i assumed my most active new users would stick around longest so i checked 1,641 signups on post bridge: the people who cancelled had connected MORE accounts (3.6 vs 3.3) and posted MORE in week 1 (51% vs 45%) than the people still paying 6 months later (which i know is not statistically significant, but it did shatter my intuition as data often does) the only early behavior that predicted staying was something else entirely