Clustering

Clustering is a quick, easy way to organize data. In the process, you will intuitively, collaboratively spotlight relationships amongst your ideas.
This is usually a team sport. You will need a stack of 30 to 50 ideas produced using the Ideation Card, some markers and some more sticky notes. Give yourselves 60 minutes.

  1. Grab a handful of 10 to 15 ideas and read each one aloud. Post them up, discuss any that you don’t understand, and eliminate duplicates.
  2. Begin identifying pairs of similar or related ideas. Stick these paired ideas side-by-side on the wall. Find as many pairs as you can. Once you’ve got a few pairs started, you can also add ideas to existing pairs. Now you’ve got “clusters”.
  3. Grab another handful of ideas, read aloud, eliminate duplicates, and begin to post them up. Repeat this process of introducing new ideas into the mix and then adding to or making new clusters, until all ideas have been clustered.
  4. Feel free to reconsider and shuffle the clusters while you work. As a general rule, don’t let one cluster get too “full”—when a cluster contains more than 8-10 distinct ideas, it’s time to start asking if there is enough variation to break it into a couple smaller clusters.
  5. When finished, give each cluster a descriptive 3-5 word title.
ClusteringImage.jpeg

Clustering is an intuitive, collaborative activity. Group members add value as much through how they sort and discuss the ideas as with the titles they assign. Relationships and meaning in the ideas, pairs and clusters emerge naturally through the process.

You will sometimes find that there is more than one way to assign ideas to clusters. You might ask “where does this idea add more?” to choose a cluster to place it in. Or, you might find that there are two different interpretations of the idea at hand, which is why it fits into multiple clusters. Give each interpretation its own sticky note and add those to clusters. The process has revealed that, in the group’s best judgement, the difference matters. You’ll highlight it when you name the clusters.