Caterpillar Cowboy

journo-geekery:

Retention, Cohorts, and Visualisations | The Intercom Blog

When I started designing cohorts for clients applications, something always bugged me about the visualisation. When you’re looking at a cohort you have a few different questions:

• How is our retention rate overall?
• Where do we lose customers?
• Is the rate at which we lose customers getting better or worse?

Whilst the above grid of squares can be useful, and augmenting them with sparklines helps, there are still comparisons that are difficult to make.

…

Cohorts are useful for visualising where customers are lost, if the results of a cohort analysis shock you, then you’re not in regular enough contact with your customers. Solve that problem first; there’s no use knowing exactly what’s happening if you don’t know why.

This, I like.  Definitely read further if mapping cohorts as cycle plots (the sparklines underlying the trend line) are new to you.  There’s a LOT of information packed into this one end-result graphic and the post breaks down the steps before it.  Good stuff.