This is general learnings from qualitative research that I followed during the last 2 months, involving more than 10 times of FGD (Focus Group Discussion) from several market research agencies (Mindspace, Deka, Synovate, and one research agency that I forget). I also experienced the wide range purpose of these research from concept test until packaging test. The specific findings is secret but I believe I can share to you the general findings, this is few of it:
1. Moderator is very important: If you conduct a qualitative research, in this case FGD, make sure your FGD moderator is good. This is the most significant KPI that make the whole process success or fail. If you want a little detail you can also try to find who is the research manager behind your moderator. Because that research manager is the one who briefing your moderator, it is important to make sure that your agency’s marketing manager and moderator is brand-minded, have good knowledge about your industry, and have exceptional capabilities in directing the flow of the discussion.
2. Respondents are very important: respondents should be represent your target respondents. I find quite bad experience my self when one of the research agency is done a fatal blowing mistake by letting two of respondents is not our intended respondents. But because of the project itself is a regional project and lack of coordination with our side. A important learning, ensure your coordination with all the stakeholders when you conduct your research. The number of respondents is also critical in the process. I found that 8 respondents is too much to be handled by a moderator. There is always blind spot (silent respondent that the moderator forgot to explore) in 8 respondents FGD’s, but again this is very dependent on the skills of the moderator. In my first FGD’s I see that the moderator is quite superb and didn’t find any difficulties in managing 8 respondents FGD’s but again if you don’t sure with your moderator qualities then 6 respondents is a safe bet.
3. Stimulus is important: stimulus is a support tools that help respondent in doing their job, giving us meaningful information. In packaging test, the packaging can be shown to respondent so they have clear experience on how is the real packaging. Others stimulus including: video, enlarge picture, magazine or newspaper and etc. If you or your research agency can provide the stimulus it is recommended for using it for optimal result.
4. Settings is important: time and place is important. If you conduct the research in workday and worktime it’s quite good settings because your respondents is in the good shape and have fresh mind (and most of them taking on leave holiday) to give their input. But if you conduct the research in workday and after time than there is serious risk with your respondents because your respondents is already in tired situation due to the working all day and traffic jam situation. So it is best to pick on worktime or weekend time for conducting your research. Place is also important. Make sure your FGD’s place is comfortable so the respondents can focussed on the process instead of everything else. And one other thing, make sure you take their communication device (handphone, BB, and etc) to ensure focus on the process.
After your take care all the important factors your job as the FGD’s witness is make sure that the moderator don’t miss any important questions, you can check it with Discussion Guide (DG). It is also important that you dig the important findings that maybe not occurs in your discussion guide because that is your job (in this case our job 
) to make sure that this process result is meaningful findings not just not important findings.
December 13, 2010 at 12:48 pm
Good information, but I think you forgot some other tips that are crucial to the success of any FGD, except if is a deliberate attempt due to time and space constraints, which could be considered as general finding.
For example, the span of the discussion is crucial. A moderator should be able to keep to time frame; he/she cannot afford to keep respondents beyond certain time regardless of the skill and experience. Fatigue will set in, and then if vital questions that are critical to the research are not probed, then the quality of the result from such project will not be qualitative enough. Subsequently, decision made as a result of such outcome will mislead and divergent of customers interest.
December 15, 2010 at 12:23 pm
yes I agree with you Akin, but imho that is the mistake of 2 sides, first is the moderator for not able to control the flow (time and span) and witness who didn’t remind the moderator when she/he makes mistakes. That is why discussion guide is very important for the moderator as well for the witness because that is script for the show!
.
Thanks for your visit
December 22, 2010 at 6:13 pm
My God… Freddy, should you need learn anything about Qual MR, you know you should ask me… and WHY DIDN’T YOU CAME TO ME???
Let me get some points straight:
1.Moderator: A very good moderator, may not need his/her manager to guide her in every step. As for briefing the moderator – I believe it is the client’s responsibility to convey their objective clearly and not trying to put EVERYTHING in a single research. In my experience, when the client knows what they want, the research outcome would definitely gives clear answer – of course you need to have a good research question to begin with. The moderator is a researcher, he/she should have do their homework of learning the subject – some times, not knowing anything of the subject can give advantage too… therefore, he/she will not go into that fieldwork with pre-assumptions.
2. I’ve done 10 respondents in a group. The key of a good group dynamic is to control the energy flow – when a moderator can keep high energy (even up to 3 hours), I can assure you the group will always be good.
– of course another advantage of my speed of talking, they need to concentrate on me to understand me, hence purposelessly I force them to keep up 
There is no guarantee that we will always get good respondents – not all people can cope with a group of strangers.
I might suggest you to ask the moderator to call up to the silent respondents by calling them by the names. It always worked on my groups
3. Stimulus. Is only important to the objective. Why would you need a moodboard when you’re testing a STB?
It is excellent to get a live experience, but don’t overwhelmed the consumers with a million things to look at. And DON’T, let me repeat that DON’T use wordy stimulus for Indo consumers!!!! They read like grandmas!!
4. Setting
how can you invite busy ppl like managers to come at working hours? They will never come. It is customised to the type of consumers… Tired, worn out… we deal as much as we can.
As for the place – adjust them as necessary, I can’t call the director of government dept to my office, what shall I do? I go to them. There is always a perfect setting and time for different type of consumers, we just need to understand their nature.
5. be a nice client. No one wants to serve bitchy or ass-licking clients. When you are nice, believe you me, they will go the extra miles for you.
6. Understand the nature of Qual, we give you depth of a research, we can’t give you the scale.
7. I’ll be at IPSOS on Jan, give me work and I’ll show you good moderation
December 23, 2010 at 2:24 pm
hahaha I thought you’re quanti researcher!
. Waw! thanks for the advices. All the knowledge is registered! 