How to Conduct Focus Groups | Real-World .NET Applications

Before you launch into a focus group series, it's important to determine several things.

A schedule. The best results come from situations where there's been enough time to examine the contingencies. A good schedule (Table 9.1) provides sufficient time for everything, especially recruiting and guide writing, and enough flexibility to be able to make a mistake or two.

Table 9.1: A TYPICAL FOCUS GROUP SCHEDULE
Timing	Activity
t - 2 weeks	Determine audience and scope; start recruiting immediately
t - 2 weeks	Determine broad topics to be investigated; start writing guide
t - 1 week	Write first version of discussion guide; discuss exact topic wording with development team; check on recruiting
t - 3 days	Write second version of discussion guide with timing; discuss with development team; recruiting should be completed
t - 2 days	Complete guide; schedule run-through; set up and check all equipment
t - 1 day	Run-through in the morning; check times and adjust guide questions as appropriate; do final recruiting check
t	Conduct groups (usually 1–3 days, depending on scheduling); discuss with observers; collect copies of all notes
t + 1 day	Relax; do something else.
t + 3 days	Watch all tapes; take notes
t + 1 week	Combine notes; write analysis

The target audience. These are the people you're going to invite to the test. Specifically, you need to know the subset of the target audience that is likely to give you the best feedback.
The scope of your research. Focus group series can have a few groups of a handful of people or as many as a dozen groups with ten or more participants apiece. The number of groups and people will depend on the complexity of your questions, the depth to which you want to explore the answers, and the certainty with which you want to know these answers.
The topics that you want to research. Not all groups feel equally comfortable talking about all subjects, and not all subjects lend themselves to group discussion. Choosing topics carefully and thoughtfully writing a discussion guide helps you get the most information out of the group process without sacrificing the depth of research or the clarity of the results.

Setting a Schedule

A typical schedule for a focus group series takes about three weeks from beginning to end and should provide sufficient time for recruiting and writing the discussion guide.

Picking a Target Audience

It is not the actual differences between participants, but whether they perceive each other to be different, that determines their willingness to discuss a topic together.

—David L. Morgan, Focus Groups as Qualitative Research, p. 47

In focus groups, maybe more than in any of the other methods in this book, picking the right audience is crucial. For people to feel comfortable talking about their experiences and their values, they need to know that they are not going to be judged by the others in the group, that they are among peers. Thus, unlike most other kinds of research, focus groups concentrate on homogenous audiences.

Defining what "homogenous" means depends on the context of the research and the group. The participants don't need to be the same in all ways, just in the ways that are important to the research. For example, generational differences can keep a group from completely opening up about their musical tastes.

From your ideal target audience, you should choose a subset or several subsets that are likely to give you the most useful feedback. The right group will vary from situation to situation. First, you need a solid profile of your target audience, complete with a thorough understanding of their demographic, Web use, and technological makeup. For example, if you're just looking to find out what existing users value about your service, you want to pick the people who represent the largest subset of your actual audience. However, if you're looking to find out why there is a lot of turnover in your users, you must be much more specific: you must focus only on the people who are just like your standard population but have used your site once and never again.

Often, it's useful to choose several groups to get an understanding of the breadth of experience and as a way to concentrate on specific groups' experiences. For example, if your system doctors are supposed to use your system, but you see only medical students, then you probably want to talk to both groups. The doctors will tell you why they don't use it, and the medical students will tell you why they do. Maybe you need two separate sites.

Define the group similarly from the perspective of the members of the subgroups. If you feel that certain groups of people would not feel comfortable with each other, then don't put them together. Start with demographic and Web use terms. Income, race, sex, class, age, job, and computer experience all can play a role in how people interact in a group situation and how they react to a given user experience. In order to create a group that can comfortably discuss our topic, we have to take these differences into account. Once the basic outlines are defined, further divide the groups based on behavior: Who does what? What else do they do? Sometimes differences that matter in one group don't in another.

Here are some examples.

When picking subgroups for a study researching automobile buying experiences, it was pointed out that the men tend to talk a lot more than women in mixed groups when talking about cars. Since the researchers were interested in the opinions of both groups, the choice was made to have separate groups of men and women and a combined group.
When researching how to expand the customer base of a technology news service, it was decided that apart from matching basic demographics and Web use, it was important that the audience should spend a couple of hours reading news every day.
A medical services site decided that doctors and doctor-administrators would likely have trouble feeling comfortable with each other because of their roles in the hierarchy of a hospital, even though they would likely be using the same services. Thus, the groups were segregated based on how much time a doctor spent on administrative work.
A Web site aimed at elderly computer users decided that people living in the San Francisco Bay Area would likely have been exposed to a disproportionate amount of information about the Web when compared to people with otherwise identical profiles who lived elsewhere in the United States. Research was conducted in Washington, D.C., and St. Louis, Missouri.

You should feel free to define the subgroups in whatever way you feel comfortable. However, do not define subgroups based on their opinions or preferences. Prescreening participants for their stated values is likely to introduce unnecessary bias into the groups and defeats the core purpose of the focus groups, which is to determine those values. For example, don't screen for people "who like shopping online"; instead, focus on people who shop online and then determine the full range of reasons why these people shop.

Recruiting

Once you've picked a target audience (or audiences), it's time to find them and invite them. Recruiting for focus groups, much like recruiting for other kinds of research, should be done early and should begin as soon as you've picked your audience.

A couple of things distinguish recruiting a group of people from the recruiting process described in Chapter 6, which concentrates on finding individuals.

The exact profile of the participants is very important. With much research, it's possible to have participants who are close, but not exactly ideal target audiences. With focus groups, it's critical that all the people in a given group fall into the audience profile since one person outside it can derail a whole discussion. Owners of small shops, for example, could potentially feel uncomfortable discussing their problems and feelings with an owner of a large store or a franchise. And although good moderation can sometimes overcome such problems, it's best to avoid them in the first place.
Never recruit people who know each other. When two or more people know each other—sometimes even if they just know who the other person is—it creates an unbalanced dynamic in a group discussion. A raised eyebrow or a sigh says a lot between two people when they know each other, and can disrupt the discussion and inadvertently conceal important information. Thus, when recruiting using the friends and family method, avoid recruiting people who were all referred by the same person, and when using a database, select people who were initially recruited at different times or using different methods.
Avoid people who know about how focus groups work. The focus group process depends on candid answers and a level of comfort in the discussion group. Veterans of the process may attempt (often subconsciously) to give the kinds of answers that they feel are expected, or they may try to predict the next topic of discussion. Although not fatal, this can skew the discussion, necessitating extra moderation and analysis efforts. If you have to recruit people with focus group experiences, the experience should be neither frequent nor recent. For these same reasons, you should never include people who work for organizations where they could be exposed to a lot of focus groups, even indirectly. So no marketing companies, no advertising agencies, and so on.
Screen out people who have significantly more knowledge about any of the topics than the other participants. If there's an "expert" in a group, that person's knowledge can intimidate the other participants, and his or her views can bias the whole group's perspective.

Once you've decided on your target audiences, write a profile of the kind of people you want to recruit.

The following could be a profile for potential users of a home improvement site. It defines a group of people who are actively thinking about home improvement and would be likely to look for information about it online.

A SAMPLE FOCUS GROUP RECRUITING PROFILE

Demographics

Ages 20–55
Income not important if behavior criteria are met, otherwise, $100K + household.
Gender unimportant, as long as primary decision maker.

Web Use

Have a personal computer at home or work.
Use Internet at home or work.
Have 1 + years of Internet experience.
Use the Internet 5–10 hours per week for personal tasks: Shopping for products Comparing products or services Gathering information about a topic (e.g., Sidewalk, BabyCenter, ImproveNet, CarPoint)

Behavior

Have completed home improvements in the last 9–12 months or intend to perform improvements in the next 3 months.
Total improvement cost at least $20K (estimated if in the future).

Defining Scope

Focus groups must be small enough for everyone to have the opportunity to share insights and yet large enough to provide a diversity of perceptions.

—Richard A. Krueger, Focus Groups, p. 17

You need to decide two things when determining the scope of the focus groups you're going to run: how many groups you're going to have and how many people you're going to have per group.

Never do only one group for the same reasons you shouldn't base conclusions about a class of people on the words of a single person: there may be factors that seem important to the group, but that are not representative of the views of your audience at large. That said, more than four groups is rarely necessary. The first group is, essentially, a dress rehearsal. By the third group you should see confirmation of the views and statements (sometimes verbatim) from the first two groups. The fourth group should be used for confirmation of the trends in the first three. If there is still a lot of new information (especially dissension) coming in during the fourth group, it may be a sign that further groups are necessary or that the makeup of the groups is too broad.

Likewise, although most marketing texts recommend 8–12 people per group, it's generally better to have fewer when examining user experiences and attitudes. The degree of detail that is appropriate for user experience research is often greater than the typical marketing focus group. By reducing the size of the group, it's possible to go into greater depth with each person. Six to eight is a good size, balancing the collection of fine detail with a breadth of perspectives. In cases where you'd like to get a lot of depth from each participant, or if the individual focus groups are short, it may be useful to reduce the number to four. Use fewer than that, and the discussion feels more like an interview and doesn't produce as dynamic a situation as with more people.

Choosing Topics

For an average focus group, you should have three to five main topics to investigate. These should be phrased in terms of the project as a whole. "Understanding the mental model people use when researching insurance" could be a goal for an insurance brokerage site, while a service that recommended home building contractors could be interested in "Knowing at which point people turn to an external service when doing home repair." A high-end online auction site doing a competitive analysis could have "Understanding what repels high-end car sellers from listing their cars on eBay" or "Uncovering what factors will help our product idea to be seen in the same class as Sotheby's" as two of their goals.

These objectives should be sufficiently focused so that a group could adequately discuss each one in about 10 minutes. They should not be so focused as to be the actual questions you would ask the participants, nor should they be issues that are better answered by other means (such as a survey)."Making a list of our competitors" would generally be too broad and probably better answered by a survey, whereas "Discovering what factors make the user experiences of our competitors more compelling than ours" is probably more appropriate.

Note

Not all people are comfortable talking about all topics. Choose questions that your target audience will be comfortable discussing. Different audiences for the same product may react differently to the same topics.

Writing a Guide

The discussion guide is a script for the moderator to follow. It creates a consistent framework and a schedule for the focus group series. Groups are asked the same questions in the same order with much the same context, and all the topics are given enough time. This allows a discussion to bring out the subtleties of the participants' views without shortchanging any of the topics.

Before you start writing the whole guide, it's useful to think about the core, the questions. Focus group questions should be

Carefully ordered. Questions put the participants in a certain frame of mind, thinking about certain issues and remembering certain events. A careful sequence of questions takes advantage of that frame of mind to make the flow of the group discussion feel more "natural," which in turn helps the participants to maintain a creative flow of ideas and produce better insights. In general, questions should flow from the most general to the most specific, with each question narrowing the discussion a bit and concentrating on a subset of what was discussed before. There should be planned transitions between topics unless a brand-new topic is introduced and discussion begins from a general question again.
Nondirected. As described in Chapter 6, questions should not imply an answer or a value judgment. They should focus on allowing the participants to fill in their own thoughts and values. For example, asking "Which do you think is a better search service, Google or Lycos?" assumes that the participant feels there are advantages of one over the other. Instead, the question should be framed neutrally," Are there any things you like about using the Google search service? Are there things you like about using Lycos? What are they? Are there any ways in which you can compare them? How do they compare?"
Open-ended. Questions should not constrain the answers to fixed responses. They should encourage people to open up and share experiences. Longer responses tell a greater part of the story and tend to be less ambiguous than shorter responses. Thus, rather than phrasing a question in the form "What's your favorite recipe site?" you could ask "What are you looking for when you search out recipes online?"
Focused on specifics. The questions should encourage the participants to be specific in their answers. Richard A. Krueger, in his book Focus Groups, recommends breaking down "why" questions into multiple "what" questions, explicitly asking for the influences that informed the participants' decision and the attributes of their decision. For example," How did you decide to go shopping online for forks?" and "What factors went into picking this site?" will provide better insight than asking "Why did you pick this site?"
Personal. People will often attempt to generalize their experiences to the public at large or some hypothetical audience that they are not part of. Since you want to know individual views, values, and experiences, emphasize individual experiences. Questions should be formulated so that they concentrate on people's current behavior and opinions, without presenting the option to project their experiences. Thus, "If you had to redo your kitchen right now, which of these features would you use to find a home contractor?" is preferable to "Which of these features do you think are useful?"
Unambiguous. There should be as few shades of meaning as possible, especially when the participants are introduced to new terminology.

Granted, fulfilling all these criteria with all questions is often difficult (writing questions that are simultaneously specific and open-ended is a particularly tricky challenge), but they should be kept in mind as guidelines that should be followed whenever possible.

Sample Discussion Guide

The guide is broken up into three major sections: the introduction, the main discussion, and the wrap-up.

The guide that follows is from a focus group for an online news site that is (primarily) interested in understanding the criteria its existing users use to pick the news sites that they read.

Warm-up and Intro (10 minutes)

The introduction sets the tone for the discussion, breaks the ice for the participants, and explains the process.

Hi. Welcome.

My name is _______________ (first name). I am a researcher working with _______________ (company), who has asked me to help them get some of your thoughts and opinions about some products and some ideas they have. I am not associated with the development of any of the products we'll be talking about today, and I have no emotional attachment to any of these things, so you can say whatever you want.

We invited you here because you all read a fair amount of news online and you've spent a lot of time on the Web. What we're going to do today is talk about some of your experiences so that they can create a service that is best tailored to experiences of people like you.

Telling people how they were chosen helps them feel comfortable with one another. Informing them of the end goal of the research helps them start focusing their thoughts.

The discussion is going to be a pretty casual conversation, but there will be a couple of times when I will ask you to concentrate on certain things.

While we're talking, it's important that you be as candid as possible. You won't hurt anyone's feelings with anything you say, so please say exactly what you feel.

Furthermore, we want YOUR opinion. No opinion is right or wrong here—especially about the things we're going to talk about—it's just an opinion, so even if you disagree with someone in the room we'd like to hear that.

But we'd like you to speak one at a time.

Set out the ground rules for conversation ahead of time and explicitly allow disagreement. This encourages people to feel comfortable voicing their thoughts later.

Also, since we have a lot of ground to cover today and none of us want to be here for hours, I may have to ask you to wrap up a thought or put it aside so that we can move on.

Again, people don't take being interrupted as personally if the expectation has been set early on.

Behind that glass wall, as you can imagine, are a couple of people from the company whose idea we'll be talking about and _______________ (assistant moderator's first name) who is working with me. Occasionally, _______________ may come in here with a note or something we need. Regardless, feel free to ignore them.

Explicitly noting the mirrored wall (assuming there is one in the room) acknowledges the participants' anxiety about it and helps to diffuse that anxiety. Sometimes it's even appropriate to have the participants wave or make faces at the people behind the mirror during the introduction. However, once mentioned, it shouldn't be brought up again.

As I said, we brought you here to hear what you think, so you won't hurt anyone's feeling by whatever you say. This videotape, in case you're wondering, is here so that _______________ and I don't have to sit here feverishly scribbling notes and can concentrate on listening to you. It's purely for research purposes. It may be seen by members of the product development team, but it's not for any kind of publicity or promotion or broadcast.

If the session is videotaped, that should be mentioned.

Now I'd like to read you what's called a statement of informed consent. It's a standard thing I read to everyone I interview. It sets out your rights as a person who is participating in this kind of research.

As a participant in this research

You may stop at any time.
You may ask questions at any time.
You may leave at any time.
There is no deception involved.
Your answers are kept confidential.

Here is a form that gives us permission to videotape this discussion and to use the videotape in our research.

It's critical to inform participants of their rights and to get releases that allow you to videotape them. Doing otherwise is both unethical and, in some situations, illegal. Don't be dour when introducing these ideas, however, since it can create an overly formal atmosphere right off. I usually joke that what we're talking about is what we're interested in, and (pointing at the snack dish) that this isn't a secret candy taste test.

Any questions about any of that? Let's start!

Now I'd like all of us to introduce ourselves. By way of introduction, I'd like you to tell us four things.

Your first name
Which city you live in
What TV shows or publications you absolutely can't live without every week
Anything that regularly irks you about the Web

(______ does introduction first, everyone goes around).

The introductory questions introduce everyone and break the ice by sharing something that's fairly personal, such as a favorite TV show or a pet peeve. As such, it's appropriate for these questions to be specific, (somewhat) directed, and unrelated to the topics of inquiry.

General News Reading (20 minutes)

How do you get the news most frequently?

Probe: Are there some other ways you regularly get news?

Which news sites do you regularly read? (make a list on the whiteboard)

Probe: Are there some that you read more than others?

Probe: What attracts you to the sites you read more often? (Ask individuals about specific sites.)

Have you ever switched from one favorite to another? How often (did it just happen once or have you had several favorites)?

Probe: What made you decide to switch?

Probe: How did you pick the site you switched to?

Probes are follow-up questions that dig deeper into a given topic.

Switching gears, what kinds of news are there? (make a list).

Are there any good sites for sports? How about for politics? Technology?

Even though the primary topic is technology news, asking about several different kinds of news keeps the discussion from focusing on technology prematurely.

Are there sites that specialize in any of these (business, entertainment, technology)? Can you name some, even if you don't read them regularly?

Changing gears again, have you ever come across a situation where you think that advertising has affected the news you're reading?

Probe: Can you give an example?

Prioritization Exercise (20 minutes)

Pass out paper with the following exercise:

Rate each of the following things based on how important they are to you. Rate them from 1 to 5, with 1 being least important. Also, if you can think of a site that does that one thing better than any other site, write its name or URL next to the statement. You don't have to put a different site down for every one, and you may not want to put any sites down for some.
- The number of different stories on a given topic
- The number of different topics covered
- How quickly the page downloads
- Who is presenting the news
- How comprehensively each story is covered
- Regular columnists
- The quality of the site's search engine
- The visual appearance of the site
- How quickly stories are covered after they happened
- How easy it is to get around in the site

(going around) Say which ones you picked as the most important.

Probe: What about it makes it so important?

Probe: Is there a site that does it well? Does anyone else have that site written down?

Are there any attributes of news sites that you think are important but that are not on the list?

This exercise is designed to start a discussion about what makes a good news site by focusing on specific features.

Competitor Site Review (20 minutes)

We've been talking about general news reading, but let's talk about a couple of specific news sites for a few minutes.

(turn on projector)

Are any of you familiar with _______ (competitor site)?

Probe (for people who are): Have you ever read it? How often? Under what circumstances do you read it versus other news sources?

Describe ______'s personality.

Does _______ fall into any kind of specialization?

What kinds of stories would you expect to find here?

Probe: Are there situations where you think you would read it versus other sites?

Probe: Are there kinds of technology news that you wouldn't go here for?

Probe: What about columns? Are there any that would entice you to read them on a regular basis?

(go through questions again for another competitor; present competitors in a different order to every group)

Blue Sky and Wrap-up (10 minutes)

We're almost done. I'd like to do a quick brainstorming exercise. I only have one question, but I'd like everyone to think about it for a few seconds and say what comes to mind.

Not thinking in practical terms at all, what would you really like a news service to do that none of them currently offer?

(wait 30 seconds, then go around and discuss, write ideas on board)

All right, that's the last of our questions about news.

I know you're all excited about talking about online news, but we really have to wrap up. If there's something you think of on your way home that you'd really like to tell us, please feel free to send a note to the following email address (write email on board).

I have one final question: is there anything that we could do better, either in terms of scheduling or in terms of running these groups? Should we have different food, a case of beer, whatever?

Once the content for the site had been established, subsequent focus groups can concentrate on the desirability of specific features and how people use news sites in their daily lives.

After you've written the guide, it's important to review it and time it. Members of the product development team, especially the product and project managers, are good guide reviewers. They can point out technical ambiguities, serve as a second perspective, and then prepare the rest of the team for observing the groups.

Guides should always be tested. An easy way to test the guide is to get a couple of people (it doesn't have to be as many people as the actual group) who haven't seen it and walk them through it, paying attention to how they answer the questions and how accurate the timing is. In addition, treat the first group as a dress rehearsal, reviewing the effectiveness of the guide and adjusting it appropriately.

Conducting the Group

The Physical Layout

The groups should be held in a comfortable, large room with good ventilation, good temperature control, and few distractions. Typical setups look like conference rooms or living rooms, depending on whether the participants are supposed to feel they're in a comfortable work area or at home. Actual conference rooms or lounge areas are easily converted to focus group facilities. If the groups are to be held in a conference room with windows to the inside of the office, the windows should be covered or the group should be held as far away from the windows as possible. Outside windows are fine, unless they're on the first floor, in which case they can be too distracting. After the group has started, no one except for the assistant moderator should enter the room (and a sign should be posted on the outside door to this effect). I once observed a focus group that was mistakenly interrupted by a pizza delivery for the observers. After some initial confusion, it became clear that the pizza was for the observers. Although the participants joked about it, it set up an unfortunate dynamic.

Eating is an informal activity that can break the tension in a group that's just forming, so provide food and drink when possible. For late evening groups, people are likely to have eaten dinner, so providing high-energy snacks (such as cookies) can often keep a sleepy group talking. Make sure to provide vegetarian choices and plenty of water. Crunchy food can be quite loud when recorded, so things like celery and potato chips are best avoided. Having plenty of noncarbonated drinks is handy; people drinking carbonated beverages have a tendency to stop themselves from commenting because they're afraid they'll burp in the middle of a sentence.

How to Create a Seating Order

Krueger recommends the following procedure to place people around the table:

Print table tentswith the participants' names on them. Table tents are folded sheets of paper that sit upright on the table. Each participant's first name is printed on both sides of the tent, large enough that it can be read from the observation room and on camera (1.5-inch letters printed in black seem to work pretty well).
Determine who are the quiet and outspoken people by visiting participants in the waiting area and making small talk with the group.
Figure out a placement for each person (long-winded people next to the moderator, so he or she can break eye contact without disturbing the group dynamic; quiet people across, so the moderator can make eye contact easily and elicit their comments).
Organize the table tents according to this order.
Then, when bringing people into the room, drop the table tents seemingly randomly on the table, but actually according to the predesignated order.

The room should have a way for observers to watch the proceedings. This can be a traditional two-way mirrored wall, with a soundproof room behind the mirrors that observers can sit in. The observer's room should have its own entrance, so that observers can come and go without entering the discussion room. If such a room is unavailable, closed-circuit video is an inexpensive and easy solution. A long video cable and a television in a nearby room generally suffice, as long as the observation room is acoustically isolated from the discussion room.

Focus groups should be videotaped, when possible. Unlike an audiotape, it's easy to see who's saying what on video and to see the body language that sometimes betrays a negation of a concept or discomfort with a topic. If the video process is made unobtrusive, it quickly disappears into the background of the conversation.

There needs to be only one camera for a focus group, roughly behind the moderator, and it needs to have great sound. If it has a good wide-angle lens, it's not necessary to have a camera operator, which reduces the cost of the process, as well as the hassle and intimidation of having a person swinging a camera constantly between one side of the table and the other.

Two cardiod microphones are generally sufficient to capture all the participants' comments without a lot of extraneous noise. Cardiod microphones are directional microphones that have a broad area of sound capture in front and a much smaller area facing back. Thus, unlike the omnidirectional microphones that are attached to most video cameras, they can be pointed away from sources of noise such as outside windows and air vents (see Figure 9.1).

click to expand
Figure 9.1: Cardiod versus omnidirectional microphone sensitivity.

For temporary setups, two opposing ones set on the table work well, but they're vulnerable to vibration (so a pencil tapping or a coffee cup being set down or a computer humming will sound much louder to the microphone than they do in the room). Putting some cloth or a mouse pad between the stand and the table can reduce this. For permanent situations, two hanging microphones eliminate most ambient noise from above (such as air conditioners and fluorescent light transformers) while recording conversation well (see Figures 9.2 and 9.3).

click to expand
Figure 9.2: Temporary conference room-style focus group layout.

click to expand
Figure 9.3: Permanent living room-style focus group layout.

The Moderator

Focus group moderation is a skill. The moderator must balance the needs of the participants in the group so that they feel comfortable in the discussion and the needs of the research. To do this without biasing the group takes practice, aptitude, and the right frame of mind.

The basic skills that a moderator must have are a respect for the participants, the ability to listen closely, and the ability to think fast. Often, a moderator must be able to predict where a conversation is headed and either drive it that way or move it in a more desired direction, without the participants realizing that they are being moderated. This can involve many subtle cues in what the moderator says, the tone he or she uses, and even his or her body language.

More specifically, the moderator must be

Always in control. Most of the time the moderator can use body language and verbal emphasis to maintain control, subtly directing the discussion to certain participants and topics. However, if a digression is moving in an unproductive direction for too long, the moderator should not hesitate to exert more overt control and refocus the discussion on the necessary topics. For example, if in the course of discussing a news Web site, the discussion digresses into the current headlines, the moderator should refocus it on the product rather than on the content.
Always moving forward. The moderator should monitor the flow of the conversation and introduce topics at appropriate times, making the transition feel natural rather than controlling the flow by stopping discussion or abruptly changing the topic. That way the discussion doesn't turn into an interview, with the participants answering the questions the moderator poses, one after another.
Nonjudgmental. The moderator acts as mediator, helping the participants express their views without stifling their eagerness to do so. Therefore, the moderator should not express his or her views, but facilitate the views of the group to come out. This sometimes involves suppressing the habits we have learned in maintaining civil conversation. For example, many people nod while someone is talking as encouragement, whether they agree or not. This communicates agreement and an endorsement of the viewpoint being expressed. Since the moderator is seen as the "official" at the table, the participants may feel that this is official endorsement of a certain position and, as such, may feel uncomfortable voicing a dissenting opinion.
Respectful. The moderator must have the utmost respect for the participants at all times, even if he or she does not agree with them. Every participant in a focus group has a perspective that's useful to the development team, even if it doesn't match with the team's reality. For all participants to feel comfortable speaking their minds, they need to know that the moderator will treat their input with as much validity as everyone else's. This can be accomplished through the use of nonjudgmental statements and strict control of body language, but it's best communicated through the honest belief that everyone needs to be heard.
Prepared. The moderator should know enough about the topics discussed to be able to follow up on participants' statements with specific questions. This does not mean acquiring expertise in the subject matter, but the moderator should have a good working knowledge of the general concepts, the terminology, and the implications of the issues being discussed. The moderator should also have some knowledge of the habits and environment of the people in the group.

In addition, the moderator should have a good sense of timing (knowing when to transition from one participant and one topic to another), a good short-term memory (referring to people's earlier statements and using their own words to describe concepts), and an aptitude for diffusing uncomfortable situations, preferably with humor.

The Assistant Moderator

Although it's certainly possible to conduct a focus group without an assistant moderator, having one helps the process.

The assistant moderator is a key analyst and the connection between the focus group and the outside world. The assistant takes care of the needs of the focus group participants and collects information, leaving the moderator to focus on maintaining a productive discussion.

Before the discussion begins, the assistant should greet the participants when they first come in, bring them into the discussion room, present any initial paperwork (such as nondisclosure participation agreements), and bring them refreshments. As the discussion progresses, the assistant can bring in notes from the observers and take care of requests.

During the discussion, the assistant moderator should take extensive notes on the interesting parts of the discussion (such as key quotations, issues voiced by the participants, and his or her own observations) and manage the observers' discussion (taking notes on that, too). After the group ends, the assistant moderator can use these notes to spur a debriefing with the observers and the moderator.

Note

If there is no assistant moderator, the burden of moderating the participants, managing the observers, and analyzing the final results falls on a single person. In such instances, it's useful to create situations where the moderator can leave the participants alone for five to ten minutes and walk back to the observation area to review the discussion with the observers, helping frame the discussion for them.

Moderating the Discussion

Moderating a discussion is the process of balancing the participants' comfort level and keeping the discussion producing useful information for the research. There are few strict rules for how to moderate a discussion since every group and every topic will demand a different approach, and different moderators have different styles that work equally well.

There are, however, several general guidelines that apply to most user experience focus groups and that most moderators seem to follow.

Spend some time with the group beforehand. Talking informally to the participants for five minutes before the group begins can give you a good feeling for the personalities of the participants. You can guess who is going to be quiet, who is going to dominate the discussion, who is going to be obstinate, who is going to lecture, and so on. This helps prepare you for how the group is going to progress. The assistant moderator can also do this and brief the moderator on the personalities just before the group begins.
Stick to the guide, but be opportunistic when necessary. If the discussion turns to a question that's supposed to be asked later, but the participants are comfortable talking about it now, go with it. Ask the questions that were planned for later (unless order is absolutely critical, in which case ask the group to postpone the discussion), and after the topic is exhausted, move the discussion back on track.
Engage everyone in the discussion. Be wary of people who are losing interest or who don't seem comfortable speaking. Try to figure out what is causing their boredom or discomfort and attempt to alleviate it. Sometimes all it takes is to call the person's name and ask explicitly what his or her thoughts are. When the amount of effort needed to engage someone in a discussion is detrimental to the effectiveness of the discussion as a whole, however, it's best to spend the energy on getting the rest of the group into the conversation instead.
Avoid introducing new terminology and concepts. New words and ideas tend to frame and dominate discussion once introduced. If the group wants to use its own words for a concept or to create its own models for ideas, let it. Don't use industry jargon, even if it sounds as though it should be completely comprehensible, unless it's also the group's jargon. For example, asking "What factors influence your choice of ecommerce portal?" will likely be met with confusion since most people don't characterize the reasons for their choices in terms of "factors," and "ecommerce portal" is likely going to be an alien term.
Restrict your body language. There are lots of body motions that subtly communicate a preference for a particular person's viewpoint or a particular viewpoint, and people are very good at picking up on them, even if they don't do it consciously. Nodding signals to the other participants that the view being expressed is "right." Wincing signals that it's wrong. Consistently leaning toward one participant says that person's comments are more important (even if you're doing it because he or she is speaking quietly). You may want to watch videotapes of yourself moderating to see what your natural body language communicates and then try to reduce the behaviors that communicate a preference.
Clarify comments. When someone says something that's either unclear or contradicts an earlier statement, restate the idea or ask for clarification. Ask for examples of what the person means or definitions of terms he or she is using.
Restate ideas. To make sure that everyone (including the moderator and observers) knows exactly what a participant is trying to say, it's often useful to restate what was just said and see whether it correctly represents the participant's ideas. When heard in different words, the core of the idea is clarified. The technique also reinforces the relationship of the moderator to the group, subtly communicating to the participants that although they're the experts, the moderator is in control.
Probe for alternative perspectives. If someone says something that contradicts earlier statements, or if he or she introduces a new interpretation of what's being discussed, probe the group for agreement or dissent. Let the person finish his or her thought and then ask the rest of the group to volunteer thoughts or experiences, especially different experiences or thoughts on the same topic.
Don't dominate the discussion. Let the group carry on the discussion as much as possible. When one participant makes a comment, give the rest of the group some time (Krueger recommends five seconds) to respond to the comment before asking follow-up questions.
Provide time to think. Not everyone can immediately articulate his or her thoughts or instantly grasp a concept. Explicitly providing opportunities to think or remember before a discussion begins can help participants contribute. Five to ten seconds is all that's generally necessary and can be introduced as the question is asked, "Remember the last time you ordered spare parts? Think for a couple of seconds about what situation led up to that."
Use humor when appropriate. Don't make fun of the topic, of the participants, or of their statements, but feel free to lighten a situation with humor. Don't try to be funny—it seems artificial and sets up the expectation that you're there to entertain—but when something is funny, laugh and don't be afraid of saying something that's funny. Use self-deprecating humor carefully. Humility is valued, but only if it doesn't undermine the subject or the situation.
Keep the energy level high. There will be times when the group is meandering or isn't interested in a particular topic. Maintaining (and inspiring) a high energy level can salvage such a group from boredom and frustration.
Give people a break. Sitting in a room for two hours in intense discussion can be difficult and draining. If a focus group runs for more than 90 minutes, give people a 5-minute break halfway through to collect their thoughts, put change in the meter, go to the restroom, and the like.

Finally, be relaxed. Smile. Really, truly, empathize with people's experiences. Enjoy the opportunity to share in people's lives. Have fun with it.

Asking Questions

The process of asking questions is a subset of moderating, but there are enough elements that strictly relate to questions that I feel it deserves its own section. Several techniques help in asking questions and in clarifying the responses to questions.

Prioritize issues. When coming up with lists of issues, always have participants explicitly prioritize, either through voting on the most important thing on the list (randomize when walking through the list!) or through a spontaneous finger tally for each item (everyone holds up a hand rating the item, the moderator quickly comes up with a rough average of the group's ratings).
Write down opinions before discussing them. When people have written something down, they're much less likely to change their minds and "go with the flow" during discussion. For example, participants who first rate a list of topics on paper have more individual and earnest responses in a subsequent group discussion of the ratings.
Use the participants' exact thoughts and words, when possible. When the group has introduced terminology or defined certain concepts, modify follow-up questions to use those same terms and concepts.
Key questions need to be asked verbatim. For secondary questions, it feels more natural to use the same terminology as the group, but main concepts should be presented in the same way to all groups in order to reduce doubt when analyzing the results.
Be ready with examples. Questions may not always make sense to participants, especially if the concepts are new or the terminology is unfamiliar. Prepare several good examples for questions where you feel that may be the case. First try to get responses without the example, but if people aren't responding well, provide the example. For instance, if the group doesn't know where to start a feature brainstorm, give them an example of what is considered a new feature ("The site emails you whenever items you like go on sale.").

A Feature Buying Exercise

Chauncey Wilson, Director of Bentley College's Design and Usability Testing Center, suggests the following technique for feature prioritization: Give people a "menu" of features on a page with a description of each feature and "price" that's based on an estimate of how much each feature costs to develop, in fictitious dollar amounts ($100 for something that's relatively straightforward to implement, $500 for something that's very involved). Then tell each person to review the feature list and "buy" up to $1000 dollars worth of features, writing them down on a piece of paper. After all participants have made their choices, go around and discuss why people made certain choices.

Common Problems

There are several problems that occur in many focus groups.

Group-think. People are conditioned (or predisposed, depending on whose theories you read) to avoid conflict when in small groups. One of the best ways of avoiding conflict is by rationalizing agreement with the apparent consensus of the group. Thus, people have a tendency to want to agree with the people in the room, even if they don't really agree with the conclusions, because "belonging" is felt to be more important than "being right." Many of the techniques described earlier are designed to minimize group-think, but it still happens. Always be on the lookout for it, and attempt to minimize it by making it clear to the participants that disagreement is not only tolerated, but encouraged, when it's in earnest.
Faulty assumptions. There are situations where the development team and moderator have fundamentally misunderstood the nature of the user experience, and the whole focus group is based around an incorrect assumption. The best way to avoid this is by interviewing a representative of the target audience beforehand. If the focus group series has already begun and this is discovered, then the guide should be rewritten as soon as possible and, if necessary, new participants scheduled.
Latecomers. Unless the introductions are still going on, don't let people in once the group has begun. Give them their full honorarium and say good-bye. The process of bringing people up to speed is disruptive, and the presence of a "stranger" puts people back on guard, negating the effects of the ice-breaking that happened earlier.
Reticent groups. Sometimes groups of people just don't want to talk. The reasons for this vary from group to group, but sometimes you find yourself in a room full of people who aren't really interested in talking about the topic. Depending on the situation, it may be possible to reintroduce energy into the discussion and still keep with the guide and schedule. Sometimes, though, it's necessary to take drastic measures if the group isn't discussing the topic at all. At this point, the creativity of the moderator comes into play, and he or she has to invent a way to get the participants talking (for example, by turning the session into an open-ended brainstorm or moving on to questions that the group seems to be more interested in).
Quiet participants. The reasons certain people don't talk are varied. They may feel uncomfortable with the subject, or that their opinion is not shared, or that they have nothing to add. They may just be shy. Getting everyone's perspective is critical, so drawing out taciturn participants is important. The easiest way to have people talk is to call on them directly, but when over-done it sets up a "call and response" dynamic to the discussion. Asking the whole group for additional comments, but looking directly at the quiet participant, signals that the comment is directed to him or her, without communicating that this person is being singled out. Finally, nonverbal signals can communicate when someone has a thought even though he or she may not feel comfortable voicing it. The inhalation before speech begins, a lean forward toward the table, eye contact with the moderator, a wince when another participant says something, all of these can be indicators that there's a thought brewing.
Overly talkative participants. Some people just don't know when to shut up. They generally come in one of two breeds: the person who has an interesting idea, but is having trouble articulating it, and the person who has no idea and is hoping that talking will make one appear. The former can be overcome by judicious interruptions and restatements of the participant's position; the latter is somewhat harder. When people are clearly talking without purpose, ask them to wrap up their thought and move on, calling on someone else or asking a different question. If they continue to talk aimlessly, politely remind them that they need to let some of the other people in the room speak.
Group dominance (the "Alpha Jerk" effect). A single dominant or bullying participant can ruin the utility of a focus group. Ideally, the best way to avoid the intimidation caused by such a person is to remove him or her before the group even starts. Such behavior patterns can be obvious when meeting and observing participants in the waiting room. However, once the group is going and it's obvious that there's a person who is vying for the "control" of the group, it's critical to draw attention away from him or her and into the group as a whole and to reinforce the notion of the moderator as the driver of the group. The techniques are basically the same as for a talkative person, but they need to be applied more firmly (and often, more frequently) than for someone who's merely wordy. In addition, physically dominating the discussion space by standing up and walking to a whiteboard helps to regain the floor and control of the discussion.
Unqualified participants. Sometimes people misunderstand what the participation criteria are, or they misrepresent their experience. Weeding these people out early is important since they will likely not be able to contribute to the discussion in a meaningful way, and their presence will tend to make the other participants uncomfortable. Asking the key screening criteria early in the session and gently requesting participants who don't match to leave the room may make the rest of the group much more productive on the whole.
Tangents. Groups will have a tendency to spiral off on tangents. These can be useful discussions of values and ideas, which when important to the topic, should be allowed to run for a while. When they're on points of marginal interest, they should be wrapped up as quickly as possible. After noting that people seem to be interested in a topic—no topic should ever be discouraged flat-out—the moderator should encourage the group to wrap up the discussion or to postpone it.
Hostility. Open hostility in focus groups is rare, but certain people (and certain ideas) can cause it. Vehement disagreement or offensive ideas can lead people to grow hostile. At this point, the moderator should remain impartial and focus the discussion on the ideas rather than the emotions. Thus, a statement full of emotional content should be treated as if it were emotionally neutral. For example, if someone says another person's statement is really dumb, ask what it is that he or she disagrees with rather than dealing with the concept of "dumb."
Offensive ideas. Occasionally, someone will say something that offends the majority of the other participants. When this happens, either it can be used as a trigger to get all participants to share their opinions or it can polarize the group. To guide the discussion in the former direction, rather than letting it degenerate into partisanship or the ostracism of the speaker, ask if anyone else shares that view. Then ask if anyone has a different view and probe how their views are different. It's especially important for the moderator not to side with one view or another—even if you strongly disagree with the stated view—and concentrate on the facts of the statement.
Highly emotional topics. Much like controversial topics and group hostility, emotional topics can get tricky. Again, the best way of dealing with them is by focusing on the ideas behind them, as long as emotional release doesn't become the only topic. This should be taken into account in the guide, and the participants should be given sufficient time to discuss their emotions if emotional responses are expected. If a topic unexpectedly veers in an emotional direction, the emotions of the participants should be respected and discussed, even at the detriment of a couple of questions in the guide. However, the moderator should move away from such a discussion as soon as possible while trying to get enough information about the experiential factors that led to the emotional state.

Managing Observers

You can extrapolate in any direction from one point, but two points determine a line.

—Carolyn Snyder, user experience consultant and principal, Snyder Consulting

As many people from the development team should attend as many focus groups in a series as possible. This gives the team instant information about their ideas, and it provides the analyst access to valuable technical expertise. For the same reasons that there should be more than one group per topic, people should be encouraged to observe at least two groups if they're planning to observe any. It's much easier to know which phenomena are unique to a given group of people and which may be more general phenomena when you've seen several groups discussing a given topic.

Since observers are an active part of the analysis process, it's important that they know how to approach the observation process. There's a right way to observe a focus group and many wrong ways. Make a point of meeting with new observers beforehand and prepare them for observing the focus group. The following instructions help prepare observers and can be presented as part of the initial orientation or as a written list.

FOCUS GROUP OBSERVER INSTRUCTIONS

Listen. As tempting as it is to immediately discuss what you're observing, make sure to listen to what people are really saying. Feel free to discuss what you're seeing, but don't forget to listen.
Don't jump to conclusions. Use people's statements as guides to how they think about the topic and what their values are, but don't treat the specifics of their statements as gospel. If everyone in a group says they like or hate something, that doesn't mean that the whole world thinks that way, but it is a good indicator that there are enough people who do so that you should pay attention.
Focus groups are not statistically representative. If four out of five people say something, that doesn't mean that 80% of the population feels that way. It means that a number of people may feel that way, but it doesn't mean anything in terms of the proportions found in the population as a whole. Nothing. Zilch.
Focus group participants are experts. The participants in a focus group know what they want to do and how they currently do it. Listen to their needs and their experience, and treat them as consultants who are telling you what your customers need, not as the consumers of a product or the targets of a sales pitch.
Focus groups are not a magic bullet. A couple of good ideas from every group is enough to make that group worthwhile, but not every statement that the group participants make should be followed to the letter.
Feel free to pass questions to the moderator, but don't overdo it. Occasional questions to the group are OK, but there should not be more than a couple in a session. Write your question clearly and concisely, and phrase it as if you were talking to the moderator. Then give it to the assistant moderator, who will then take it to the moderator. When appropriate for the flow of conversation, the moderator will introduce the question. However, the moderator may decide never to introduce it, if the timing or topic is inappropriate.
People are contradictory. Listen to how people are thinking about the topics and what criteria they use to come to conclusions, not necessarily the specific desires they voice. A person may not realize that two desires are impossible to have simultaneously, or he or she may not care. Two people may think they're agreeing, when they're actually saying the exact opposite.
Don't write people off. Sometimes, a participant may say things that indicate that he or she isn't getting it. Never assume that someone has nothing important to say just because they aren't interesting or insightful from the start. Understanding why one participant "doesn't get it" can hold the key to understanding the perspectives of everyone who "does."
Save some pizza for the moderator.

Tips

On rare occasions, it will be necessary to eject someone from the group because he or she is disruptive to the discussion or completely unqualified. The assistant moderator should do the dirty work and ask the person to leave, but only when both the moderator and the assistant agree that continuing the group with the person is impossible. In such a case, a good way of approaching the process is to find a natural break in the group and ask the person to leave. At the beginning, a demographic form or a nondisclosure agreement is a good time to duck out and discuss it with the assistant moderator. Later on, there can be a "5-minute stretch" or a similar pretext for breaking the action. During this time, the assistant moderator should come into the room and politely ask the person to go with him or her. The person should then be thanked for participating, given the promised honorarium, and excused. Everyone will know what happened, but the excitement of seeing someone "get kicked out" will quickly fade.
Sometimes it's useful to have multiple moderators. If there are a lot of groups, or they're particularly long, or they're done in different languages, it's sometimes useful to use several people as moderators. If multiple people are going to be moderating, it's important that they go through every question and determine how they're going to ask it and what they will probe on.
Two-way pagers or wireless chat devices (such as the Cybiko handheld) can be used instead of having the assistant moderator running in with notes. The moderator can have the device on the table and occasionally glance at it to see if there are new comments or questions from the observation room, where the assistant moderator (or one of the observers) operates another of the devices. When using a pager, use one that doesn't get a lot of messages so as not to distract the moderator, and turn off sound and vibration to keep new message alarms from interrupting the group.

Hiring Experts

Focus groups are an involved process, and you may choose to hire a company that specializes in organizing them rather than expending the time and energy to do the process in-house. There are several things to consider when hiring a company to run your focus group.

Does the moderator know the subject? A moderator needs to be able to think on his or her feet about the subject and probe appropriately. Thus, the company should have some experience with either the specific subject or one that's closely related. If he or she does not, then you should prepare to brief the moderator in depth on the subject, the terminology, and the issues associated with the topic.
Can they recruit the right people? Getting the right people for the focus group is critical to that group's success, thus the recruitment method is important. Ask the company how they get their database and how they will screen people for your focus group. Make sure to review the screener.
How long is the turnaround time? The analysis method varies from company to company. Depending on the final deliverable, it can be from days to weeks. Make sure to find out what they are planning to do and how long it will take. You may want to forgo a fancy report and slide presentation for a list of findings and a discussion with the moderator and analyst.

As with any contract work, ask for references to the last couple of projects that the company has completed and follow up on them. Do not be shy about asking for satisfaction. If anything is amiss—the recruiting, the moderating, the analysis—call them on it. If necessary, request additional groups be run at their cost.