3.7 Focus Groups
At a Glance
Other names Focus Group · Group Interview
In Brief
A focus group is a facilitated discussion with 6–10 participants on a specific topic. A moderator guides the conversation while participants react to each other’s ideas. The output is a set of themes, customer language, and insight into how people influence each other’s opinions.
Common Use Case
You want to explore how a group of potential customers talk about and influence each other’s thinking on the problem you are investigating. You bring several participants together for a guided discussion and observe how one person’s perspective shifts what others say. The group dynamics reveal social influences on decision-making that one-on-one interviews would miss.
Helps Answer
- How do participants influence each other when making decisions together?
- What problems do participants describe when they talk openly?
- What do people think about the current alternatives available to them?
- What language do participants use to describe their needs?
- How do group members react to each other’s ideas?
Description
A focus group is a facilitated small-group discussion for hearing the “voice of the customer.” The conversation chains associations among participants, surfacing shared language and unarticulated needs that one-on-one interviews often miss. It fits open-ended exploration — unpacking a complicated buying journey, teasing apart the factors that drive a decision, reacting to a concept or stimulus, or hearing how people talk about a brand.
The method draws heavy and deserved skepticism, because groups are widely misused for evaluative work — confirming a design, measuring willingness to pay — where one-on-one interviews and behavioral tests give far more reliable signal. The social setting amplifies the loudest voices and invites groupthink, and purchasing decisions are often partly unconscious, which limits what any verbal method can surface. Treat a focus group as a way to hear how people talk about a problem, not as proof they will buy a solution.
You do not need a recruited panel, a venue, or a paid moderator to run a lightweight focus group: a 45–60 minute video call with 4–6 participants from your own network, with an AI tool handling transcription, captures most of the value cheaply. A common setup is a hybrid, where a human facilitator runs the live session while AI handles transcription, theme extraction, and cross-session comparison. Some platforms run AI-Moderated Interviews at Scale, where a chatbot guides participants and follows up on interesting threads.
How to
Prep
- Pick a single, clear purpose. One product or one issue per session. If you cannot state the learning goal in a sentence, you are not ready to run the group.
- Narrow your target audience. Talk to one segment at a time. Mixing distinct participant types in one room produces conversation that averages over real differences.
- Decide whether you need a comparison group. Running a parallel group from a different segment lets you contrast opinions against a wider context.
- Refrain from ulterior motives. Do not run a focus group to validate a decision you have already made or to soft-launch a sales pitch. The signal will be skewed before the first question lands.
- Find a co-facilitator. A second person handles note-taking, equipment, and side conversations so the moderator can stay present in the discussion.
- Choose a comfortable venue and recording method. Comfort and psychological safety shape how openly people speak. For virtual sessions, default to a meeting tool participants already know.
- Write the discussion guide. Prepare up to ten open-ended questions. Establish rapport early, avoid jargon, and save anything embarrassing or threatening for late in the session, once participants feel comfortable. Decide the opening framing here too: that this is a brainstorming-style discussion to hear opinions rather than reach consensus, and that there are no right or wrong answers.
Execution
- Get consent. Pass out consent forms at the start and confirm everyone understands the session is being recorded.
- Keep introductions short. Have everyone introduce themselves briefly — long introductions set the expectation that everyone will speak at length.
- Open with your framing, then guide loosely. Deliver the framing you prepared in Prep, then ask your prepared questions while staying flexible, adding follow-ups when something interesting surfaces. Ideally participants begin talking among themselves and you can step back.
- Stay neutral and draw people in. Keep your reactions even and empathetic, make eye contact with the participants who are speaking less, and gently invite them in.
- Capture more than words. Your co-facilitator takes written notes, capturing non-verbal cues as well as what was said. AI transcription handles the verbatim record automatically; some platforms also flag in real time when engagement spikes or when a participant is being drowned out.
- Prevent any one person from dominating. Use prompts like “Does anyone else have a different perspective?” or “I want to make sure we hear from everyone — [name], what do you think?”
- Wrap within 90 minutes. Keep the session between 45 and 90 minutes; energy and signal both decline sharply after that.
- Offer a private feedback channel. Provide a short feedback form so participants can flag anything they did not feel comfortable saying out loud.
- Repeat if you can. Run a few groups with comparable composition — one session is a single data point and is unusually vulnerable to one strong voice.
Analysis
- Transcribe the session. AI transcription tools handle this in minutes; clean obvious errors before coding.
- Code and cluster the responses. Group related quotes into themes using affinity mapping — sticky notes or a digital equivalent that lets you cluster similar reasons together. Capture the participant’s actual language, not your paraphrase.
- Look for divergent opinions and outliers. A focus group’s strength is showing how people influence each other; pay particular attention to moments where someone changes their mind, pushes back, or goes quiet.
- Separate group-influenced views from individual beliefs. When a theme looks important, plan a follow-up one-on-one interview to check whether participants hold the view independently or only adopted it under social pressure.
- Triangulate with other data. Compare themes against your existing customer interviews, support tickets, and behavioral data before treating any single theme as a finding.
- Confirmation bias Do not run a focus group to confirm what you already believe. Listen for divergent or unexpected opinions, and write down the moments that surprise you.
- Small-sample generalizability bias One focus group can produce a vivid, articulate theme that does not generalize to the wider market. Treat themes as hypotheses, not conclusions, until triangulated across multiple sessions or methods.
- Say-do gap Assume participants cannot accurately describe why they make decisions. Use focus groups to surface language and concerns; use behavioral methods to measure intent.
- Social-desirability bias Participants understate behaviors they think the group will judge (price-sensitivity, brand-switching, lapsed habits). Anonymize answers where you can, and triangulate sensitive claims with private follow-ups.
- Observer effect Participants behave differently when they know they are being recorded or watched through a one-way mirror. Acknowledge the recording up front, then move on quickly so it fades from attention.
- Moderator-induced bias An inexperienced moderator may steer participants toward leading questions or cede the room to one dominant voice, distorting which views surface and how they are weighted. If facilitation is not your strength, partner with someone who has done it before or run a pilot session first.
Learn more
Case Studies
Procter & Gamble: FemCare shifts to “heart-to-heart” conversations
P&G’s FemCare division shifted from behind-the-mirror focus groups toward face-to-face “heart-to-heart” conversations with consumers, citing richer insight than the traditional one-way-mirror format.
California Milk Processor Board: the “Got Milk?” deprivation insight
In 1993, focus groups asked participants to go a week without milk; their genuine anxiety in the “what if there’s no milk?” scenario became the deprivation strategy behind the iconic campaign — a reaction the numbers had missed.
Coca-Cola: New Coke (1985)
Before reformulating its flagship, Coca-Cola ran taste tests with roughly 200,000 consumers who preferred the new recipe; its focus groups, however, surfaced strong opposition from loyal drinkers, which researchers dismissed as a minority opinion. The quantitative data won, the backlash forced a reversal within months — a classic case of group qualitative signal being overruled by the numbers.
Further reading
- Krueger, Richard A. & Casey, Mary Anne — Focus Groups: A Practical Guide for Applied Research (5th ed., Sage, 2015)
- Merton, Robert K., Fiske, Marjorie & Kendall, Patricia L. — The Focused Interview: A Manual of Problems and Procedures (Free Press, 1956)
- Morgan, David L. — Focus Groups as Qualitative Research (2nd ed., Sage, 1997)
- Morgan, Carol-Ann — Using the Focus Group in Market Research (B2B International)
- Nielsen Norman Group — The Use and Misuse of Focus Groups
- Zaltman, Gerald — How Customers Think: Essential Insights into the Mind of the Market (Harvard Business School Press, 2003)
Got something to add? Share with the community.
