Affinity diagrams is a analysis technique that involves hundreds of post it notes, and using them to organize them out to find out new information about the data you collected.
The nature of this method is definitely data analysis, since you use it to analyze data. The context that you will use this method is if you have a lot of data you need analyzed, you could use this to get that done.
Description of the Method
"Affinity diagramming is a process used to externalize and meaningfully cluster observations
and insights from research, keeping design teams grounded in data as they design." - Martin and Hanington, Universal Methods of design
This method, much like Elito, first starts out with you listing out all the observations you had whilst researching your subjects. Though instead of straight up putting them into a spread sheet, you put them on post-its. You then organize these post its out, grouping them into groups and sub groups, until you have them all in three or four huge groups. This, like Elito, is an inductive approach, aka it goes bottom up. You go into it not knowing the issues, and look at the raw data and naturally place them in chunks, which eventually forms the issues.
Each Group up should have a name, and as your groups get grouped together, they should have higher level names.
- the lowest ones should be instances showing those issues.
- The ones above that are the aspects of that issues
- Above that, it is the specific issues of that problem space
- The highest level should be the problem space you have
After your done doing this, you should be able to see your major problem areas to the top, and from those, you can start building out ideas from them. You could also move this into a spreadsheet online after you are done, but this is not needed, this is mainly to look back at it after its done, so you don't have to set it all up again, and pictures can get confusing.
Actually Doing the method
To prepare for this method, get a lot of different post it notes, you are going to need them, and get all your research ready so you can look at them as you write the observations down.
The first thing you should do in this method is write down as many observations as you can think of based off your research. You and whoever you are working with should do this at the same time, so you can get as many down as possible. Each one goes on a different post it, and for now just put all of them in a pile. When you think you are done, stop. Try to get as many observations before you stop though, but if you miss some its fine, you can always add more as you go.
The next step is to completely randomly put up all the post its on a board, and spread them out. The point of this is so you can start seeing connections that you haven't thought of before. Doing it as you put them up on the board could start to become more rigid then you want, making you miss new connections.
After you have them up on th baord, start grouping them together one by one, from smaller to bigger as you go. This can take awhile to do, and it is easy to get stuck while doing this, with people argueing what goes where, but if you stick with it you can get it done. From here, you can put them up in a spreadsheet, or just take pictures and base it off those.
Positives and negatives
- Helps you analyze a lot of data and make sense of it.
- Forces you to look at each piece of data, piece by piece.
- Builds connections you would never see on your own
- Helps you find issues that were not readily apparent.
- Can take a long time to do
- Could get stuck
- Uses a ton of post it notes.
Tips and Tricks
- Try to make the observations not that complicated, make them as bare as they can.
- If you get stuck grouping them, have someone else come by not attached to the project, many times new eyes can help you see connections you did not.
- You can always put the same post it in two groups if it fits equally in both.
- Have different colored post it notes for group names, so that way you can easily tell the differences
Project we used it for
This went better then our Elito method in my opinion, though I still think we could have done better. We did this in conjunction with our Jigs research from back in the contextual inquiry, and we wanted to use it to find more data with it. The thing that I think we messed up the most on was the groupings, we did not stick with the issues grouping, we just kind of grouped them up as we went, connecting them any way possible.
When we got stuck, I ended up helping out another team with theirs, and it ended up getting new insights for them and helped them group it better, and then they came over to ours and helped us, it was a pretty nice relationship between groups.
I liked affinity diagram, especially when compared to the Elito method. Thought they were pretty useful overall, and I will definitely use them again if I need them. In fact we used a pseudo affinity diagram in the final project to help analyze the data we got.
Fun Fact, for way too long first semester, thought this technique was called infinity diagrams.