Redko Labs editorial notes · Updated October 13, 2026
Choose a narrow first audience
“Researchers” is too broad for a useful first interview round. Choose a role, an organization type, and a recurring task. For example, a computational biologist preparing a reproducible analysis has different constraints from the laboratory director approving software. Write down your best current guess about the user, the buyer, and the person who controls access to data. Keep these as hypotheses until conversations provide evidence.
Use these six questions
Ask: When did you last do this task? Walk me through what happened. What tools or workarounds did you use? Where did the process slow down or fail? What did that delay change for your team? Who decides whether to try a new tool? Let the interviewee finish describing the current process before showing your solution. If possible, ask them to demonstrate a permitted example rather than reconstructing everything from memory.
Record evidence separately from interpretation
Use a simple table with five columns: interview role, recent example, current workaround, consequence, and next step. In a separate column, write what you think the observation means. “Spent two afternoons reconciling exports” is a reported observation. “Will pay for our platform” is an inference that needs a separate test. Ask permission before recording, and remove personal or confidential information from notes shared outside the team.
Look for disagreement, not just repetition
If users describe different problems, compare their context before combining their answers. A task performed once a year may not justify the same product as a daily bottleneck. Ask a skeptical person to explain why the current workaround is good enough. That answer can reveal switching costs, procurement requirements, or a constraint your prototype ignores.
End with a small, specific commitment
Ask for one next step that fits what you learned: an introduction to the buyer, a review of a mock-up, or a bounded test using permitted data. Record who will do what and when you will follow up. NSF I-Corps treats direct customer discovery as part of research commercialization; an interview itself is not a sale. Bring your interview notes and unanswered questions when discussing founder support with Redko Labs.
Common questions
How many interviews prove demand?
There is no universal interview count that proves demand. Look for a repeatable problem, a clear buyer, and progressively stronger commitments.
Do I need a finished prototype?
No. Start by understanding the current workflow. A prototype becomes useful when it helps test a specific question.
Further reading
Program details are proposals until confirmed in an offer. These notes are general information, not legal, investment, medical, or immigration advice.