4 Customer Discovery and Validation
Learn how to turn venture assumptions into testable hypotheses, gather useful customer evidence, and decide whether to continue, revise, or abandon an idea.
Purpose of
is a systematic way to learn how a specific group experiences a problem, what they do about it, and what might persuade them to adopt or buy a solution. Its purpose is to reduce uncertainty with , not to collect compliments for an idea. Interviews help explain people’s context and past behavior, while experiments can show whether they take meaningful action.
Research is iterative: form a hypothesis, test it, interpret the , and revise the next test. Interviews and experiments are complementary parts of this ongoing learning process.
Turn assumptions into hypotheses
Start by listing the beliefs that must be true for the venture to work. Turn important beliefs into a : a statement specific enough to investigate rather than a broad assumption such as ‘Retailers need better software.’ For example, ‘Independent neighborhood grocers lose sales because they run out of popular items’ identifies a group, a problem, and a possible consequence.
Prioritize assumptions that are both uncertain and consequential. Ask whether the venture would need to change direction if an assumption proved wrong. Business-model assumptions can concern desirability, feasibility, and financial viability.
Common assumption categories include:
Customer and problem: A particular group experiences a problem often and considers it important.
Solution and adoption: The proposed approach addresses the problem, and customers can adopt it.
Business viability: Customers will pay, or another sustainable source of revenue exists.
Feasibility: The venture can deliver the solution with available technology, skills, and resources.
Recruit and interview customers
Define the before recruiting. For example, recruit store owners who manage inventory themselves rather than ‘anyone in retail.’ Seek people who fit the segment and have direct experience with the problem. Friends and family may be convenient, but their supportiveness or lack of relevant experience can distort the .
Prepare a short guide, but allow the conversation to follow relevant details. Begin with the participant’s role and situation, then ask for a recent, specific example of the problem. Questions about past events are more useful than asking participants to predict whether they would approve of a proposed idea.
Useful prompts include:
‘Tell me about the last time this happened.’
‘What did you do next?’
‘How often does this occur?’
‘What does it cost you in time, money, or missed opportunities?’
‘What tools or workarounds do you use now?’
‘Who else is involved in choosing or paying for a solution?’
‘What makes it difficult to change the current process?’
Listen for concrete stories, existing workarounds, consequences, and decision processes. Avoid leading questions such as ‘Would you use our app?’ or ‘Don’t you think this would save time?’ Hypothetical praise is weaker than specific past behavior. Early interviews should focus on understanding people’s world, not selling a concept or asking them to design the product.
A practical interview includes a welcome and brief explanation, questions about relevant experience, exploration of the problem and current responses, and a closing invitation to share anything missed or continue the conversation. Record what the participant actually said and did, and keep observations separate from your interpretation.
Test behavior with experiments
Interviews can explain needs and context, but people may not be able to predict what they will do. Use an when the question is whether customers will act—for example, try a prototype, share information, sign up, schedule a demo, or pay. Choose the smallest, lowest-cost test that can provide credible for the assumption. Possible tests include observing a workflow, showing a simple prototype, testing a landing page, or offering a limited pilot.
Before running an , write down five elements:
Hypothesis: What do you believe about a defined customer group?
Test: What will you do to expose that belief to ?
Measure: What observable behavior or result will you record?
Threshold: What result will count as enough to continue, revise, or stop?
Next step: What decision will you make based on the result?
For example, a team believes independent grocers will pay to reduce stockouts of popular items. It interviews store owners about recent stockouts and current tracking practices, then offers a simple weekly inventory report to a small set of qualifying stores. Before testing, the team defines a meaningful signal, such as several stores agreeing to a paid pilot, rather than counting general interest as validation. If store owners describe the problem but decline the pilot, the team investigates whether the problem is not urgent, the offer is unsuitable, or the buyer lacks authority or budget.
Interpret and choose a next step
Review notes across participants. Look for recurring problems, workarounds, and costs, and notice where accounts conflict. For an , check whether participants took the action it was designed to measure. Compare findings with the original hypothesis rather than only with what the team hoped to hear. Objective listening and repeated testing help counter founder bias, which can lead teams to hear the answers they want.
Interpret within its limits. A small set of interviews can reveal patterns and sharpen questions, but it does not establish how common a problem is across an entire market. A positive response to a concept is not the same as adoption or willingness to pay. Record uncertainty, revise the hypothesis when needed, and test again.
Build only as much as necessary to learn. Each should connect clearly to the assumptions behind the proposed value and business model. Use the results to decide whether to continue, revise, or abandon an assumption.