How Many Personas Per Interview?

Short answer: 5 is the standard starting point. The system supports up to 10 per interview. More isn't always better.

The rules of thumb

Goal	Recommended count
Quick concept validation, single user segment	3-5
Standard product/feature test	5-8
Multiple market segments, complex decision	8-10

The cap is 10 per project. If you need more than 10 personas, run multiple interviews rather than piling them into one — the marginal insight drops past 10, and the report becomes harder to read.

Why 5 is the usual answer

5 personas is enough to see patterns repeat without being flooded with transcripts
Most research questions hit "saturation" around the 6th or 7th persona (no new themes emerging)
5-person interviews typically finish in 50-70 minutes; 10-person ones take 80-100 minutes
A 5-persona report is short enough that someone will actually read it

When to push to 8 or 10

You genuinely cover distinct segments. Examples:

A fitness app testing across beginners, advanced users, weight-loss focused users, rehab users, and seniors — that's 5 segments × 2 personas = 10.
A B2B SaaS pricing test across small business, mid-market, enterprise, and individual buyers — 4 segments × 2 = 8.

The pattern: personas per segment × number of segments. Two per segment is the minimum to spot whether a finding is consistent or just one person's quirk.

When 3 is enough (and when it isn't)

3-5 works for early concept validation where you mainly want to confirm a direction. Don't go below 3 — anything smaller feels anecdotal rather than pattern-based, and one contrarian persona can dominate the report.

Composition matters as much as count

5 personas who all think the same way gives you 1 opinion × 5, not 5 opinions. The diversity question is at least as important as the count question. Aim for a mix:

2-3 personas matching your core target user
1-2 personas representing adjacent or potential users
(If budget allows) 1 edge-case persona to test boundaries

A good test: would someone reading your report learn something different from each persona? If not, you have too much overlap.

Watch for saturation

Don't set a fixed count and walk away. Watch the running interview (or check the partial output). If personas 6 and 7 aren't surfacing anything new, you can end the interview early. Most platforms charge by the persona, so this saves real money.

Common mistakes

Ten personas, one mindset. Picking 10 personas who all happen to be price-sensitive gives you one perspective multiplied by ten. Vary the dimensions.
Three personas for a complex multi-segment decision. You'll never know which segment disagreed with what.
Assuming more = more rigorous. For a single-segment concept test, 10 personas mostly produce noise around the same signal.

Starting your first interview
Setting up Discussion Chat