Transcribe focus groups
Upload a focus group recording and get a timestamped transcript with speakers separated, ready to check, code and quote.
How it works
- Upload the session recording, from a single central microphone or a mixed feed.
- Choose the language and start the transcription.
- Rename the speakers using your seating plan or moderator notes.
- Correct the text, then export DOCX or TXT for coding.
A focus group is a recording problem before it is a transcription problem. Six or eight people around a table, one microphone, laughter and the occasional interruption: the audio is as hard as speech gets. A good transcript starts with good placement of the microphone, and a good analysis starts with accepting that the machine's first pass will need a human to finish it.
Scribewren separates the voices and timestamps each turn, so you spend your time on the parts that need judgement. Rename the speakers as you learn who is who, tag the passages that answer each question in your guide, and listen again wherever the labels look doubtful. Where two people speak at once, the transcript will rarely capture both, which is why the playback link on every line matters.
From transcript to themes
Once the text is checked, export it to your analysis tool or a document and begin coding. Because the timestamps are preserved, every quote can be traced back to its moment in the session, which makes reporting quotations to a client or committee far easier to defend.
Tips
- Seat participants in a circle with a single microphone in the middle, and ask people to speak one at a time.
- Note the order in which each person first speaks; it makes renaming the labels much quicker.
- Expect more corrections where several people laugh or talk over each other, and listen to those passages again.
- Keep the moderator's guide next to the transcript so you can mark the question each answer responds to.
Questions
How many speakers can it separate?
Several voices are separated automatically. Accuracy of the labels falls when voices are similar, far from the microphone or overlap, so check the busier passages.
What about participants talking over each other?
Overlapping speech is the hardest case for any transcript. Expect to listen back and correct those passages by hand.
Can I export timestamps for analysis?
Yes. Timestamps are in the editor and travel with SRT and VTT exports.
Is participant privacy protected?
Originals are deleted within 24 hours of processing, nothing is used for training, and transcript text is not sent to analytics or error reports.
How long can a session be?
Unlimited accepts files up to 10 hours. On Free, only the first 30 minutes are transcribed.
Try it on your own recording
Free includes a few transcriptions a day. No card needed. Compare plans.
Start free