A cohort-based survival model groups customers, users or token holders by the period in which they started, such as the month of sign-up or of a token claim, and tracks what share of each group is still active at each later period. The survival rate for a cohort at month is the number still active divided by the number who started, so a cohort of 1,000 with 420 active after six months shows 42% survival.
It is a refinement of cohort analysis and gives a truer picture than a single average churn rate, because early cohorts, users acquired through incentives and later organic users often behave very differently. Subscription businesses use it to estimate customer lifetime value. In token projects, analysts apply it to airdrop recipients or active addresses to see whether activity holds after rewards end.
Ref 1010001416