Short answer: validate the segment, situation, alternatives, proof needs and next action before you treat a market as ready. The right method depends on the decision and cost of error.
There is no universal number
The common question is how many interviews are enough. The honest answer is that the number depends on what you are testing. A narrow hypothesis in one clear segment may need fewer interviews than a broad market question across several roles, countries or buying situations.
The goal is not to reach a ritual count. The goal is to gather enough evidence to make the next decision with lower uncertainty. Sometimes that means five focused interviews. Sometimes it means fifteen interviews plus desk research, sales evidence and a survey.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
Start with the hypothesis type
A problem hypothesis asks whether a segment experiences a painful situation. A segment hypothesis asks which group should be approached first. An offer hypothesis asks whether a promise and proof structure make sense. A buying-process hypothesis asks who participates and what blocks the decision. Each type needs different evidence.
If the hypothesis is broad, interviews may reveal too many directions. Narrow it first with market mapping. Then use interviews to test the parts where human reasoning matters.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
Segment clarity changes the count
Five interviews with similar buyers in the same situation can produce useful patterns. Ten interviews split across unrelated audiences may produce noise. Before counting interviews, define the segment by situation, role, alternative and urgency.
If you need to compare segments, treat them separately. Four interviews in segment A and four in segment B are not eight interviews for one hypothesis. They are two small evidence pools that may need different interpretation.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
Look for saturation, but define it carefully
Saturation means new interviews stop changing the pattern materially. It does not mean every detail repeats. In launch research, practical saturation appears when the same triggers, alternatives, objections and proof requirements keep returning within a segment.
Do not stop only because the team heard something it likes. Stop when additional interviews are unlikely to change the next decision, or when the remaining uncertainty should be tested with another method.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
Risk and reversibility matter
A reversible landing page message test may require less evidence than a product investment, market expansion or enterprise positioning change. When the cost of error is high, increase evidence depth and combine methods. Interviews can explain why, while surveys, analytics, sales data or live offer tests can add scale.
This is especially important for healthcare, B2B infrastructure, manufacturing and education products where trust, compliance or switching costs shape buyer behavior.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
Use a stopping rule
Before the first interview, define what will happen if patterns repeat, if they conflict, or if the segment appears weak. A stopping rule might say: after six interviews in one segment, continue only if new objections appear; run three more if two roles disagree; switch methods if respondents cannot recall recent behavior.
The final report should state the number of interviews, segment definition, recruitment source, evidence limits, repeated patterns and unresolved questions. That makes the result usable without pretending it is a statistical forecast.
- Write the decision this section affects.
- Separate verified facts, strong signals and open hypotheses.
- Choose the smallest next test that could change the plan.
FAQ
Are 5 interviews enough?
Sometimes, for a narrow qualitative hypothesis in one segment. They are not enough to quantify demand across a market.
Can AI replace interviews?
No. AI can help synthesize sources and prepare hypotheses, but it cannot provide fresh respondent experience by itself.
When should we add a survey?
When the qualitative hypotheses are clear enough that structured answer options will mean something.
Conclusion and next step
Good research does not promise certainty. It helps the team spend the next unit of time, design attention or media budget with a clearer view of audience, demand, offer and risk. If the next decision is expensive, start with a diagnostic and only deepen the research where the evidence boundary requires it.
Useful next pages: Market & Offer Diagnostic, Audience Research Sprint, Glossary and Research proof.