Activity⚡ Medium

Affinity Mapping

Turn a wall of research observations into patterns your team can act on. Affinity mapping writes individual observations on separate notes and groups them by theme — finding the signal in qualitative data.

Research SynthesisDiscovery WorkshopsTeam AlignmentFeature PrioritisationProblem FramingSprint Planning

01 — TL;DR

Two sentences.

Affinity mapping is a research synthesis method where individual observations — each on a separate note — are grouped by theme through a collaborative sorting process, producing a visual map that reveals patterns, frequencies, and relationships in qualitative data invisible in raw transcripts. The output is not a summary of what users said; it is a structured representation of what kept coming up.

Developed by Jiro Kawakita in the 1960s (the KJ Method), it was adopted into design practice for synthesising large qualitative datasets. Its value for product teams is the collaborative dimension: when designers, PMs, and engineers sort observations together, every person has processed the raw data and arrived at themes collectively, rather than receiving a researcher's pre-digested summary.

02 — When to Use

Apply this when…

→You have completed five or more user interviews and need to identify themes before presenting findings

→Usability testing produced a large volume of behavioural observations needing synthesis

→A cross-functional team needs to process shared data together rather than receiving a summary

→Support tickets, surveys, or feedback has accumulated and needs structured analysis

→A discovery sprint produced observations from multiple sources and you need signal across all of them

When NOT to apply it

Skip it when you have fewer than ten observations (patterns too speculative), when findings are already synthesised and the team needs to act, when the question is quantitative, or when time constraints demand a faster method like a verbal summary.

03 — How It Works

The mechanism

Affinity mapping works by externalising data and making it physically manipulable. When observations are on separate notes that can be moved and regrouped, the synthesis becomes a visible, spatial, collective act. The sorting reveals relationships that reading alone would not surface.

Pattern recognition requires externalisation

A researcher trying to identify patterns across forty observations while holding them in memory will miss connections obvious when notes are arranged on a surface. The affinity map is not a representation of thinking that has already happened — it is the environment in which the thinking happens.

One observation per note, not summaries

Each note should contain a single, specific observation — something a user said, did, or expressed. "Users were confused by the pricing page" is an interpretation. "Sarah could not tell if the Pro plan included API access" is an observation. Interpretations predetermine groupings; observations allow groupings to emerge from the data.

Silence during sorting produces better maps

Silent sorting prevents the most vocal person from anchoring everyone's thinking. Each participant follows their own interpretation, and disagreements surface visually as notes that keep being moved between groups. Those contested notes are often the most interesting findings — genuinely ambiguous experiences that do not fit the expected narratives.

Theme stability and frequency are the primary signals

The map is complete when groupings stabilise. The two key outputs: theme frequency (how many observations per cluster — a proxy for how common the experience is) and theme severity (how emotionally charged the observations are — a proxy for how much it matters). Both dimensions inform prioritisation.

Synthesis tool, not decision-making tool

An affinity map showing "navigation confusion" as the most frequent theme does not mean the navigation should be redesigned next sprint. It means navigation confusion is the most represented issue in this dataset. Whether to act on it depends on business priority, effort, and corroboration. The map answers "what did we find?" not "what should we do?"

04 — Real Example

Spotify's cross-squad synthesis reveals systemic patterns

Spotify's autonomous squads regularly use affinity mapping to synthesise cross-squad research. When the discovery team and the library team both conduct research independently, affinity mapping reveals whether frustrations are local (one feature area) or systemic (appearing across multiple touchpoints).

Without shared synthesis, two squads produce separate findings that rarely get compared — each prioritises its own data and cross-cutting themes get missed. A user who cannot find a saved podcast and one who cannot find an old playlist are experiencing different symptoms of the same information architecture problem — and that connection is only visible when observations are sorted together.

Spotify · Cross-Squad Research Synthesis

Cross-stream synthesis reveals systemic patterns invisible in isolated research

Cross-squad synthesis → systemic insight

05 — In the Wild

Test yourself & see real examples

No examples yet — be the first.

Spotted a team that clearly synthesised research into patterns before building — or one whose product reveals observations were never connected? Submit what you observed.

✓ Reviewed before publishing✓ Your name on every example you submit✓ Violation or fix — both welcome

Seen Affinity Mapping skipped in a real product? Help grow the evidence base.

06 — Common Mistakes

Where teams go wrong

✗Writing interpretations instead of observations. "Users find navigation confusing" is pre-synthesised. "Marcus spent 40 seconds scanning the nav before clicking Settings instead of Reports" is a raw observation. Interpretations predetermine groupings; observations let patterns emerge.

✗Running the session solo rather than collaboratively. A researcher who synthesises alone and presents themes has produced a summary, not a shared synthesis. The team needs to handle the raw observations to develop ownership of the findings.

✗Stopping at theme identification without prioritisation. Ten themes presented equally is not a completed synthesis. Add frequency ranking (observation count per theme) and severity assessment (emotional intensity of observations) to extract actionable signal.

✗Treating every observation as equally valid. Mixing interview observations from recruited participants with a single power user's email and a stakeholder's intuition loses analytical integrity. Label the source on every note so groupings can be evaluated with appropriate weight.

07 — Variations & Related Principles

Connected ideas

Affinity mapping is a synthesis method — it sits downstream of research collection and upstream of design decisions.

The most important pairing is affinity mapping with journey mapping. Affinity mapping reveals what users experience; journey mapping reveals when and where. Together they produce problem statements specific enough to be defensible and measurable.

08 — 10-Min Exercise

Run it right now

⏱ 10 minutes · Team · No prep

The Quick Cluster

1. Think of the last five user interactions your team has had. For each, write one specific thing a user said, did, or expressed. Aim for 10-15 observations total. Do not summarise.

2. Silently, spend two minutes grouping observations by whatever feels similar. If alone, do this mentally and write group names as you go.

3. Label each group with the simplest description of what the observations share — one to four words. Avoid solution labels ("fix the navigation") or product-area labels ("checkout module"). Label the user experience.

4. Count observations per group and rank by count. The largest group represents the most frequent pattern. Ask: does this match what the team is currently prioritising? If not, you have found a gap worth discussing.

10 minutes