Cohort grain that fits Korean billing calendars

Many Korean consumer subscriptions renew on calendar months or on the same day of month as signup. That rhythm matters when you choose cohort grain for user retention charts.

Signup-week cohorts help when acquisition campaigns change weekly and you want to see early survival differences. First-paid-date cohorts help when free trials of uneven length scramble signup-week survival.

A workable compromise we often recommend:

1. Build primary retention views on first paid date.
2. Keep a secondary view by signup week for marketing retrospectives.
3. Exclude staff, refunded-within-48-hours, and gift accounts from both — and write that exclusion in the chart subtitle so leadership does not forget.

If your warehouse already stores “subscription_started_at” and “first_charged_at” as different stamps, document which one feeds the board metric. Ambiguity here creates more confusion than any sophisticated model later.