Focus group research: when a group works better than one-on-one interviews

Focus group research brings a small group of similar people together for a moderated discussion of one topic. It works best when you want a range of views, the words people use with each other, and a sense of how opinions shift when people hear one another. One-on-one interviews work best when you need a detailed account of what one person actually did, how a decision was made, or anything people wouldn't say in front of peers.

University of Wisconsin-Extension describes focus groups as "focused" in two ways: the participants are similar in some way, and the discussion follows a focused set of questions. Because participants hear and respond to each other, a group produces different information than the same people would give alone (Quick Tips: Focus Group Interviews). The goal is perceptions and a range of opinions, not consensus.

What a focus group is good for

• A range of reactions in one session. Hearing seven people react to the same concept, message or name in 90 minutes is efficient.
• Shared language. You hear the words people use with each other, which helps with positioning and with wording a later survey.
• Needs you didn't know to ask about. In The Use and Misuse of Focus Groups, Jakob Nielsen says the proper role of focus groups in product development is to discover what users want. He describes running one with system administrators at Sun to learn how they wanted large documentation files distributed across servers, questions a usability test would never have raised.
• Group dynamics and organizational issues. Nielsen notes that groups let you observe these directly.
• Work that spans days or weeks. People can describe long processes that would be expensive to observe.

Where a focus group misleads

• Say versus do. Nielsen's central warning is that focus groups capture what people say they do, not how they actually behave. He recommends always supplementing them with direct observation of one user at a time.
• Usability. Watching a demo is different from using a product. He calls focus groups "a rather poor method for evaluating interface usability."
• Wanting versus needing. People may think they want one thing when they need another. Nielsen suggests showing concrete examples or prototypes to reduce this.
• The group changes the answer. Richard Krueger's analysis guidance notes that participants sometimes change or even reverse their positions after hearing others (Designing and Conducting Focus Group Interviews). That is worth observing, but it means a group's conclusion is not each person's view.
• One session is not enough. Nielsen and Krueger both recommend running more than one group, because any single discussion can be sidetracked or unrepresentative.

Choose by the question

The senior-buyers row is our judgment, not a finding from the sources. When participants work for competitors, or the topic touches budgets and internal decisions, expect less candor in a group. Interviews also avoid the scheduling problem of getting several senior people into one session at the same time.

A worked example

Suppose you sell ingredients to independent restaurants and are considering a new ordering service. This example is hypothetical.

You have two questions. First, how does each kitchen decide what to order from which supplier? That is a sequence of decisions made by one or two people in each restaurant, with details about prices and relationships. It calls for one-on-one interviews, perhaps five or six. Qualitative market research walks through a similar case.

Second, once you have a short description of the service, how do chefs and owners react to it, and what questions does it raise? That is a good group question. Two groups of six or seven, one of chefs and one of owners, would show the range of reactions and the words each group uses. Krueger recommends putting similar types of people together, which is why chefs and owners are in separate groups.

Running the interviews first gives you better material for the groups: you know the real problems and can describe the service in terms kitchens already use. Running a group first can also work when you don't yet know which topics matter, as long as you follow up with individual interviews on the ones that do.

Size and format

Sources differ on the ideal size. Krueger suggests five to ten participants, with six to eight preferred. Nielsen describes six to nine users for about two hours. UW-Extension says eight to 12 participants for one to two hours. For a business audience, the smaller end is easier to recruit and gives everyone time to speak.

A group needs a skilled moderator and, ideally, an assistant who takes notes. Focus group questions covers the question route and moderation in detail.

There is also a middle option. GOV.UK's interview guidance notes that you can interview people in pairs or small groups when they use a service together, such as members of a team who work on the same tasks (Using in-depth interviews). For a B2B product used by a small team, a joint interview with the people who share the workflow can show handoffs that neither a big group nor a single interview would.

Your next step

Write your main research question on one line. If it asks what someone did or decided, plan one-on-one interviews and use customer interviews as your guide. If it asks how a group of similar people reacts to or talks about something, plan at least two focus groups. If it asks how many people think something, interviews vs. surveys covers when to count.

Either way, you need people who do the job. Instant Expert can find people who match a description you write, such as "chefs who order for independent restaurants." You review who it finds, it sends your invitations, and you pay only for calls that get booked. The procurement professionals in restaurants directory page is one place to start.