Blog

From highlight to insight: tagging and coding interview notes

July 10, 2026 · 5 min read

An interview transcript is raw material, not a finding. The value is in what you pull out of it — the moments that carry signal — and in making those moments findable and reusable later. Highlighting and tagging (what researchers call coding) is how you get there, without ever losing the quote behind a claim.

Highlighting: capture the moment it matters

Reading a transcript, you feel the sentences that matter. Highlighting is simply selecting one and turning it into an insight on the spot — while you're still in context. The highlighted quote becomes the insight's evidence, so the finding is anchored to what a real person actually said, not a paraphrase you'll second-guess next month.

Tagging (coding): give every highlight a place

A highlight only becomes reusable once it's classified. Four lightweight dimensions do the work:

• Type — is it an insight, a pain point, an opportunity, a feature request, a user need, or an observation?
• Themes — the recurring topics your research keeps circling back to.
• Tags — lighter, cross-cutting labels (behaviours, product areas) for slicing later.
• Personas — who the finding is about.

Do this at capture time, not in a big coding session weeks later. A few seconds of classification per highlight is what keeps the whole repository searchable.

Why the evidence link is the point

A finding with no source is just an opinion. When every insight keeps the quote and the interview behind it, anyone can trace a claim back to the moment it came from — in a different team, months later, when the decision it informed is being questioned. That traceability is the difference between research people trust and research people relitigate.

From highlights to patterns

One highlight is an anecdote. Tagged consistently, highlights roll up into something bigger: themes reveal what your research keeps surfacing, affinity mapping clusters loose notes into candidate groups, and a grounded AI summary can draft the narrative — all still traceable to the individual quotes underneath. Coding is what turns a stack of interviews into a body of understanding.

Keep your tags consistent

• Prefer a small, shared taxonomy over ad-hoc tags invented per study.
• Reuse an existing theme or tag before creating a new one — near-duplicates are the enemy of search.
• Write insight titles as claims, not raw quotes ('technicians distrust offline sync', not 'sync issues').
• Merge duplicates so the same finding doesn't exist five times.
• Give one person or a small group ownership of the taxonomy so it doesn't fragment.

How Lens helps

In Lens you highlight straight in an interview's notes or transcript — select any passage and capture it as an insight in one action. Set its type, themes, tags, and personas right there, and it stays linked to the quote and interview it came from. From there it feeds your themes, affinity boards, and AI summaries. For the wider picture, see insight management and what a research repository is.

← All posts