Customer discovery is the process of testing assumptions about a specific customer, their problem, current behavior, and willingness to change before committing to a solution.
It’s one of the most important (and most ignored) steps in startup validation. Instead of guessing what people want, you go directly to them to understand their real problems, how painful those problems are, and whether they’d actually pay to solve them.
Not there yet? Start one step back with how to come up with a startup idea.
Ready to put it into practice? Follow the practical customer discovery guide for a recruiting message, interview script, evidence log, and decision rules.
A startup can execute well and still fail if it solves a problem customers do not consider important.
Customer discovery helps you avoid this by answering the questions that actually matter:
If you skip this step, you’re basically building in the dark.
Here’s how most successful founders approach it:
Talk to people in your target market and find out what problems they actually have.
Confirm that the problem is real, painful, and worth solving.
Test whether your idea actually solves their problem better than existing solutions.
See if they would actually pay for it (or switch from what they’re using now).
Even smart founders mess this up. Here are the biggest ones:
While customer discovery requires talking to real humans, Roast Rocket helps you prepare for it.
Before you spend hours interviewing people, use our AI to:
Think of it as the pre-flight check before you go talk to real customers.
Customer discovery is the hands-on process of talking to real people. Idea validation (like Roast Rocket) is the strategic stress-test you do first to make sure you’re even asking the right questions.
There is no universal target. Continue until qualified interviews stop materially changing your understanding, recurring patterns are clear, and the important user and buyer roles are represented. Use the practical customer discovery guide to plan and evaluate each round.
Yes. That’s actually the point. You should validate the problem and solution before writing code or designing anything.
As early as possible — ideally before you write a single line of code or design a single screen. The whole point is to find out whether the problem is real and worth solving before you spend months building a solution nobody wants.
Focus on past behaviour, not hypotheticals. Ask “When was the last time you ran into this problem?” instead of “Would you use this?”. People are great at remembering what they did and terrible at predicting what they’ll do. Rob Fitzpatrick’s book The Mom Test is the canonical playbook.
Find weak spots, sharpen your pitch, and walk away with a clear next step — free.
LAUNCH YOUR ROCKET