Store optimisation used to be a keyword exercise. It is now closer to conversion rate optimisation on a page you do not control, judged by an algorithm that increasingly cares about what happens after the install.

Both major stores have shifted weight toward post-install engagement signals. Retention, session depth and the ratio of installs to meaningful first sessions now influence ranking as much as keyword relevance. In practice this means a listing that over-promises will rank worse than a listing that under-promises and delivers, even if the first one converts better in isolation.

Custom listings are underused

The ability to serve different screenshots, descriptions and preview videos to different traffic sources is the highest-leverage tool most teams are not using properly. A visitor arriving from a paid campaign about one feature should see that feature first. A visitor arriving from a search for a competitor should see the comparison that matters to them.

Teams that treat these variants as real experiments — one hypothesis, one changed element, enough traffic to conclude something — routinely find double-digit conversion differences between audiences they previously served identically.

The first screenshot is the single most valuable piece of design work in a consumer app. It is seen by everyone and read in under two seconds.

What still moves the needle

  • The first frame of the preview video. Most viewers never reach the second one, so it has to carry the entire proposition.
  • Ratings velocity, not just the average. A steady flow of recent reviews outranks a higher lifetime average with no recent activity.
  • Localisation depth. Translated metadata beats translated keywords, and native screenshots beat both.
  • Honest categorisation. Competing in a category you do not belong to buys impressions and loses ranking once engagement data arrives.

Where discovery is going

Editorial placement and AI-assisted store search are both growing shares of discovery, and both reward clarity over keyword density. A description written for a person — what the app does, for whom, in the first two sentences — now performs better in machine-assisted search than a description engineered for a crawler. That is an unusually pleasant alignment of incentives.

The practical takeaway is simple: your store listing deserves the same review cycle as your onboarding flow, because it is the first screen of your product.