I am using “active subscriptions” segmented by the first purchase month to look at blended retention (all product duration). I find it very convenient. What I realized recently is that if you compare:
- yearly retention chart
- active subscription chart filtered yearly and segmented by first purchase month
You get a different result. In particular, the yearly retention chart initial value for a cohort is around 10% higher than the active subscription chart initial value. I am trying to understand why.