Affinity Diagramming

Description

A bottom-up synthesis method in which individual observations are clustered by affinity into labelled groups, revealing themes from raw qualitative data.

Affinity diagramming organises large numbers of individual observations, quotes or interpretation notes into groups based on their natural affinity, building a hierarchy from the bottom up. It derives from Jiro Kawakita's KJ method, developed to make sense of anthropological field data, and was adapted in Contextual Design to synthesise contextual inquiry notes. Notes are grouped without predefined categories, and each group receives a label written as a first-person statement of the insight it contains, then groups are clustered into higher-level themes. Rigour comes from writing atomic notes that each carry one idea with a source code, resisting pre-set categories, writing insight labels rather than topic labels, and keeping a traceable link from every theme back to participants.

Attributes

Loss type
Active
Effort
Days

Steps

  1. Write each observation, quote or interpretation as an atomic note tagged with a participant code.
  2. Clear a large wall or digital board and place notes one at a time, grouping those that share an underlying meaning.
  3. Keep groups small (about 4 to 6 notes) and split larger ones.
  4. Label each group with an insight statement written in the user's voice, not a topic heading.
  5. Cluster groups into higher-level themes and label those.
  6. Walk the wall to check coverage and identify gaps or contradictions.
  7. Derive design implications and questions from each top-level theme.

Sources

  • Rapid Contextual Design
  • Lost Methods in UX
  • Lost Methods in UX — A Complete Treatise (Gopinath)
  • Beyer H & Holtzblatt K (1998) Contextual Design
  • Kawakita J (1967) Hassoho (Abduction)
  • Scupin R (1997) The KJ method: a technique for analyzing data derived from Japanese ethnology. Human Organization

Also known as

  • KJ method
  • Affinity mapping
  • Affinity clustering
  • Contextual Design affinity
  • KJ method (lineage)
  • Hierarchical affinity
  • Affinity Diagramming (Beyer-Holtzblatt Hierarchical Method)

Categories

  • User Research
  • Qualitative Analysis & Synthesis