Marketing owns the listing. Engineering owns the install experience. Both move the same number — install conversion.
App Store Optimization is usually framed as a marketing function — keywords, screenshots, ratings, A/B tests. The marketing half matters. But there is an engineering half that moves the same numbers and is rarely owned.
Install size is a conversion lever
Apps over 200 MB on the App Store warn users on cellular networks. Apps over 500 MB are increasingly returned without install. Every engineer should know the size of their app's binary and the trend over time. Asset catalogs, on-demand resources, and App Thinning recover surprising amounts.
Crash-free session rate predicts retention
If your crash-free rate is below 99.5%, your store rating ceiling is a half-star lower than it could be. Firebase Crashlytics or Sentry Mobile, integrated from sprint one, are the price of entry.
Cold start is a first impression
Cold start over 2 seconds is the difference between "app feels snappy" and "app feels slow" — and users decide in the first session. Profile cold start with Instruments (iOS) and Macrobenchmark (Android) every release. Defer initialization that does not block first frame.
Background battery and data drain
iOS and Android both surface "this app uses a lot of battery" warnings. Hitting one of those is a one-star review generator. Audit background tasks, location updates, and network polling intervals.
Permission prompts: ask less, ask later
Every permission prompt on first launch costs install completion. Defer push permission until the user does something that warrants it. Defer location until they tap "find nearby." Conversion goes up measurably.
Update cadence as a ranking signal
Both stores reward apps that update frequently. A monthly cadence is the floor; weekly internal builds via TestFlight and Play Beta keep the team honest. Stale apps lose to fresh competitors.