Turn user interview recordings into findings a team can act on

Before the interviews: plan for the analysis

Synthesis is far easier when the interviews were designed for it. Nielsen Norman Group advises treating interviews as a research study, not a casual chat. Write your research goals first. Then draft a guide with a few open questions and follow-up probes, and use it with every participant so answers compare.

Decide where recordings and transcripts live, who can see them and how participants are identified. Codes like P1 to P8 keep the analysis readable without spreading names around. Ask for consent to record, and say how long you'll keep recordings.

During each interview: listen, then probe

Let the participant talk, and use follow-ups to get at motivations and specific past examples. Avoid leading questions. NN/g notes that interview data depends heavily on the interviewer's skill, and leading questions undermine it.

If you're recording with a transcript, skip verbatim notes. Jot down moments worth revisiting with a rough time, so you can find them later.

After each interview: code the transcript

Within a day, read the transcript and tag passages with codes. NN/g describes a code as a short word or phrase that labels a passage, like a keyword for an article.

Some codes describe (“exports data to a spreadsheet”). Others interpret (“doesn't trust the dashboard figures”). Keep a code list with a one-line definition each, so a second person codes the same way.

Pull short quotes that show each code in the participant's own words. Note the participant code and timestamp next to each one. A finding without traceable quotes becomes an opinion the moment someone disagrees.

After the last interview: cluster, then prioritize

Put the coded observations on a board and group them as a team. NN/g's affinity-diagramming method has three steps. Write notes individually, then cluster and label the themes. Finally, prioritize them and agree next steps with an owner.

Don't drop a small cluster just because it's small, and don't force notes into a theme they don't fit. Write each theme as a finding: what you saw, how many participants, why it matters and what to do next.

Using Bridge to record and find interviews

Treat the Assistant's answers as leads to check in the transcripts yourself, not as the coding. Deciding what a pattern means is still the job of the researcher and the team.

• Capture interviews on Zoom, Google Meet or Microsoft Teams with Bridge Agent (the host admits it), record bot-free with the Mac desktop app, or host them on Bridge.
• Meetings you host or capture in Bridge get a searchable, speaker-labeled transcript.
• Export transcripts as PDF, Word or Markdown to code them in another tool.
• Tag every session in a study with a label, and keep the study's meetings in a space.
• Ask the Assistant inside a space where a topic came up across all its meetings.

What interviews can't tell you

Interviews tell you what people say they do. NN/g cautions that some people present themselves in a certain light, and you need observation or behavioral data to see what they actually do.

Pair key findings with analytics or observation before treating them as settled. Report counts as numbers of participants, not percentages of your users.