Quick answer
A focus group is a moderated qualitative discussion among selected participants about a defined topic, experience or stimulus. It is useful for exploring language, shared norms, disagreement, concept reactions and the way views form socially. Define the research question, create compatible groups around relevant differences, recruit purposively, obtain consent, use neutral open prompts and activities, prevent dominance without forcing consensus, and analyze both content and interaction across several groups. A focus group cannot estimate population prevalence, predict sales or determine a winner by show of hands. Validate important claims with individual research, surveys, behavioral evidence or experiments.
What a focus group is
A focus group is a guided discussion among a small group selected for relevant experience or characteristics. A moderator introduces topics and stimuli, encourages participants to respond to one another and explores both shared and divergent views.
Interaction is the distinctive source of data. Participants compare experiences, challenge claims, borrow language and reveal social norms. The group is not merely a cheaper way to run several interviews simultaneously, because what people say changes in response to the room.
That same interaction creates limitations. Dominant voices, conformity, status and impression management can suppress views. Design and analysis must examine how an opinion emerged, not simply count how many people voiced it.
Use focus groups for the right question
Focus groups are useful for exploring category language, social meaning, shared routines, reactions to an early concept and reasons for disagreement. They can help teams hear how people collectively negotiate a new idea or controversial trade-off.
Use individual interviews for highly personal histories or when power, confidentiality and sensitivity make group disclosure unsafe. Use observation for actual context, usability testing for task performance, surveys for prevalence and experiments for causal effect.
Do not ask a group to forecast market demand or vote a final creative winner. Expressed enthusiasm in an artificial room lacks price, competition, attention and real consequence. Treat it as a clue about meaning and mechanism.
Build a focus-group study
Define the decision and why interaction adds value. Create a group matrix around relevant contrasts, plan several groups to compare patterns and specify which people should not be placed together. Recruitment quality matters more than filling every seat.
Design a guide that moves from experience to stimulus and reflection. Use private writing or ranking before conversation when independent reaction matters. Rotate or randomize concept order where fatigue and sequence could shape response.
Moderate for balanced participation, debrief after each group and analyze transcripts, notes and interaction across groups. Stop when relevant group types are covered and new sessions add little decision value, not when one room produces a compelling quote.
Question
Define why social interaction is useful and what decision the group should inform.
- What can discussion reveal?
- What must be learned individually instead?
Composition
Create compatible groups while separating differences that could silence or distort discussion.
- Who can speak safely together?
- Which contrasts need separate groups?
Design
Build neutral questions, private tasks and stimulus sequence that prevent premature consensus.
- What should be captured before discussion?
- Does the stimulus answer the question?
Moderate
Create balanced participation, probe reasoning and make disagreement safe.
- Whose voice is missing?
- Is consensus real or social?
Analyze
Compare interaction and content across groups, then choose the next validation method.
- How did the view form?
- What cannot the groups establish?
Compose groups for safety and discussion
Participants need enough common ground to understand one another and enough difference to generate useful comparison. Separate groups when role, seniority, customer status, language or experience could inhibit candor. Employees and managers rarely discuss organizational failure freely together.
Consider whether participants know each other. Existing communities can reveal real norms but create confidentiality and relationship risk. Strangers may speak more freely yet lack shared context. Record the choice and its implications.
Recruit beyond enthusiastic users. Include rejecters, lapsed users or noncustomers when relevant, and support accessibility, interpretation and fair incentives. Over-recruit for predictable no-shows without turning the room into an unmanageable crowd.
- Interaction need justified
- Decision and limits written
- Group matrix covers relevant contrasts
- Power and safety reviewed
- Recruitment behavior-specific
- Consent explains group confidentiality limits
- Private reaction captured where needed
- Stimulus order controlled
- Moderator remains neutral
- Dominance and dissent noted
- Cross-group analysis completed
- Validation method selected
Design questions and stimulus
Begin with lived experience and category language before exposing brand concepts. Ask open questions that invite stories and comparison. Avoid asking whether participants agree with the sponsor's premise or explaining the concept until its unaided interpretation has been captured.
Stimuli should be complete enough to answer the research question but not polished beyond the decision stage. Clearly label placeholders and keep differences controlled when comparing variants. A rough idea and a finished advertisement are not fair competitors.
Use activities purposefully: private notes, card sorting, timeline construction or paired discussion can broaden participation. Activities are data-collection structures, not entertainment, and their outputs need the same contextual analysis as talk.
Focus group example
The rehydration study captures private comprehension before conversation, then uses the group to expose language, trust and social influence. If one parent confidently misinterprets a preparation step and others adopt the account, that interaction is a safety finding rather than majority evidence.
The company does not select the pack with the loudest support. It identifies unclear concepts and candidate wording, then moves to individual comprehension and usability tests where each participant must perform the task without group help.
A hypothetical health-products company explores packaging and instructions for a children's oral rehydration concept before formal comprehension and safety testing.
Run separate groups for parents of younger and older children and for different prior-use experience, with language-specific moderation and no clinical power imbalance.
Collect private first reactions and an individual explanation task before discussion so the group does not erase initial misunderstanding.
Explore how parents interpret purpose, preparation, warning and trust; rotate pack order and invite counterexamples rather than asking which design wins.
Map when opinions change, which words participants borrow, where one voice dominates and which safety interpretation persists across groups.
Revise concepts, then run individual comprehension and usability testing with clinical and regulatory review. Use a survey only for appropriately defined prevalence questions.
A group reaction cannot establish that packaging is safe or that the market will buy. The method generates language, mechanisms and design risks for later validation.
Moderate participation and disagreement
Set expectations that different experiences are useful and confidentiality cannot be guaranteed by the researcher once participants leave. Use names, eye contact and invitations to include quieter voices without forcing anyone to disclose more than they wish.
Probe the reason and example behind a reaction. Ask who sees it differently, what condition would change the view and whether the group is describing experience or expectation. Do not reward praise or defend the stimulus.
Manage dominance by redirecting, using round-robin or private tasks and taking a break if necessary. Notice alliances, jokes, silence and borrowed language. The moderator should not manufacture equal airtime when some participants lack relevant experience, but should prevent status from deciding truth.
Analyze interaction, not vote totals
Create group-level summaries immediately, then code transcripts and observations against the research question. Preserve who introduced an idea, how others responded, whether a position shifted and what social condition surrounded it.
Compare within and across group types. A theme appearing in every room may be robust within the sampled contexts, while a view voiced only after one participant's story may still reveal an important mechanism. Do not translate mentions into population percentages.
Search for contradiction and silence. Absence may reflect irrelevance, embarrassment or moderator design. Use source-linked quotations and describe the group context so stakeholders do not mistake one articulate participant for the market.
Report findings and limits honestly
State the participant criteria, group composition, dates, location or platform, stimuli, moderator, guide and analysis method. Separate finding, evidence, interpretation and implication. Explain where group dynamics may have shaped the result.
Use language such as some participants in these groups, not consumers prefer. A focus group can show reaction range and vocabulary, but the sample and interaction do not support statistical prevalence or a confidence interval.
Turn findings into questions for the next method. Test comprehension individually, estimate attitude distribution with a representative survey or test market behavior through an experiment. Preserve unpopular evidence rather than reporting only the sponsor's favored concept.
Limitations and common mistakes
Focus groups are vulnerable to conformity, dominance, social desirability, moderator influence, artificial context and selective interpretation. Participants may discuss imagined behavior more confidently than they would act under price and competition.
Common mistakes include mixing incompatible participants, showing leading stimulus, asking for forecasts, counting hands, reporting percentages, treating consensus as truth, ignoring body language and dissent, and choosing a concept because the client liked the room's energy.
The method is valuable when interaction itself is evidence. If the research question does not benefit from people responding to one another, choose a method that protects individual depth or observes actual behavior.
Use focus groups to learn the language, norms and range of reactions that form in discussion. Never treat the loudest voice, or even the room's consensus, as a market forecast.
Frequently asked questions
What is a focus group?
It is a moderated qualitative discussion among selected participants used to explore experience, language, norms, reactions and disagreement around a defined topic.
How many people should be in a focus group?
The right size depends on topic, depth and participant needs. It should be small enough for everyone to contribute and large enough to generate useful interaction; no universal number guarantees rigor.
How many focus groups are needed?
Run enough groups to cover relevant participant contrasts and compare patterns. Stop when additional groups add little decision value, while documenting which perspectives remain missing.
Can focus groups predict sales?
No. Group reactions lack representative sampling and real purchase conditions. Use them to generate mechanisms and language, then validate demand with suitable quantitative or behavioral methods.
What is the moderator's role?
The moderator creates safety, introduces neutral topics, balances participation, probes examples and disagreement, manages dominance and avoids defending or selling the concept.
Sources and further reading
- SAGE: Focus Groups, A Practical Guide for Applied Research ↗Publisher source on planning, questions, moderation, analysis and reporting
- SAGE: The Focus Group Guidebook ↗Publisher guidance on appropriate uses, expectations, ethics, timeline and cost
- CDC: Conducting Focus Groups in Small Communities ↗Peer-reviewed applied guidance on group familiarity, community context and participatory research
- GOV.UK Service Manual: User Research ↗Official method-selection, recruitment, consent, accessibility and analysis guidance