Affinity Diagrams are used to organize unstructured ideas and data into related groups for further analysis or action. It is one of the Seven Management and Planning Tools[1] that came from Post-World War II operations research conducted in Japan. It is also frequently known as the KJ Method after its inventor, Jiro Kawakita.
Affinity Diagrams are commonly used by Business Analysts, Project Managers, researchers, business management, and others who need to sort a large amount of information into groups.
Note that there are two important points I want to make about Affinity Diagrams up front.
Affinity Diagrams are used not just to organize ideas and data, they are also used to build consensus among a group for:
Affinity Diagrams are most commonly known though as the tool used during Brainstorming [2] sessions to capture and organize the team output. However, Business Analysts may find them useful in other situations as well, such as when analyzing:
And similar types of large data sets where the information provided can cover a wide area and where organizing the information into similar data sets can help further analysis.
When completed, the Affinity Diagram results can be used to create Cause and Effect Diagrams (aka Ishikawa or fishbone diagrams); guide stakeholder analysis efforts; or can feed into the creation of a feasibility study or business analysis plan.
The process of creating an Affinity Diagram is relatively simple. However, many descriptions of the technique that exist conflate the early part of a creating an Affinity Diagram with the Brainstorming process. The best description of this combined process comes from Jared Spool[3]. However, in order to clearly separate the two I am going to assume the process of creating the Affinity Diagram begins once the unstructured data set exists. For example, once the Brainstorming session has finished.
Creating an Affinity Diagram is a three step process that includes:
The first step is to sort the data set into groups based on common themes or other relationships. This process should continue until every item in the data set has been placed into some group, even that is a group
of one item.
If the Affinity Diagram is being created as part of a group exercise, then use the following rules
:
The location of the items for this process is irrelevant, as long as all participants can easily access and see all of the items. This is most commonly done by placing all items on index cards or Post-It Notes and moving them around on walls or tables. But it can also be done via some software that allows collaborative interaction or writing (Google Docs for example allows collaborative editing of a document).
The second step is to give the groups that have just been created names that describe the common theme of the group, or name a category that the items fall into. This is done by re-reading all of the items that comprise a group and looking for common themes and relationships. For example, if reviewing the feedback from User Acceptance Testing the group names might be Interface changes
, Performance problems
, Functional errors
, Missing Functionality
, and similar groupings.
If the Affinity Diagram is being created as part of a group process, all participants should remain silent during the naming process and each participant should have the ability to name each group on their own. This might be done by giving each participant index cards of their own that they can write each group name down on.
The last step of creating the Affinity Diagram is to analyze (and discuss if in a group) the groups and re-organize them into broader mega-groups
or break some groups into smaller sub-groups
. Once organized in this way the diagram can be adjusted so that the groups are prioritized in relation to the larger goal. If the diagram is being created in a group setting, you might ask each participant to write a number from 1 to 3 on the label of the three groups they think are the highest priorities.
The idea here is to gain a deeper understanding of the relationships among the data and to improve the Affinity Diagram so that it reflects the goals and priorities of the analysis target or the group.
The result should be an organized and prioritized data set that can be used to guide further analysis efforts. For example, the data might:
The key thing to remember is that Affinity Diagrams are usually just part of the first stage of an analysis or group collaboration effort.
The results of the Affinity Diagram process are completely dependent on the individual(s) involved in creating them. While they may help bring common themes to light, the do not improve the analytical capability of their creators and important relationships can still be missed.
Parking Lotstrand of the mind map, and then creating categories as new strands and moving the ideas from the Parking Lot to the appropriate category strand. This has the benefit of showing both the sorted and unsorted ideas. Or you can
closestrands of the map selectively if you want to try and sort to only one category at a time.