Use the river diagram to communicate data

I run a lot of workshops, including planning workshops and retrospectives.  Sometimes the crew votes on things like “what is the best feature, or how did we go this time.”

But sometimes we collect data and then want to discuss it as a team.  And this presents a challenge – I like data in a spreadsheet and I like pictures on the wall but sometimes it seems hard to capture numbers in a useful way in the workshop.

But do not fear – the River Diagram is here and this is exactly what it is for.

So, how does the river diagram work?

Lets say I run a fortnightly retrospective (team meeting to look at improvement).  A while ago we decided that we wanted to focus more attention on “becoming more awesome,”  or as out tester put it “let’s try to be a bit less lame and embarrassing.”

While we did not really agree on whether we were pretty embarrassing or pretty awesome, we did agree on where we could focus our energy in order to improve.

We had been working on our improvement plan for for a few weeks and Delphi, our data analyst, said she would run a survey.

Delphi came up with some questions and sent them out.  The team then completed the survey and sent their results back.

Delphi wanted to go through the results in our retro.  She wanted to project the spreadsheet on the wall and then discuss the results.  But Fang, our UX guru, suggested we use the famous river diagram for this purpose (It is apparently famous but nobody had actually heard of it).

This is what Fang did:

Step one, write the questions on a white board or handy wall.  Write them one under each other in a column.

Step two, add the possible scores across the top to form a row of numbers.

Now we had this on the wall:

Then he drew the “river.”

Step three involves putting a mark against the lowest and highest score for each question.

Step 4 involves drawing each “river bank” by connecting the dots for both lowest and highest scores.

Now we had this on the wall:

Delphi said is was great data analytics for a UX artist type and then wanted to do some normal distribution or something.  But everyone else started to immediately express views on the data.

Where the “river” is fat, it means we disagreed – there was a high variance of opinion.  People started discussing the fat part of the river and why some people thought were great at breaking work down and others thought this was our area that most needed improvement.

Then we noticed that anything narrow meant we pretty much all agreed on the score.  But we also noticed that anywhere the river flowed to the left (sticking to the plan) seemed low and anywhere the river flowed to the right (learning new things and building something we are proud of) was high.

So without any real analysis, our eyes were drawn to patterns and our discussion followed the flow of the river.

In hindsight we could have probably all just written our scores on the wall instead of doing a survey.

But the approach works well either way.  I have used this for health checks, “engagement survey” results and retros. It always seems to lead to a worthwhile discussion.

This is not a precise analytical tool, but it is great for use in a workshop.  Give it a try and let me know how you go.