From YC User Interviews to Purchasing Evidence: A Greater Bay Area Worksheet

Bring back an event, not a compliment

An encouraging response to a university prototype is a useful opening. It does not establish demand. A productive interview should reveal a recent incident, the current workaround, its consequences and who can authorize a next step. That evidence is useful whether a research team is based in Hong Kong, Shenzhen or elsewhere in the Greater Bay Area.

In his YC user-interview lecture, Eric Migicovsky recommends discussing specific past experiences, listening carefully and avoiding a sales pitch disguised as research. The worksheet below is Jimhang Capital's original regional application, not a YC programme.

Source：Eric Migicovsky — How to Talk to Users

Map the people around the problem

Consider a hypothetical tennis-ball collection robot. A coach may care about interrupted lessons. A venue manager may care about storage and maintenance. A budget owner may compare the annual expense with existing arrangements. Interviewing only the coach risks confusing enthusiasm with purchasing authority. Map who experiences the problem, who pays for it and who approves a trial.

Start with one recognizable setting and then interview an adjacent group for comparison. An interview count is a scheduling choice, not a validation threshold. Keep supportive university contacts separate from independently recruited potential customers so their responses remain interpretable.

Keep observation separate from interpretation

Ask the person to walk through a recent task before presenting the proposed solution. With permission, observe the workflow or review a relevant record. Capture only information needed for the research. This compact worksheet makes missing evidence visible.

Test across cities without assuming equivalence

A Hong Kong interview cannot represent Shenzhen, and one Shenzhen customer cannot represent the region. Treat operating hours, available space, procurement steps, language and maintenance arrangements as variables to investigate. When responses differ, examine the workflow before attributing the difference to a city's character.

After each interview, write three separate entries: what the participant said, what the team thinks it means and how that interpretation could be checked. If someone calls the product expensive, identify their comparison and budget responsibility before changing the price. Leave unanswered questions visible.

Let the evidence change the next version

Review repeated obstacles alongside observations that contradict the original hypothesis. Choose one question for the next round: for example, whether a venue will provide an off-hours test window. Permission to observe, a scheduled trial, a signed pilot and an actual payment are distinct events. They should never become one inflated customer count.

When approaching Jimhang Capital or another Hong Kong angel or seed investor, attach an anonymized page explaining the roles interviewed, the product decision that changed and the remaining uncertainty. A clear account of learning makes a more useful discussion possible than a page of uniformly positive feedback.