Answers
Synthetic panel vs AI focus group
A synthetic panel is a census-grounded set of simulated respondents asked the same instrument and aggregated into option shares. An AI focus group is a small tracked conversation. Lewsearch is a panel that can seat a room. The calibrated 50-state benchmark is a 7.07-point average miss on the best 80% of 1,521 scored questions (1,216 retained). 10.02 points across every scored question. Every study is calibrated against published survey data before it reaches you. The panel is the statistic. The room is the reread.
The artifact you walk out with
Search results for both phrases mix interview tools and tabulated panels. The useful test is the artifact you walk out with. If it is a distribution with an n and a published miss, you bought a panel. If it is a transcript and a theme map, you bought a group. The calibrated 50-state benchmark is a 7.07-point average miss on the best 80% of 1,521 scored questions (1,216 retained). 10.02 points across every scored question. The room inherits that panel. Product notes are on /focus-groups and /panel-studies.
Worked example: one claim, two reads
A grocery claim tests in a state market. The panel returns option shares. The calibrated 50-state benchmark is a 7.07-point average miss on the best 80% of 1,521 scored questions (1,216 retained). 10.02 points across every scored question. The room of 12 highlighted simulated respondents tells you which phrase sounded like a lawyer wrote it. The room alone does not give you a share. The panel alone can miss the wording that will fail in a live facility. Lewsearch seats the room on the panel, and the panel stays the sample.
| Question | Synthetic panel | AI focus group |
|---|---|---|
| What you can quote | Option shares and crosstabs | Quotes, themes, contradictions |
| n that is the sample | The full panel (often 500+) | The room (often 5–30) |
| Lewsearch role | The production artifact | Optional tracked room on top of the panel |
| Published miss | 7.07-point average miss on the best 80% of 1,521 scored questions (1,216 retained). 10.02 points across every scored question | Same panel underneath; quotes are not scored as an average miss |
FAQ
- What is the difference between a synthetic panel and an AI focus group?
- A panel returns percentages you can tabulate. A focus group returns a small conversation you can follow. Lewsearch runs the panel first and optionally seats 5–30 named simulated respondents on top of it.
- Which one is Lewsearch?
- Panel-shaped. Studies typically sample 500 to 10,000 simulated respondents from a chosen market. The optional room is 5–30 simulated respondents with memory across questions. See /focus-groups and /study.
- When do I use each?
- Use the panel when you need shares, ranks, or crosstabs. Use a room when you need to hear the same 12 simulated respondents contradict themselves after the second screen. Use a human facility when the sample itself is the deliverable.
- Can I convert 12 quotes into a poll?
- No. Twelve quotes are not 500 option shares, synthetic or human.
Check the work
Related answers
Canonical: https://lewsearch.com/answers/synthetic-panel-vs-ai-focus-group