Every business decision rests on assumptions about customers: what they want, what they will pay, how they decide, what words they use. When those assumptions are wrong, you build features nobody uses, write copy that does not land and price offers that do not sell. Customer discovery is the discipline of testing those assumptions cheaply before you spend on building and marketing.
What discovery is and is not
Discovery is structured learning about the people you want to serve: their situation, their problems, how they currently solve them and what would make them switch. It is not asking people whether they like your idea. People are polite and poor at predicting their own behavior. Discovery focuses on what they have actually done, what they are doing now and what they have already paid for.
The cost of skipping it
Consider the typical path without discovery: three months building, two months marketing, then silence. The same money spent on fifteen conversations in week one would have revealed that the target customer already has a workaround they are happy with, or that the real pain is adjacent to what you planned. Discovery does not guarantee success; it removes the most expensive mistakes early.
Three levels of evidence
- Opinions. What people say they want or would do. Weakest evidence, but useful for vocabulary and direction.
- Reported behavior. What people say they have done, especially recently. Stronger, particularly with specific examples.
- Observed behavior. What people actually do: clicks, purchases, time spent, workarounds you watch them perform. Strongest.
Good discovery moves you from level one toward level three as quickly as budget allows.
Worked example
A consultant planned a course on productivity for freelancers. Before building, she ran twelve 30-minute calls with freelancers from her mailing list. She asked what they had tried in the past year to manage their time, what they had paid for and what they abandoned. Ten of twelve had bought a planner or app and stopped using it within a month. Their real frustration was not planning; it was saying no to low-value clients. She reshaped the offer around client selection and pricing. The first version sold to five of the twelve interviewees.
Common mistakes
- Treating one enthusiastic friend as validation.
- Asking would you buy this instead of what have you bought.
- Starting research after the product is built, when findings are painful to act on.
- Confusing a survey with discovery. Surveys measure what you already know to ask about.
Action Step: Write down the five riskiest assumptions behind your current offer or idea. For each, note which level of evidence you currently have. This list becomes your research agenda.