JS
REPLAY / ex-013
Personalized weekly digest
Reviewed as if it had not run yet. The coach sees only experiments that read out before 2024-07-08.
01 THE EXPERIMENT, OUTCOME HIDDEN
HYPOTHESIS
If lapsed users get a digest built from their own numbers rather than a generic newsletter, more of them come back.
PRIMARY METRICReactivation rate11.2pp baseline
EXPECTED LIFT+3ppas written at the time
AUDIENCEreturning userspersonalization
READ DATE2024-07-08started 2024-06-10
02 COACH REVIEW
Review this blind?
The coach will see only what this team knew before 2024-07-08. You answer, then the real outcome is revealed and the call is scored.