Label Topics in Focus-group Transcription
A time-consuming but very important step after conducting focus groups is to identify where in the transcription each topic was discussed and what was said. Relevance AI's "Label topics in focus-group transcription" workflow speeds up this process by looking for section identifiers.
How to use Label Topics in Focus-group Transcription
Requirement
- Conducting focus groups and using section identifiers. Section identifiers are simply unique sentences said by the moderators when starting a section. For instance,
- What do you think is the most concerning point about XYZ?
- How do you explain the experience you had ...
- Transcribing the audio files. Relevance AI provides you with several workflows to transcribe audio, one of them is Transcribe Audio and Identify Speakers. If your audio file is already transcribed, it is possible to upload the transcription as CSV as well. Read more about audio processing here.
Label Topics in Focus-group Transcription workflow wizard
Select your dataset that includes the transcription. Go to workflows and locate "Label Topics in Focus-group Transcription".

Relevance AI - Label topics in focus group transcription
On the workflow wizard:

Relevance AI - Label topics in focus group transcription setup wizard
- Select the text field containing the transcription
- Enter the section identifiers and the corresponding label you wish to assign to each section
Example:
Note 1: You can have multiple section identifier for a single section| Section identifier | Label | | --------------------------------------------------------- | ------- | | What are the necessary skills for this job? | Skills | | What do you think is the most concerning point about XYZ? | Concern |
Note 2: There is no limit on the number of entered section identifiers - If your dataset includes transcription of more than one audio file, you must specify the field showing the file names otherwise, results will not be be accurate.
- Click on "Run workflows"
Access the results
After the workflow is finizalized, a new field called "Topic/Section" is added to your dataset. You can set up a category view page to analyse each section topic separately. You can also export your data to a CSV file.
Tips for defining section identifiers
- Avoid sentences or phrases that could be said more than once during the interview/focus group
- Avoid sentences or phrases that could be easily said differently. For instance
That is
andThat's
are easily reversible - Avoid too long section identifiers. The longer they are the more the chances that they are said differently by the moderator
- You might need to use a subset of your section identifier when using this workflow
Updated 4 months ago