William Jones·May 10, 2026·6 min read

When to Stop Interviewing: Saturation in Synthetic Persona Research

research methodologyinterviewsproduct research

Every qualitative researcher knows the ritual. You conduct interview after interview, and at some point the transcripts start sounding familiar. New participants confirm what you already heard. You declare "theoretical saturation" and stop recruiting.

The problem is that most teams declare saturation after talking to 8-12 people who are remarkably similar to each other. They didn't reach saturation. They reached an echo chamber.

Synthetic personas don't automatically fix this. But they give you a tool that human recruitment never could: systematic personality variation across interviews.

The saturation problem in traditional research

Theoretical saturation — the point where new data stops generating new insights — was formalized by Glaser and Strauss in the 1960s. It's a sound concept. The execution is where things break down.

Guest, Bunce, and Johnson (2006) found that saturation typically occurs within 12 interviews for homogeneous samples. The key word is "homogeneous." If your participants share similar demographics, psychographics, and personality profiles, 12 interviews will feel like plenty. You'll hear the same concerns repeated. You'll stop recruiting confident that you've captured the landscape.

But you haven't. You've captured one slice of it.

The reason is straightforward: traditional recruitment optimizes for availability and willingness to participate. This produces a self-selected sample biased toward people who are relatively agreeable, open, and extraverted — the personality profile most likely to say yes to a research interview. You're missing the introverted skeptic who wouldn't respond to your recruiting email. You're missing the neurotic perfectionist who thought your screening questionnaire was too invasive. You're missing the disagreeable contrarian who doesn't see the point of giving free feedback.

Why personality variation matters more than sample size

Adding more interviews from the same personality cluster doesn't move you toward saturation. It moves you deeper into redundancy.

Imagine testing a new onboarding flow with 15 personas who all score high on Agreeableness and Openness. By interview 6, you'll hear consistent feedback: the flow is fine, minor suggestions, generally positive. You'll declare saturation at interview 10.

Now run the same test with a persona scoring low on Agreeableness and high on Conscientiousness. Suddenly the onboarding flow has problems you never heard about. Unclear labels that the agreeable personas glossed over. Missing confirmation steps that the conscientious persona finds unacceptable. A pricing page that the disagreeable persona tears apart.

That's not a new finding from a 16th interview. That's a fundamentally different perspective you missed because your first 15 interviews sampled from the same personality neighborhood.

The Big Five personality model gives you five independent dimensions to vary. Each dimension runs from low to high. Even sampling just two points per dimension (high and low) gives you 32 distinct personality combinations. That's more behavioral diversity than most research teams achieve in a year of recruiting.

A practical framework: OCEAN spectrum coverage

Here's a concrete approach to reaching genuine saturation with synthetic personas.

Step 1: Start with your target archetype. Build a persona that matches your best understanding of the target user. Set realistic OCEAN scores. Interview it 3-5 times across different scenarios.

Step 2: Systematically vary one dimension at a time. Take your baseline persona and create variants where a single OCEAN score is pushed to the opposite extreme. High Openness becomes Low Openness. Low Neuroticism becomes High Neuroticism. Interview each variant 2-3 times.

Step 3: Combine the extremes. Create personas that combine unusual trait combinations — high Neuroticism with low Agreeableness (the anxious critic), or low Openness with high Conscientiousness (the by-the-book traditionalist). These edge-case combinations often surface the most unexpected feedback.

Step 4: Declare saturation when personality variation stops producing new themes. Not when you've heard the same thing from similar personas, but when fundamentally different personality profiles converge on the same issues.

This is the difference between false saturation and real saturation. False saturation is hearing agreement from similar people. Real saturation is hearing agreement from people who have every reason to disagree with each other.

How Synthicant's template library helps

Synthicant ships with 22 persona templates spanning 6 research frameworks. These templates aren't random — they're designed to cover the OCEAN spectrum systematically.

The technology adopter framework includes personas from eager early adopters (high Openness, high Extraversion) to skeptical laggards (low Openness, high Conscientiousness). The decision-making framework spans from analytical deliberators (high Conscientiousness, low Extraversion) to impulsive risk-takers (low Conscientiousness, high Openness).

You can use these templates as-is for broad coverage, or use them as starting points, adjusting the OCEAN sliders to create targeted variants around your specific user base.

The point isn't to interview all 22 templates for every research question. The point is to ensure you're sampling from across the personality spectrum before you declare that you've heard enough.

The convergence test

Here's a heuristic that works well in practice: run your key research question past three personas with maximally different personality profiles. If all three surface the same concern, it's almost certainly a real issue. If only the high-Neuroticism persona flags it, it might be an edge case worth noting but not prioritizing. If only the high-Agreeableness persona says it's fine, that's a yellow flag — they might be telling you what you want to hear.

Traditional research can't do this because you can't control your participants' personality traits. You take who you get and hope for diversity. Synthetic personas let you engineer diversity into the sample, which means your saturation point actually means something.

Twelve interviews with the same personality type is an echo chamber. Twelve interviews across the OCEAN spectrum is research.

References

Guest, G., Bunce, A., & Johnson, L. (2006). "How Many Interviews Are Enough? An Experiment with Data Saturation and Variability." Field Methods, 18(1), 59-82. — Foundational study establishing that thematic saturation typically occurs within 12 interviews, with the caveat that this applies to relatively homogeneous samples.

Glaser, B.G. & Strauss, A.L. (1967). "The Discovery of Grounded Theory: Strategies for Qualitative Research." Aldine Publishing. — Introduced theoretical saturation as a formal concept in qualitative methodology. The stopping rule that most research teams reference, whether they know it or not.

Jiang, H., Zhang, X., Cao, X., et al. (2024). "PersonaLLM: Investigating the Ability of Large Language Models to Express Personality Traits." Proceedings of NAACL 2024. — Demonstrated that assigned Big Five personas produce consistent, distinguishable behavior in LLMs, validating the approach of using personality variation as a sampling strategy.

John, O.P. & Srivastava, S. (1999). "The Big Five Trait Taxonomy: History, Measurement, and Theoretical Perspectives." Handbook of Personality: Theory and Research. — The most-cited overview of Big Five personality taxonomy, establishing the independence and comprehensiveness of the five dimensions used as variation axes in this framework.