What affinity diagramming is
You have finished a round of user interviews.
Your notes are full of quotes, observations and half-formed hunches, and now someone on the team wants to jump straight to a decision.
Before deciding what to build, consider how you will compare those notes and make room for different perspectives.
This article helps you decide whether a short affinity diagramming session is worth running before that decision, and what such a session actually involves.
Affinity diagramming means organizing related observations, ideas or findings into distinct clusters.
Physical or virtual sticky notes are commonly used, and the method is also called affinity mapping, collaborative sorting or snowballing.
According to nngroup.com, it is a flexible activity used across the UX process, from initial research to final design refinement.
It typically organizes observations or ideas from research, concepts from design-ideation sessions, and ideas about UX strategy.
For your situation, the relevant use is the first one: turning interview observations into named themes.
Instead of a wall of unsorted quotes, you end up with clusters such as onboarding confusion or pricing objections, each backed by the notes inside it.
When the session is worth running
The clearest signal is a pattern problem.
If your team currently decides what to build based on whoever argues hardest or on the freshest conversation, a clustering session changes the input.
Use the clusters to compare selected interview observations and identify questions to discuss before deciding.
For this proposed use, decide which notes to include and record any observations you leave for later review.
It is also worth running when several people collected notes separately.
If each interviewer holds a private mental model of what users said, the team is not actually looking at the same research.
For a shared session, ask each interviewer to contribute relevant notes to one visible board, then discuss how to group those notes.
It may be less useful when you have only one or two interviews and a single decision-maker.
With very little data, clustering can become ceremony around a conclusion you had already reached.
In that case, a written summary of what you heard may serve you better until you have more interviews to compare.