How I Synthesise a Week of User Interviews in One Afternoon

For a long time, synthesis was the part of research I dreaded. I would finish eight interviews on a Friday and spend the next week drowning in sticky notes, transcripts and half-written themes. By the time I presented, the team had already moved on.

Now I do most of it in a single, focused afternoon. Not because I cut corners, but because I set up the work so the afternoon is mostly assembly.

Before the afternoon: capture as you go

The secret is that synthesis starts during interviews. After each session, I spend fifteen minutes writing a short debrief while it is fresh:

• Three things that surprised me
• Direct quotes worth keeping, with timestamps
• Any behaviour that contradicted what the person said
• One open question I want to check with later participants

By the end of the week, I have eight short debriefs instead of eight hours of unprocessed recordings.

The afternoon, step by step

Here is how the four hours usually break down.

Hour one: pull observations

I go through each debrief and copy every distinct observation onto its own card, digital or paper. One idea per card. A typical week produces 80 to 120 cards. I tag each with the participant number so I can trace it back.

Hour two: cluster without naming

I group cards that feel related, without naming the groups yet. Naming too early forces cards into categories that may not fit. I just move things around until clusters stabilise. Usually I end up with 8 to 12 clusters.

Hour three: name and sharpen

Now I write a one-sentence insight for each cluster, phrased as a finding, not a topic. Not pricing, but people compare our price to a spreadsheet, not to competitors. Then I check each insight against the cards: does it hold for most participants, or just one vocal person?

Hour four: prioritise and draft

I rank insights by two questions: how many participants did this affect, and how much does it matter to the decisions the team faces now? The top five go into a short draft.

What the output looks like

My final document is rarely longer than two pages:

1. The research question, in one line.
2. Who we spoke to.
3. Top five findings, each with one supporting quote.
4. Recommended next steps.
5. Open questions.

If your findings need a 40-slide deck, you probably have not finished synthesising yet.

Why timeboxing helps

Giving synthesis a hard boundary forces decisions. Without it, I would keep moving cards around indefinitely, chasing a perfect structure. With it, I deliver findings while the team still remembers why they asked the question, which is when research has the most influence.

Cover: Photo by Vitaly Gariev on Unsplash.