Cohorts are such a great way of representing retention. We used a similar methodology to determine how many customers churned away from the business I worked for. The next step is always determining the root cause of cohorts leaving.
I feel like the most valuable of these charts is the cohort vs reward, but the grouping gets a bit busy when we look at quarters compared to years for the earlier users. Makes it difficult to do direct comparisons.
Excellent analysis. :)
I split the more recent years into quarters to be able to see things in a more granular fashion for newer users. I don't think there's as much value in doing the same for pre-fork users, as I still have a fairly manual process and not much new is happening with the older users being hardcore users continuing to stick around long term.
If you want I can give you access to the Google Sheet I used to collate the data and create the graphs. It would be fairly straightforward to join the quarterly cohorts together and make a cleaner graph. You can message me on stingchat if you'd like access.
Thank you for the offer, but at this stage, I do not have a lot of time. I have a lot of things to clean up before I start my new job the 2nd week of November - way too much stuff! :D