What changes when a scenario costs minutes.
We don't replace fieldwork. We change what reaches it — and how much the next hypothesis costs to test.
| Focus group | Survey / panel | SyntheticPerson | |
|---|---|---|---|
| Time per study | 3–6 weeks | 1–3 weeks | Minutes per scenario |
| Marginal cost of 2nd scenario | Full cost again | Nearly full again | Approaches zero |
| Stated vs. real gap | High (group effect) | High (desirability bias) | Modeled explicitly |
| Longitudinal memory | None | Partial, expensive | Native |
| Brazil calibration | Depends on recruitment | Depends on sample | At the model's core |
| Uncertainty reported | Rarely | Margin of error | Bands on every output |
Simulation doesn't retire the focus group or the survey — it comes before them, reducing the number of expensive hypotheses that reach the field. The field gets sharper focus.
Backtest updates
Get every new backtest by email.
When a population earns or loses a seal, or a new real outcome is measured against the engine, we write once. No newsletter cadence — only new entries in the ledger.