Choose a Panel Size
Set respondent counts by research method, audience coverage and budget.
Respondent count controls the size of a simulated study. It is different from the population size shown on an audience's detail page.
Recommended starting points
The composer uses these recommendations for the corresponding research types. They are starting points, not guarantees of accuracy or required minimums.
| Method | Recommended respondents |
|---|---|
| Closed question or survey | 400 |
| Open question or interview | 50 |
| Preference test | 200 |
| Creative test | 150 |
| URL feedback | 30 |
| Focus group | 8 at one table |
| PR test | Set by the test |
An audience can seat at most the number of people shown in the audience picker. A segment defined by a criterion people are asked about may seat fewer than the recommended number; the run then seats everyone the segment has, and you are charged for the people seated.
For a first experiment, a smaller panel can help check the question and stimulus before a larger run. Set counts per audience when comparing groups. Focus groups seat 3–8 participants; increasing an audience count does not enlarge one table beyond that. A PR test needs enough respondents per version to compare the direction of change; the displayed estimate is the source of truth for the current setup.
Scale for the question
Use more respondents when you need coverage of several meaningful subgroups. Inspect the denominator after filtering: a large overall run may still contain a very small selected group.
Adding simulated respondents does not fix a poorly defined audience, missing source evidence or leading questions. It also does not turn a synthetic panel into a random sample of real people with a conventional survey margin of error.
Estimate the cost
Review the estimate shown for the selected method, audience and respondent count. Reports are separate.
See Pricing & credits, then run and review. For exact population claims, plan Validate with real world data.