Evidence Collection: Testing Your Hypotheses With Web Research and Interviews

Your critical hypotheses are the assumptions your startup depends on. Evidence Collection is where you stop treating them as plausible ideas and start looking for signals that can support, challenge, or reshape them.

The page offers two ways to learn: Web Research, which looks for relevant public signals across the internet, and Interviews, which lets you collect answers directly from people who resemble your target customers. You can use either method first, but the strongest validation usually uses both.

Public discussion helps you understand the landscape quickly. It can reveal the language people use for a problem, existing alternatives they complain about, objections to a category, and the situations in which a pain point becomes urgent. It is a useful way to find patterns before you ask people to give you their time.

Interviews add something public sources cannot: direct answers from the customers you want to serve. They let you probe for context, compare experiences, and test whether a pattern you saw online is real for your specific segment. Treat web evidence as a way to form sharper questions, and direct customer evidence as a way to test those questions properly.

Choose Web Research when you want zigzag to look broadly for signs that each hypothesis may be true or false. The research looks for relevant public discussion across sources such as Reddit, X, review sites, industry-specific forums, news, and other publicly available material that fits your project.

The result is organized by hypothesis so you can see which signals support it, which challenge it, and where the evidence came from. Once research is complete, the source count is expandable: open it to inspect the public sources that were taken into account rather than relying on a summary alone.

Use this material critically. A lively public discussion is not the same as proof that your target customer will buy. It is evidence to weigh alongside the quality, relevance, and recency of the source — and a prompt for the interviews or tests you should run next.

Choose Interviews when you are ready to invite potential customers. Add the email addresses of people who match the customer segment you want to validate, then send them a written questionnaire or use the interview guide for a live conversation. The interview guide belongs here because it supports the direct-customer path, not web research.

By default, a written questionnaire focuses on the three hypotheses with the highest priority scores. That keeps the request short and focused. You can switch to the full questionnaire when you need coverage across every hypothesis, but expect longer questionnaires to receive fewer completed responses.

Invite people because they fit the segment, not because they are likely to agree with you. The aim is to learn what they do today, what is painful enough to matter, and how their real behaviour compares with the assumptions in your project.

Evidence Analysis becomes available as soon as web research has completed or at least one interview has been completed. A green checkmark on the Validation Framework shows that there is enough evidence to begin reviewing the results.

The analysis is updated automatically as new evidence arrives. It does not treat public research and interviews as two unrelated scorecards: each hypothesis receives one combined view of the evidence, with the source context kept visible so you can understand what shaped the result.

In Evidence Analysis, start with the hypotheses that carry the most risk for the business. Look at the supporting and challenging signals, inspect the underlying interview answers or public sources, and decide whether the hypothesis is confirmed, at risk, still needs more proof, or remains unscored.

Do not wait for perfect certainty. The goal is to reduce the chance of building around a weak assumption. If the evidence is mixed, collect another focused round. If it consistently challenges a high-priority hypothesis, update your Lean Canvas and decide whether your customer, problem, or solution direction needs to change.