How does entropy effect your cohort analysis? If over time the shape of a user’s experience becomes less about their onboarding, and more about their general use, should after a while all cohorts become a general cohort, and it be less important when they on-boarded or when they signed up?
Does that mean maybe after 1 year of using our learning chat product any changes since that student on-boarded become less different and they become more “generally used to it”.
Or should we still divide our metrics by which version they on-boarded to?
[entropy in my loose case is, “a decline from the shape of original perceptions.”]