How to use the app retention calculator
The denominator is the install cohort, not the count still active the previous day. Dividing by the users who installed on the same day is what makes D1, D7 and D30 comparable on one scale. Enter the cohort size and the active users at each point, and the table fills in; days you leave blank do not appear.
The day 1 to day 7 and day 7 to day 30 holding rates below the table use a different denominator: the previous checkpoint rather than the cohort. They are the right view for spotting which stretch loses the most people. Users lost within 30 days is simply cohort size minus day 30 active users.
Retention levels differ so much by app category, acquisition source and measurement method that this calculator proposes no industry figure. Enter your own target D30 retention and it shows the gap in percentage points against the value displayed. Comparing cohorts measured the same way is the safest read.
Frequently asked questions
Classic retention counts only users active on exactly that day, while rolling retention counts anyone active on that day or later. Rolling values come out higher for the same cohort, so the two should never be mixed. This tool uses the counts you enter, so keep to one method.
Under classic retention, yes. Each day is counted independently, so a user who skipped day 1 and returned on day 7 lifts the day 7 figure. This calculator does not treat that as an error.