What changed on the app store in the first half of 2026.
A quarterly review of the structural shifts in mobile app discovery — featured placements, the new review-prompt rules, the slow-motion consolidation of certain categories. With observations from the marketing teams who have had to respond to all of it.
The first half of 2026 has been, in app-store-discovery terms, a quieter period than the first half of 2025. There have been no headline policy changes of the kind that defined the previous two years. The major platform owners have not, in the past six months, announced any new revenue-sharing structures. The published guidelines for what an app can and cannot do have been, by recent standards, stable.
What there has been, however, is a series of smaller, less-announced changes to the underlying mechanics of how apps are discovered, surfaced, and reviewed on the two major platforms. The changes are not, individually, dramatic. Taken together, they are reshaping the work of app marketing in ways that the marketing teams I have spoken to are still adjusting to. This is a report on what those changes are, what we have been able to confirm about them, and what the early data is showing.
The change in featured placements
The first change, and probably the most consequential, is in how featured placements work. The "featured" sections of both major app stores have, for most of the last decade, operated on a relatively predictable rotation: a small editorial team would select a small number of apps each week to feature, the apps would receive a substantial but predictable boost in traffic during the feature period, and the cycle would repeat. This model worked, broadly, until late 2025.
What changed in late 2025 — and what has continued to play out across the first half of 2026 — is the introduction of a more dynamic featuring system that does not behave like the old editorial rotation. The new system, as far as we can reconstruct from the published data and from the conversations we have had with marketing teams, weighs a much larger number of signals about each app before deciding whether to feature it: not just editorial assessment, but also user engagement on similar apps, recent download velocity, review velocity, and a number of signals about the quality of the app's listing copy and imagery.
The practical effect for marketing teams is that the "featured" boost is now considerably less predictable than it used to be. Apps that would, under the old system, have been confident of a feature based on editorial relationships are now, in many cases, not being featured. Apps that would, under the old system, have been unlikely candidates for featuring are, in some cases, being featured for short periods based on signal patterns that the marketing teams did not specifically optimise for. The variance has gone up substantially.
The change in review prompts
The second change is in the rules governing in-app review prompts. Both major platforms have, since their last guideline updates in 2024, allowed apps to prompt users for reviews using built-in platform APIs that respect a per-user rate limit. The rules around when an app is allowed to use those APIs have, however, tightened materially in the first half of 2026.
The new rules are not, on a surface reading, dramatic. The main change is that the platforms are now penalising apps that prompt for reviews "inappropriately" — by which they mean, in practice, prompting in moments when the user is likely to be frustrated, distracted, or otherwise not in a position to give the prompt due consideration. The penalty is not, as far as we can tell, a fixed value. It appears to be a reduction in the weight given to the app's reviews in the discovery algorithms, applied gradually over a period of weeks following the inappropriate prompting.
This is, in operational terms, a significant change. The conventional wisdom on review prompts — that the right thing to do was to prompt every eligible user at the earliest reasonable opportunity — is no longer correct. The teams I have spoken to are, almost without exception, in the process of rebuilding their review-prompt logic to fire less often, in more carefully-chosen moments, and with a stronger filter for "the user is likely to leave a thoughtful review."
"The new review rules do not say 'don't prompt.' They say 'don't prompt badly.' The companies that figure out the difference will be in a noticeably better position by the end of the year."
The category consolidation
The third change is the slowest and, in some ways, the most consequential. Several mid-sized categories on both stores are, by every measurement we can apply to them, consolidating. The top three or four apps in each category are taking a larger share of total downloads and revenue than they were eighteen months ago. The long tail of smaller apps in each category is, in download terms, getting smaller.
The categories where this is most visible are: meditation and sleep, personal finance, fitness, photo and video editing, and reading. The categories where the pattern is less visible — and where the small-app long tail remains, broadly, healthy — are: productivity, utilities, education, and the broader category of niche professional tools.
The reasons for the divergence are, we think, mostly structural. The categories that are consolidating are categories where the user-acquisition mechanics favour incumbents. The categories that are not consolidating are categories where the user-acquisition mechanics still favour, or at least do not penalise, smaller apps. A category dominated by paid social acquisition consolidates faster than a category dominated by organic search or by word-of-mouth.
The implications for marketing teams
The three changes I have described above are, in aggregate, making the discovery side of the app-marketing job harder in ways that the underlying budget data does not necessarily reflect. The work of getting an app discovered is now less predictable, more dependent on signal-level optimisation, and more variable in its outcomes than it was eighteen months ago. The same effort that produced a particular result in mid-2024 will, in most cases, produce a noticeably worse result in mid-2026.
What we are hearing from marketing teams, when we ask how they are responding, is broadly consistent. The companies that are managing the change best are the ones who have invested in a stronger relationship between the marketing function and the product engineering team. The signal-level optimisations that the new discovery mechanics reward — review-prompt timing, in-app event tracking, conversion-event setup, the listing-copy and imagery work that the featuring system now weighs — are all jobs that sit at the boundary between marketing and product. The companies that have organised those jobs in the same team, with shared metrics and shared release cycles, are getting better outcomes than the companies that still treat them as separate functions.
What we expect for H2
I will keep the predictions brief, because predictions in this domain age badly. We expect, broadly: the featuring system to continue to be less predictable, with the variance widening rather than narrowing as the platforms refine the underlying models. The review-prompt rules to settle into something closer to a stable equilibrium, with the apps that have adjusted to the new norms in a better position than the ones that are still using the 2024 playbook. The category consolidation to continue in the categories where it is already visible, and to begin to be visible in one or two new categories — probably one of the messaging or social-adjacent categories — by the end of the year.
None of this should be read as a forecast we are committed to. The history of writing about app-store discovery is a history of confident predictions that did not survive the next platform update. What it is, instead, is a snapshot of where the underlying mechanics seem to us to be heading, based on the data we have been able to collect and the conversations we have had over the past six months.
The next report in this series will be in early October, after the autumn-platform-update cycle. We expect to have more to say at that point about how the changes described above have continued to play out. In the meantime, the patterns above are the patterns we are watching most closely.