How to Transcribe a Focus Group: Recording, Speakers and Template
Focus groups are hard to transcribe: many voices, crosstalk and no speaker labels. Record well, keep a speaker log and label the transcript.
On this page
To transcribe a focus group, record it with participants' consent and a microphone that reaches everyone, keep a speaker log during the session, transcribe the recording, then use the log to label each speaker and mark crosstalk. A speaker log makes attribution easier to check when voices overlap or sound alike.
Focus groups are harder than one-to-one interviews: many voices, people talking over each other, and transcription tools that don't tell participants apart. Planning the recording and speaker log before the session reduces the identification work afterward.
Before the session: consent and identity
- Get recording consent in writing, covering recording, transcription, who will see the transcript and how quotes may be used. If you are working under an ethics approval or a client agreement, follow its terms.
- Choose stable participant codes such as P1 to P8, with the identity key stored separately under access controls. Codes make attribution easier to manage, but a distinctive employer, location or personal story can still identify someone.
- Decide where the audio may be processed. If your consent terms or client agreement restrict which services may handle the recording, that rules out some transcription options.
Record so the transcript has a chance
- Use one good central microphone in a small room, or a recording of the video call if the group is online. A phone at one end of a long table will miss the far end.
- Seat people so nobody is far from the microphone. Note the seating plan, with each participant's code.
- Ask people to say their first name or code the first time they speak. It helps both the speaker log and anyone checking the transcript later.
- Have the moderator repeat quiet comments and name speakers occasionally: "Thanks, P3. P6, what do you think?"
- Ask people not to talk over each other, knowing they will anyway. It reduces the problem.
Keep a speaker log during the session
A second person, the note-taker, writes down who speaks and roughly when. It doesn't need to be detailed:
| Time | Speaker | First few words |
|---|---|---|
| 00:04:10 | P2 | "I usually check the app first..." |
| 00:05:02 | P5 | "For me it's the price..." |
| 00:05:40 | P2 + P7 | crosstalk about delivery times |
Afterward, you match the log against the transcript to label speakers. Without it, you are guessing from voices on a recording.
Transcribe the recording
Open the recording transcription tool, select the spoken language and choose a processing mode:
- On this device downloads a speech model and keeps the audio in your browser. Your device's memory sets a practical limit for long sessions; try a short section first.
- Online uploads the recording to the speech provider named before upload and avoids the model download. Check the file limits and usage limits the tool shows before uploading, and check your consent terms allow it.
Neither mode labels speakers. You'll do that in the next step.
If your platform recorded each participant separately, label each file before transcription and align the results against the original meeting timeline. Check whether each file includes silence, a shared-room microphone or other voices. Separate files can make attribution easier, but they do not automatically produce a correctly ordered conversation.
Label speakers and mark crosstalk
Work through the transcript with the recording and the speaker log open:
- Start a new line at each change of speaker and add the code from your log: M for the moderator, P1 to P8 for participants.
- Mark overlapping speech as "[crosstalk]" and transcribe the main thread if you can follow it.
- Mark what you can't hear as "[inaudible 00:23:15]" rather than guessing.
- Mark non-verbal reactions that matter, such as "[several participants laugh]" or "[general agreement]", but don't invent agreement you didn't hear.
- Check every passage you plan to quote against the audio and the speaker log.
For a full correction method, see the guide to cleaning up a transcript.
A five-minute check before the group arrives
Record the moderator and a person in the farthest seat speaking in turn, then briefly together. Listen back through headphones. Confirm that both are intelligible, check the spoken language, and verify that the chosen processing method is allowed under the study’s consent terms. If the far voice disappears, move or add a microphone before the session starts.
Choose your transcript standard before editing. A clean-verbatim draft can remove some false starts for readability. Interaction analysis may need pauses, emphasis and interruptions retained. Document the notation you use so another researcher can understand it.
Copy this focus group transcript template
Focus group transcript
Study / client: [name]
Session: [number, date, in person / online]
Moderator: M = [name] Note-taker: [name]
Participants: P1 to P[n] (codes only; key stored separately)
Seating plan / recording setup: [plan or description]
Consent: [what participants agreed to]
Transcript standard: [verbatim / clean verbatim]
Transcribed by / checked by: [names and dates]
[00:00:00] M: [introduction and ground rules]
[00:04:10] P2: [response]
[00:05:40] P2, P7: [crosstalk] [main thread if audible]
Unclear passages: [timestamps]
From transcript to findings
A focus group transcript is raw material. To turn it into findings:
- Read it once without coding, to get a feel for the conversation.
- Code it by topic, marking passages that relate to each question or theme.
- Separate expressed agreement from inference. Record which participants actually endorsed an idea and note dissent. Silence, nodding you did not record, or a loud speaker do not establish group agreement. A small focus group is not a population survey.
- Pick quotes that represent a view, not just the most vivid line.
An AI summary can help you see themes quickly, but treat it as a first pass. It can blur who said what and overstate agreement. On Meetings Brief, the summary step is optional: the text you submit goes to an online AI service, and daily limits apply. Check that your consent terms allow it, and read the privacy policy first.
Frequently asked questions
How long does it take to transcribe a focus group?
Longer than an interview of the same length, because of the number of voices and the crosstalk. The machine draft is quick; most of the time goes on labeling speakers and checking passages.
Can transcription software identify each participant?
The Meetings Brief transcription tool doesn't label speakers. Even tools that do can confuse similar voices in a group. Use a speaker log and check labels by ear.
Should a focus group transcript be verbatim?
It depends on your analysis. If tone, hesitation and interruption matter, use verbatim. For most market research, clean verbatim is enough.
Can I transcribe an online focus group?
Yes. Record it with participants' consent, then transcribe the recording, or transcribe it live from a browser tab with live meeting notes. Separate per-participant tracks, if your platform provides them, make labeling much easier.
Sources and scope
Product workflow checked against the Meetings Brief source on 11 November 2026. The log and template are original working aids. Check research consent and data handling against your study’s approval and organisation’s rules.
The UK Data Service recommends consistent speaker tags and documented transcription conventions. Its anonymisation guidance explains why removing names alone may leave participants identifiable. Data-level documentation, Anonymisation for text data.