarise results, keep the denominator and method clear. Do not invent a large survey sample or present informal conversations as a statistically representative study.
Create a findings table with theme, evidence, counterexample and business implication. A theme supported by several relevant conversations may justify another experiment, while a strong counterexample can reveal an overlooked segment or barrier. Include negative findings. A business plan that acknowledges what changed after research can be more credible than one pretending every early assumption was correct.
Test willingness to act, not only willingness to praise
Depending on the business, the next step might be a pilot agreement, a prototype test, a procurement conversation or a carefully designed paid experiment. Define the commitment and what it does not mean. A free trial is not recurring revenue; a letter of intent may be conditional and non-binding. Do not translate early interest into forecast sales without stating the conversion assumption.
Set a decision rule before the experiment. What result would justify building, changing the segment or stopping? Record the costs and risks to participants. Avoid experiments that require regulated activity or misleading sales claims before the necessary permissions are understood. Commercial validation must be genuine and lawful, not created solely to manufacture evidence for an immigration application.
Feed the findings into the plan
Update the problem statement, customer definition, pricing assumptions and delivery model. Link each major change to the research that motivated it. If the original segment shows little urgency or budget, do not bury that finding under a larger market-size chart. Investigate a better segment or reconsider the proposition. The founder should be able to explain why the evidence changed the decision.
Use the customer-discovery workbook to design interviews and record the resulting action. Its strongest output is a clearer business model, not a folder of flattering quotes. Keep the research honest, proportionate and attributable, then use it as one part of the wider innovation, viability and scalability evidence rather than claiming it guarantees any immigration outcome.
A worked decision example
An illustrative founder interviews friends who all say they like the idea. A second research round recruits people who actually control the proposed purchase and asks about the last real problem, current workaround and approval process. The findings show that users value the feature but buyers need a different integration before considering a pilot. The founder records both results, changes the experiment and treats expressions of interest separately from contracts. This example does not prescribe a minimum sample or claim statistical representation. It shows why participant relevance and behavioural questions matter. Evidence that changes the product or sales process is often more useful than a collection of compliments, because it gives the founder a decision that can be tested in the next stage.
Official sources checked 15 October 2026. Recheck the linked guidance before applying. This is general information, not an individual immigration assessment.
A useful next step
Use the Founder customer-discovery interview and findings pack to turn this guide into a working plan. Continue with Prove the difference: an innovation and competitor map for founders when you are ready to organise the next part of the file.