How many customer interviews are enough for market validation?

The short answer: there is no universal interview number

Five interviews may uncover important problems. Ten may reveal recurring patterns. Twenty may expose differences that were invisible in the first ten. None of these numbers automatically settles the commercial question.

The appropriate number depends on who you interview, how narrowly the target group is defined, how many distinct groups must be understood, what assumption you are testing, how much new information each interview adds and how difficult the decision would be to reverse.

Interview scope is one part of a wider validation process. Read how to validate a business idea before you build.

Interviews provide depth, not a statistical forecast

A qualitative interview sample is normally selected because participants can provide relevant information—not because every member of the wider market has an equal probability of being selected. Twenty interviews with the wrong people remain twenty irrelevant interviews.

Interviews can reveal whether people experience the proposed problem, how they deal with it today, what triggers them to act, who influences and pays for a solution, and what prevents change. Their purpose is to assess the strength, nature and conditions of demand—not to forecast market-wide revenue from a small qualitative sample.

“Enough” depends on the decision

Five conversations may be enough to repair an early interview guide. A substantial development investment demands a higher standard: recurring patterns, relevant exceptions and an understanding of differences between the people involved in buying and using the solution.

Do we have enough relevant evidence to make the next decision, and do we understand what remains uncertain?

Five factors that determine how many interviews you need

1. How narrowly the target group is defined

“Small-business owners” may include materially different situations. A useful definition may specify industry, company size, geography, buying role, current solution and the event that creates a need to act. If the group is too broad, apparent contradictions may simply reflect different markets being treated as one.

2. How many distinct segments must be understood

A business purchase may involve users, technical evaluators, managers, budget holders and procurement teams. If three groups must be understood, 15 interviews may be three small groups of five rather than one sample of 15. Analyse them separately before drawing conclusions across them.

3. How focused the commercial question is

“Will people buy this?” is too broad for a reliable interview answer. More useful questions concern the importance of the problem, the current alternative, the trigger for change, the buyer, approval criteria and barriers to adoption.

4. How much new information each interview adds

A new interview may reinforce a pattern, reveal a meaningful exception, explain a difference, contradict an assumption or add nothing that changes the present interpretation. The relevant test is whether it changes or strengthens the decision—not whether another quotation can be collected.

5. How consequential the decision is

A reversible experiment requires less evidence than a large, difficult-to-reverse investment. Consider development cost, time to market, opportunity cost, regulatory commitments and the cost of changing direction later.

What research says about interview sample size

Qualitative research does not offer one universally correct sample size. One widely used concept is saturation: the point at which further data collection produces little or no new information relevant to the analysis. Researchers have defined and applied the term in different ways.

A systematic review by Monique Hennink and Bonnie Kaiser examined 23 empirical or model-based studies. In studies with relatively homogeneous populations and narrowly defined objectives, saturation was commonly reported within 9–17 interviews. More varied populations and deeper examination of meanings could require larger samples. Hennink and Kaiser, Social Science & Medicine, 2022.

That range is not a universal market-validation rule. The reviewed studies addressed qualitative research designs, not a guarantee of future purchasing.

A complementary principle is information power. Malterud, Siersma and Guassora propose that fewer participants may be required when a sample contains highly relevant information for a focused study. Their model considers the study aim, sample specificity, use of established theory, interview quality and analysis strategy. Malterud, Siersma and Guassora, Qualitative Health Research, 2016.

A systematic analysis of interview-based studies also found that sample-size justifications and uses of “saturation” vary considerably. This supports explaining how sufficiency was assessed instead of citing a number without context. Vasileiou et al., BMC Medical Research Methodology, 2018.

How Test Before You Build scopes interview projects

The appropriate scope begins with the commercial decision, not a preferred number. The standard service includes 20 interviews and provides a clear scope for target definition, prospect selection, recruitment, interviewing, transcription, analysis and a BUILD, CHANGE or KILL recommendation.

In broader projects, the interview number, recruitment approach, schedule, deliverables and price are scoped separately. A project may include substantially more than 20 interviews when the research question requires them.

The principle is not that 20 is always sufficient—or that 50 or 100 is automatically better. The scope should cover the relevant groups and answer the agreed questions without collecting interviews that add little decision value.

Before outreach begins, Dr. Kari identifies suitable prospects using the agreed target definition. The client approves the target list. Depending on the project, participants may be realistic potential buyers, users or both.

What a well-scoped interview project can provide

• recurring customer problems and their consequences;
• current solutions, workarounds and triggers for action;
• buying criteria, approval processes and barriers to change;
• differences between relevant customer groups;
• assumptions that are supported, weakened or unresolved.

In the standard service, the client receives the 20 interview transcripts, evidence analysis, a written report, recommended changes and a BUILD, CHANGE or KILL recommendation. Larger projects can cover additional segments, markets or questions. The recommendation is a reasoned interpretation of the evidence, not a guarantee of commercial success.

What customer interviews can reveal about market demand

Customer interviews reveal whether realistic potential buyers or users experience the problem, how important it is, what they do about it today, why they might change and what could prevent them from buying or adopting a new solution.

When these signals recur across a carefully selected group, they provide meaningful evidence for deciding whether to BUILD, CHANGE or KILL the idea before committing to development. No validation method can guarantee future sales, but customer interviews can substantially reduce the risk of building around assumptions that the intended market does not share.

A practical stopping rule for customer interviews

1. Check participant relevance.Are they realistic potential buyers, users or agreed members of the buying process?
2. Review each segment separately.Are the important groups sufficiently represented for the decision?
3. Track new decision-relevant evidence.Record new themes, support, exceptions, contradictions and changes to the interpretation.
4. Investigate contradictions.They may reveal separate markets, different urgency or a missing purchasing condition.
5. Match confidence to the decision.Evidence sufficient for a prototype may not be sufficient for a full development programme.
6. State what remains uncertain.A sound recommendation makes unresolved questions explicit.

Stop because additional interviews are unlikely to change the current decision—not merely because a predetermined number has been reached.

Warning signs that you do not yet have enough evidence

• Most participants are friends, existing supporters or members of your immediate network.
• Participants are interested in the subject but are not realistic buyers or users.
• Buyers and users have been combined despite having different interests.
• An important customer segment is missing.
• The solution was presented before the current problem was understood.
• The questions encouraged praise or agreement.
• Answers concern hypothetical future behaviour rather than actual experience.
• New interviews continue to reveal major decision-relevant themes.
• Contradictory evidence has been dismissed as an outlier.
• Conclusions are based mainly on counting positive comments.

More interviews will not repair every problem. If participant selection or interview design is weak, correct the method before increasing the number.

How to judge whether the evidence is sufficient

Frequently asked questions

Questions about customer interview sample size