What Is a Focus Group? A 2026 Guide
Qualitati Research Team · 2026-07-25 · 10 min read
Last updated: July 25, 2026
Short answer
A focus group is a qualitative research method in which a trained moderator guides a small group — typically 6 to 10 pre-screened participants — through a semi-structured discussion to surface attitudes, language, and reactions to a product, idea, or topic. Its distinctive value is group interaction: participants build on, challenge, and react to each other. Its central risk is groupthink, where the group converges on a dominant view. In 2026, AI-moderated and synthetic formats reshape how focus groups are run.
Key takeaways
- A focus group is defined by group interaction, not just group logistics — the data comes from participants reacting to one another, which a one-on-one interview cannot produce.
- The standard size is 6–10 participants for 90–120 minutes, with a moderator working from a semi-structured discussion guide.
- The method's greatest weakness is groupthink and dominant-voice bias; a large part of a moderator's skill is managing these dynamics.
- Use focus groups for exploration, language discovery, and reaction testing — not for measuring prevalence or capturing sensitive individual experience.
- AI-moderated focus groups run many private, parallel conversations that remove groupthink at the cost of live cross-talk; synthetic focus groups simulate participants for early exploration only.
- Use the Focus Group Format Decision Matrix below to pick the right format for your question.
What a focus group actually is
A focus group is a moderated group discussion used to collect qualitative data. Unlike a survey, it does not count; unlike a one-on-one interview, it does not isolate. Its defining feature is that participants hear and respond to each other, so the researcher observes not only what people think but how opinions form, shift, and collide in a social setting. As ATLAS.ti's methodology guide puts it, focus groups "capitalize on group dynamics" to elicit multiple perspectives in a single session.
The method was popularized in mid-20th-century sociology and became a staple of market research, UX research, and public-health studies. A typical session runs 90 to 120 minutes with a discussion guide of open-ended questions, per Simply Psychology's overview of the method.
Focus group vs. interview vs. survey
| Method | Unit of data | Best for | Weak for |
|---|
| Focus group | Group interaction | Reactions, language, idea generation, consensus/dissent | Prevalence, sensitive topics, individual depth |
| 1:1 interview | Individual account | Depth, personal experience, sensitive topics | Social dynamics, group reaction |
| Survey | Aggregated responses | Measuring how many, how much | The "why," unanticipated themes |
How many people, and why the size matters
The conventional focus group is 6 to 10 participants (some practitioners stretch to 12). This range is a deliberate compromise: large enough to generate interaction and a spread of views, small enough that every participant gets meaningful airtime in a 90-minute window. Below six, the discussion can feel thin and stall; above ten, quieter voices get crowded out and the moderator loses control of the floor.
Most studies run multiple groups (commonly three to five per audience segment) rather than one, because a single group is vulnerable to its particular personalities. Researchers watch for the point at which new groups stop producing new themes — a form of data saturation.
Group dynamics: the method's superpower and its trap
Everything valuable about a focus group and everything dangerous about it come from the same source: people influence each other. On the upside, one participant's comment jogs another's memory, disagreement exposes the real fault lines in an audience, and the group co-constructs the vocabulary a market actually uses. On the downside, three failure modes recur:
- Groupthink. Participants unconsciously conform to the emerging group opinion to preserve social harmony, as ATLAS.ti notes. Early consensus can be an artifact of politeness, not genuine agreement.
- Dominant voices. One or two confident participants monopolize the floor and anchor the room; the moderator must actively redistribute airtime.
- Social desirability. In front of peers, people report what sounds respectable rather than what they actually do or believe.
This is why moderation is a genuine skill. The UX research field guide from User Interviews stresses that the moderator's core job is managing these dynamics — drawing out the quiet, respectfully containing the dominant, and staying neutral so the guide, not the loudest person, sets the agenda.
How to run a focus group: the basic workflow
- Define the objective. One or two decisions the session must inform. Focus groups drift without a sharp objective.
- Recruit and screen. 6–10 participants per group matched to your target segment; multiple groups per segment.
- Write a semi-structured guide. Open-ended questions sequenced broad-to-specific, with room to follow unexpected threads.
- Moderate. Warm up, pose the grand-tour question, work through the core topics, manage airtime, and probe thin answers.
- Analyze. Transcribe and code the discussion into themes — ideally noting where consensus was genuine versus manufactured by the room.
Where AI changes the format in 2026
Two AI-driven variants now sit alongside the traditional room, and they solve different problems.
AI-moderated focus groups
An AI-moderated focus group replaces the single scheduled panel with many private, one-on-one conversations run in parallel by a conversational AI moderator, then aggregates the results across a much larger sample. As Perspective AI describes it (2026), this design deliberately removes groupthink and scheduling friction — every participant gives an independent opinion, and a good AI moderator applies at least a couple of follow-up probes when an answer is thin. The trade-off is real: you gain scale and independence but lose the live cross-talk that makes disagreement visible in the same moment. Some vendors reintroduce interaction by sharing anonymized peer responses back to participants.
Synthetic focus groups
A synthetic focus group uses LLM-simulated participants instead of humans. It is useful for early, exploratory stimulus — pressure-testing a discussion guide, generating hypotheses, or previewing likely objections before recruiting real people. It is not evidence about a real market and should never be reported as such. (See our guide on whether LLM synthetic participants work.)
Focus Group Format Decision Matrix
An original Qualitati framework for choosing a format from your research goal.
| If your goal is… | Best format | Why |
|---|
| Observe how opinions form and clash in real time | Traditional (in-person / live video) | Only synchronous interaction shows live dynamics |
| Independent reactions at large scale, fast | AI-moderated, parallel 1:1 | Removes groupthink; hundreds of sessions at once |
| Sensitive or personal experience | 1:1 interviews (not a group) | Peers suppress candor; privacy restores it |
| Early hypothesis / guide pressure-test | Synthetic focus group | Cheap, fast, exploratory — not evidence |
| Multilingual audiences across markets | AI-moderated (multilingual) | Consistent moderation across languages |
Focus Group Readiness Checklist
- □ The objective names a specific decision the session will inform.
- □ Participants are screened to the target segment, 6–10 per group.
- □ Multiple groups (or a large parallel sample) are planned, not one.
- □ The guide is open-ended and sequenced broad-to-specific.
- □ A plan exists to manage dominant voices and draw out quiet ones.
- □ The topic is not so sensitive that peers would suppress candor.
- □ Analysis will distinguish genuine consensus from social conformity.
Where Qualitati fits
Qualitati is an AI user research platform for product, UX, and customer-insights teams. Its AI-moderated focus groups probe for depth, bring in quiet voices, check consensus, and are designed to counter groupthink by running independent conversations at scale. For exploratory work, synthetic focus groups let you rehearse a guide before recruiting. When sessions are done, ThemeLens and the QDA Workspace turn transcripts into coded, quote-anchored themes, and multilingual moderation covers research across ten languages. The point is not to abolish the focus group but to fix its two oldest weaknesses — groupthink and limited scale — while keeping human judgment on design and interpretation.
Limitations and trade-offs
- Focus groups measure reactions, not prevalence. They cannot tell you how many people hold a view. Pair them with a survey when you need numbers.
- Group settings suppress candor on sensitive topics. Health, money, and personal conflict belong in one-on-one interviews.
- AI moderation removes live interaction. Parallel 1:1 conversations trade real-time cross-talk for independence and scale — a real loss if watching dynamics is the goal.
- Synthetic participants are not a market. Treat them as a hypothesis engine, never as findings. Human review is recommended for any study informing high-stakes or regulated decisions.
Frequently asked questions
What is a focus group in simple terms?
A focus group is a guided group discussion, usually 6 to 10 people led by a moderator, used to explore opinions, reactions, and language about a topic or product. Its value comes from participants reacting to one another.
How many people should be in a focus group?
Typically 6 to 10 participants (occasionally up to 12). This size is large enough to spark interaction and small enough that everyone gets meaningful airtime in a 90-minute session. Most studies run several groups, not one.
What is the main disadvantage of focus groups?
Groupthink — participants conforming to a dominant or emerging group opinion rather than voicing independent views. Dominant voices and social desirability compound the problem, which is why skilled moderation matters.
Focus group or one-on-one interview — which should I use?
Use a focus group when group interaction and reaction are the point. Use one-on-one interviews for depth, personal experience, or sensitive topics where peers would suppress honesty.
Are AI-moderated focus groups real focus groups?
They are a variant. AI-moderated formats typically run many private one-on-one conversations in parallel, removing groupthink and adding scale, but sacrificing live cross-talk. Whether that is a gain or a loss depends on whether observing real-time dynamics is your goal.
Can synthetic (AI) participants replace a real focus group?
No. Synthetic focus groups are useful for early exploration and pressure-testing a guide, but their output is simulation, not evidence about a real audience, and should never be reported as findings.
Bottom line
A focus group is a moderated group discussion whose power — and whose peril — both come from participants influencing each other. Use it to explore reactions, language, and where an audience agrees or splits; avoid it for prevalence and sensitive individual experience. In 2026, AI-moderated formats let you keep the exploratory value while defusing groupthink and scaling far beyond one room, and synthetic groups give you a cheap rehearsal — as long as you never mistake the rehearsal for the show.
Start free with 30 credits — no credit card required. Create an account and run an AI-moderated focus group that probes for depth and counters groupthink, or view transparent pricing first.