What is Affinity Mapping?

Affinity mapping (affinity diagramming) is a bottom-up technique for organising qualitative data: individual observations are externalised onto separate notes (stickies, cards, digital equivalents), then grouped by affinity, their natural relatedness, with the groups labelled only after they form, and often grouped again into higher-level themes. The core discipline is direction: clusters emerge from the data rather than data being filed into predetermined buckets, which is what separates affinity mapping from sorting into a taxonomy someone brought to the meeting. Its lineage runs to anthropologist Jiro Kawakita, whose KJ Method, developed in 1960s Japan for synthesising ethnographic field data, formalised exactly this externalise-cluster-label sequence; the sticky-note workshop that design teams run today is the KJ method wearing Post-its.

Running It Well

1. Prepare atomic, sourced notes.
One observation per note (a behaviour, a quote, a pain point), specific and traceable to its participant and session. Interpretations and solutions stay off the wall at this stage; the map is built from evidence, and a note that can't be traced (via a participant tag back to the transcript or recording) is an orphaned claim when someone later asks "who said that?".

2. Cluster silently first, then discuss.
The strongest workshops open with silent grouping: everyone moves notes without argument, so the loudest voice doesn't pre-frame the clusters and independent readings surface, the same anti-anchoring logic that runs through bias-resistant analysis. Debate comes second, and disagreement about where a note belongs is signal: it usually means either the note bundles two observations or the categories aren't clean yet.

3. Label clusters as findings, not topics.
"Onboarding" is a filing label; "people abandon setup when asked for permissions before seeing value" is a finding. Insisting that cluster headers make a claim converts the map from storage into analysis, and the headers become the readout's first draft.

4. Weigh clusters by evidence, not note count.
Twelve notes might be one vocal participant; the honest weight is distinct participants per cluster, checked before any cluster is crowned a top theme, the small-scale version of the counting discipline that keeps qualitative claims honest.

Where It Fits, and Where It Strains

Affinity mapping shines at collaborative sense-making: synthesising a round of interviews or usability sessions as a team, aligning cross-functional partners by having them touch the evidence, and generating the theme structure that deeper work will formalise. Its strains are equally real: it is reliability-light (different groups produce different maps from the same notes), vulnerable to workshop dynamics, and shallow on its own for contested or high-stakes claims. The standard reinforcement is sequencing it with thematic analysis: the map for speed, collaboration, and hypothesis-generation; systematic coding across the full dataset (with its codebooks and second coders) where findings must survive scrutiny. Digital walls have also changed the inputs: when studies collect transcribed video and open-text answers in one place, as a Ballpark study does, the notes arrive pre-sourced and the path from cluster back to the moment of evidence is one click, which is precisely the traceability the method historically lacked.

The Takeaway

Affinity mapping is emergence as a method: atomic evidence externalised, clustered bottom-up, labelled as claims, weighed by participants rather than notes. Use it to turn session mountains into shared themes at speed, keep every note traceable, and graduate the load-bearing findings into systematic analysis. Kawakita built it to let field data speak before the theorist did; run properly, that is still exactly its job.