How Many Synthetic Individuals Do I Need?
Sizing guidance — when small is enough and when scale starts to matter.
More individuals isn't always better. The right number depends on what you're trying to learn.
Qualitative research — the kind iMario is built for — follows a principle of diminishing returns. After a certain point, new individuals stop surfacing new patterns and start confirming what you already know. The goal isn't volume. It's knowing when you've heard enough.
Sizing by research goal
- Small set (1–5) — Best for deep exploration or highly specific scenarios. Use this when you need to understand one perspective thoroughly, rehearse a specific conversation, or simulate a single well-defined individual.
- Medium set (10–50) — Good for pattern discovery and cross-segment comparison. Most research tasks live here. Enough individuals to surface meaningful variation, small enough to read and interpret carefully.
- Large set (50+) — Use when you need broad coverage and demographic diversity. Best for validating patterns you've already identified, or when your question requires statistical confidence across a wide population.
Start small, scale on signal
Start small. Run your task, read the results, and let the findings tell you whether you need more. Scaling too early produces noise. Scaling after you know what you're looking for produces signal.