Alternative names for this method: Concept Wall, K-J Method
Affinity diagramming is used to sort large amounts of data into logical groups. Existing items and/or new items identified by individuals are written on sticky notes which are sorted into categories as a workshop activity. Affinity diagramming can be used to:
Affinity diagramming is a simple and cost effective technique for soliciting ideas from a group and obtaining consensus on how information should be structured.
As part of research for a new design, or at an early stage in design improvement.
Review all of the material gathered to date yourself.
The items within their groupings are the output of this method. Sometimes you may find there is a small group who will create an alternative grouping, and sometimes participants will attempt to impose a structure on the groupings themselves. Always a good idea to take one or more photographs of the result as soon as the session is over.
Describe the skills or interests of the participants, and how much experience they have of this or similar applications. This is important for credibility.
The most important safeguard is not to impose your own pre-conceptions on the material, and to stress to the participants that what you want from them is how they would group the items. The best solution is usually not a purely logical one or a purely technical one (if it was, this exercise would be fairly redundant!)
If an interface is being designed, then the design team usually take this kind of output on board; however, technical issues and perhaps also policy issues may require negotiation.
Beyer, H. & Holtzblatt, K. (1998). Contextual Design: Defining Customer-Centered Systems. San Francisco: Morgan Kaufmann Publishers.
Affinity diagramming to analyse usability issues.
General method How to Cluster and Bundle Ideas and Facts
For pre-existing items, affinity diagramming is an alternative to card sorting. Card sorting finds common patterns in the way different people group information, while affinity diagramming obtains a consensus result.