A hyper-casual developer using 90-second interstitial intervals saw 4.5-minute sessions and 12% day-7 retention. After moving to 3-minute intervals and triggering ads at level completion, sessions reached 6.8 minutes and retention recovered to 19%. Although each session carried fewer ads, more daily sessions per user lifted total daily impressions.
The tradeoff between ad density and session length is measurable. The task is finding the inflection point for your genre, core loop, and users.
Why Genre Changes the Inflection Point
No single interstitial frequency fits every category because session behavior varies by app type. Puzzle games have level breaks, arcade games have death-and-retry breaks, and casual idle games can run 15 to 20 minutes without a clear trigger. Each loop needs its own density calibration.
For puzzle and level-based games, an interstitial after every 3 to 5 level completions, rather than each level or every 60 seconds, generally preserves session length without materially sacrificing revenue. Completion is a mentally neutral transition: the player is between tasks, not interrupted mid-task.
In arcade and action games, a failed run creates a natural ad break because flow has already stopped. Volume is the constraint. With 45-second runs and 8 attempts in a 6-minute session, an ad after every run produces 5+ forced ads in 6 minutes. The observed action-genre threshold is every third to fourth loss event, or at least 90 seconds between interstitials, whichever is longer.
Interstitials are usually a poor fit for utility apps. Users open them to complete a task, so a mid-workflow interruption drives higher exits than a comparable interruption in entertainment. Use a banner, or a rewarded unit tied to a workflow step, such as "watch to unlock the exported file for free." If used at all, place interstitials at session exit or clear workflow completion.
The 90-Second Rule and Where It Breaks Down
A 90-second minimum is a sensible starting floor for most genres, but it is not sufficient in every case. Ninety seconds inside highly engaging gameplay can still break flow, especially in strategy and mid-core games with high cognitive load. There, a 3-minute minimum is safer even when the 90-second floor passes the aggregate session-length test.
The rule also ignores session depth. Someone 45 minutes into a session has invested substantial time and is less likely to leave after an interstitial than someone 3 minutes in who is still judging the app. Use tiers: no interstitials in the first 3 minutes, standard frequency in the middle, and slightly relaxed rules for demonstrably committed deep-session users.
A single frequency value remains useful when data is absent. It should be treated as a default, then replaced quickly with a rule calibrated to your app's session distribution once sufficient data exists.
Measuring the Right Signal
The usual test compares session length across cohorts with different frequency settings. But session length is a lagging metric. When an ad has already caused an exit, the damage is done. Track mid-session abandonment instead: sessions ending abruptly under 30 seconds after an ad, versus sessions ending naturally through content completion or deliberate exit.
A spike in mid-session abandonment after an ad event indicates density has crossed the friction threshold for that timing. Stable abandonment leaves room to add density and capture more impressions per session, or to keep density unchanged and optimize elsewhere.
Also track impressions per user per day, not only impressions per session. Aggressive density can shorten sessions and reduce daily sessions per user, lowering total daily impressions. Fewer ads per session can therefore support longer sessions, more daily sessions, more total impressions, and a stronger signal for demand partners.
Banner Density Is a Different Category of Problem
These findings concern interstitials and, to a lesser extent, rewarded video. Banners behave differently. Persistent banners usually have little effect on session length because users develop banner blindness after a few sessions. For banners, the key measures are screen real estate and perceived app quality.
A banner taking the bottom 10% of a game screen becomes effectively invisible to engaged users after the first few sessions. In a utility app with limited usable space, that same placement can make the interface feel cramped and lower task completion. Session duration may not reveal the issue because users finish tasks and exit. Track task completion and app store rating trends instead.
Keep banner and interstitial analytics separate. Adrop's format-level reporting breaks out impressions and engagement by ad unit and format, preventing interstitial session effects from being averaged with a banner's passive signal. Separate signals allow independent decisions instead of one broad diagnosis that ads are harming the app.