What is a focus group?
A moderated discussion with a group of users that allows you to learn about user attitudes, ideas, and desires. The discussion can be centered around a new process or an existing process, topic, idea, etc.
How to conduct a focus group:
3-5 participants per session is recommended. Hold sessions until you’ve reached a good understanding of what users want. Focus groups should be conducted early in a project. A benefit to focus groups is that they don’t require anything visual (mockup, prototype, etc.) for participants to respond to.
Prepare an agenda including a list of the top-level issues to be tackled and an introduction script explaining the purpose of the day and how the day will go. Be sure to always use a quiet room with few distractions and arrange people in a circle. Ask participants to introduce themselves. All questions you ask should be open and neutral and encourage free-flowing discussion around the relevant issue(s).
A summary document should be produced for each session and include relevant profile information about the people who attended the session but keep their names anonymous.
Focus groups require a number of participants to be in the same space at the same time, so organizing them can take time and resources (compensation, obtaining a space, etc.). The research team also needs to create a number of topics for the participants to discuss, as well as determine goals for the session.
Focus groups can provide qualitative information, but the responses tend to be hypothetical. Asking someone what they think they would do vs actually watching them do something can return different results. Focus groups also have a tendency to introduce bias, as individual participants can sway the others within their group.
Coordinating a focus group has its complexities, but the real challenge is moderating the session. Learning how to deal with stronger or weaker personalities within a group is an art that benefits from practice. Synthesizing the results of a focus group requires finesse, as the conversations tend to include irrelevant “noise” more frequently than other research methods.