Mixpanel’s 2026 State of Digital Analytics report landed with a number worth sitting with: average week-1 retention for mobile apps in North America is now 4.8%, down roughly 30% year over year. Put plainly, the large majority of people who install an app today are gone within seven days, and that share is growing fast. A benchmark like this matters less as a target to hit and more as a warning that whatever “normal” your dashboards were calibrated against last year is no longer normal — a team quoting a flat or slightly-declining week-1 number against last year’s internal baseline may actually be outperforming the market while feeling like they’re losing.
The danger is treating retention as one number instead of a curve shaped by acquisition quality, onboarding friction, and channel mix. A benchmark this steep almost always reflects a mix shift — more low-intent installs from ad networks optimising for cheap installs rather than engaged users — rather than every app’s product suddenly getting worse. Teams that don’t separate retention by acquisition source will misdiagnose a sourcing problem as a product problem, and fix the wrong thing.
Data Points to Track
- Day 1, day 7, and day 30 retention, segmented by acquisition channel and campaign, not blended into one company-wide number, since channel mix shifts are the most likely driver of a sudden benchmark-matching decline
- Time-to-first-key-action, the gap between install and the first action that correlates with long-term retention in your product, since a rising gap here usually predicts a retention drop before it shows up in the day-7 number
- Cohort-level install quality signals — device tier, install source, and whether the install came through an incentivised or organic channel — logged at install time so later retention analysis can filter them out
- Week-1 churn point, the specific day within the first week where the steepest drop-off happens, rather than a single aggregate week-1 percentage
- Re-engagement triggers fired versus re-engagement triggers that produced a return session, to separate “we tried to bring them back” from “it worked”
Setup Steps
- Pull your own day-1/7/30 retention curve segmented by acquisition source before reacting to the industry number — a company-wide average can mask a healthy organic cohort dragged down by a paid channel bought for volume.
- Instrument time-to-first-key-action as a leading indicator, and set an alert when it drifts upward, since it moves before retention does and gives the team a head start.
- Tag every install with its source and campaign at the event level, not just in the attribution dashboard, so retention cohorts can be rebuilt and re-sliced later without re-running attribution.
- Compare this year’s cohorts against last year’s on the same channel mix, not the raw benchmark number, to judge whether your product or your acquisition strategy is the bigger driver of any change.
- Set a recurring monthly review of the week-1 churn point specifically, since a benchmark that moves this fast deserves a shorter check-in cycle than a quarterly retention review.
Actionable Insights
If day-1 retention holds steady but day-7 drops sharply, the problem is almost always onboarding depth rather than acquisition — users are arriving fine but not reaching the action that keeps them. If retention drops across every channel roughly in proportion, that points toward a genuine product or seasonal effect rather than a sourcing issue, and is the one case where the industry-wide benchmark decline is directly relevant to your own diagnosis. And a widening gap between paid-channel and organic-channel retention curves is the clearest signal that acquisition spend needs re-targeting toward quality over install volume, well before the blended company number reflects it.
Related Resources
Need help tracking this in your app?
Our team sets up analytics pipelines for mobile and web teams every day. Talk to us and get your first events flowing in under an hour.
Talk to an expert