Discovery & Research
The part of the job that decides whether anything else matters. Talking to customers, finding real problems, and knowing the difference between what people say and what they do.
Talk to customers weekly
Not a research project once a quarter. One or two conversations every week, forever, so you are never more than seven days from a real user.
Ask about the past
People are unreliable about what they will do and reasonably accurate about what they did. Every good question is anchored to a specific recent incident.
Separate need from solution
"We need a dashboard" is a solution. The need underneath it is usually something you could serve three cheaper ways.
Guides in this track
Common questions
How many interviews are enough?
Stop when you stop being surprised — usually five to eight within a single narrow segment. The number climbs sharply if you mix segments, because five interviews across five customer types gives you one data point per group rather than a pattern.
My company will not give me access to customers. Now what?
Start with what already exists: support tickets, sales call recordings, churn reasons, session replays, and the sales team itself. That is usually enough to form a hypothesis, which is often enough to earn the access you were denied.
Is discovery not just the designer's job?
Design usually runs the sessions well, but discovery is a team activity. The strongest predictor of good product decisions is engineers and designers having watched real customers struggle themselves, rather than reading a summary of it.