Five changes, one theme: the listing is read by models now
Between WWDC 2025 and spring 2026 Apple changed the App Store more than in the five years before. Screenshot captions entered the search index in June 2025. Custom product pages gained keyword linking in July 2025 and started appearing in organic search, and the limit rose to 70 per app in October 2025. App Store Tags, generated by Apple's models from your metadata, began driving browse placement. The ranking started weighing rating recency and post install retention more heavily. And in March 2026 App Store Connect gained over a hundred metrics, including cohort analysis and peer benchmarks.
The theme underneath: your listing is no longer only matched, it is read. A model reads your description to decide what tags you get, reads your captions to decide what you rank for, reads your retention to decide whether you deserve the rank. Writing for that reader is the job, and it happens to be a job agents are good at.
Screenshot captions: the newest keyword slot and the most visible one
Captions on screenshots are indexed, and the first three screenshots appear directly in search results on iOS. So the caption is both a ranking input and the first thing a searcher reads. Six words or fewer, the outcome first, the title terms reused, no repeats across the three, and every localized set in its own language. Apps that A/B test screenshots are still the minority; on the App Store most apps update fewer than four times a year. That is the opening.
Write three captions per locale from the keyword plan, six words or fewer, outcome first, reusing title terms without repeating between captions. Generate the screenshot sets and queue a product page optimization test with the new captions against the current ones. Report impressions and conversion per country after 14 days.
Custom product pages in organic search: build one per keyword group
Custom product pages used to be an ads only feature. Since July 2025 they can be linked to search terms and appear organically, and with 70 available per app there is room for one page per keyword group in every major country. A page for sleep sounds with timer shows the timer in screenshot one. A page for rain sounds shows rain. Conversion lifts from matched pages sit around five to nine percent in category benchmarks, and the same page serves the ad click and the organic click.
Group our target keywords into intents, at most ten groups per country. For each group build a custom product page whose first three screenshots and captions match the intent, link it to the group's keywords for organic search, and use it as the destination for any ad on those keywords. Track impressions and conversion per page and retire pages that do not beat the default listing within 30 days.
Rating recency and retention in the ranking
A great average from last year is worth less than a wave of good ratings this month, and an app that keeps users holds its rank more firmly than one that does not. This moves review replies and retention fixes from support work to ranking work. Ask for a rating after the first win, not at launch. Answer every review the same day. Fix the day 3 leak before buying anything.
Move our rating prompt to right after the first win in the app and never show it in the first session before that moment. Answer every review within hours. Every Monday show me the rating trend for the last 30 days against the previous 30, and the day 7 retention trend, next to our rank for the top ten keywords, so I can see the three move together.
The hundred new metrics, and the three that matter
App Store Connect's expanded analytics give you cohort retention, peer benchmarks and source attribution that used to need a third party. An agent with read access can now say whether your conversion is above or below the category median for your traffic source, which turns every listing test into a comparison against the market rather than against last week. Watch three things: product page conversion against benchmark, day 7 retention per cohort, and the share of impressions that come from browse rather than search, because that share is where tags show their effect.
From App Store Connect, report weekly: product page conversion against the category benchmark per traffic source, day 7 retention per cohort against the benchmark, and the share of impressions from browse versus search. When browse impressions rise after a description change, attribute it to the tag change and tell me.
