Product & Analytics / Growth
N-Day Retention ~ (variants: "bounded / Day-N", "unbounded", "rolling", "range/bracket")
Of the people who started using the product on a given day, how many came back later.
The fork, why two teams get different numbers
"Day-30 retention 22%" has at least three incompatible meanings, and tools ship different defaults for the same phrase.
In the Metric Library
- The full fork, both definitions worked all the way through
- The trap that makes the number lie, on a real export
- Every formula variant, spelled out
- The reconciliation anchor, what to tie it to and when to refuse
Reference: Product-analytics convention (no governing body); vendor retention-analysis documentation cited as convention · not reproduced