Apple shipped the largest overhaul of App Store Connect Analytics since the tool launched, and most indie developers haven't opened the new dashboard yet. That's a mistake. Buried in the announcement are capabilities that used to require a third-party analytics stack — cohort analysis, peer benchmarks, filterable subscription funnels, and a proper reports API — and they're now native, first-party, and free.
This guide walks through what's actually new, which metrics matter for small teams, how to set up cohorts and benchmarks that are worth looking at, and what to do with the data once you have it. If you're shipping an iOS app today without pulling App Store Connect data into your decisions, you're leaving installs, retention, and revenue on the table.
what changed in march 2026#
Apple announced the update on March 25 and began rolling the new metrics into App Store Connect through April. Five changes matter:
More than 100 new metrics. The old dashboard had roughly 30. The new one crosses 130, with the biggest additions in monetization, subscriptions, and user-behavior dimensions.
Cohorts. You can now group users by a shared attribute — download date, download source, offer start date, region, acquisition campaign — and track how that cohort behaves over time. This was previously a Mixpanel or Amplitude thing. It is now a free Apple thing.
Peer benchmarks with differential privacy. Two new benchmarks — download-to-paid conversion and proceeds per download — compare your numbers against an anonymized peer set. Apple is using differential privacy to publish the number without exposing any individual developer's data.
Up to seven simultaneous filters. The old dashboard capped you at one or two. The new one lets you drill "users in Germany, on iPhone, who came from Search Ads, who started a free trial between March 1 and March 31, who opened the app at least three times" in a single view.
Analytics Reports API expansion. Two new subscription reports are now exportable. You can pull them into your warehouse, join them to your own product-analytics data, and keep the whole thing consistent.
the metrics that actually matter for indie devs#
A hundred new metrics is enough to drown in. The five that earn their place in a weekly review for a small-team app:
1. download-to-paid conversion (with peer benchmark)#
This is the single most useful number Apple just gave you. It measures the share of users who downloaded your app and then became paying users within a defined window. The benchmark next to it tells you where you sit relative to comparable apps — and "comparable" is defined by Apple's categorization, not by you, so the comparison is harder to game.
If your number is meaningfully below the benchmark, you have a conversion problem. That could be onboarding, paywall placement, pricing, trial length, or a dozen other things — but the benchmark is telling you there is room to move, and the peer set is hitting numbers you're not. If your number is above the benchmark, you have a scaling problem: your conversion is fine, you need more top-of-funnel.
2. proceeds per download#
Proceeds per download normalizes across install volume and pricing model. A $4.99 one-time-purchase app and a $9.99/month subscription app with a 10% conversion rate can land in similar proceeds-per-download territory, which lets you reason about your business the way a portfolio investor would: revenue per acquired user.
Watch this over six-week windows rather than week-to-week. It's noisy on short horizons, signal on long ones.
3. subscription retention by cohort#
Previously you could see subscription retention in aggregate, which was close to useless because the aggregate number is always dominated by your longest-lived cohort. The new cohort view lets you see retention for users acquired in a specific campaign, a specific month, or a specific offer — and compare that to other cohorts.
This is where you catch the onboarding regression that has been quietly degrading your retention for four months. Before, it hid inside the aggregate. Now, a newer cohort that's retaining worse than an older one is a visible line on a chart.
4. impressions → product page views → downloads funnel#
The full App Store funnel — impression on a search, click to the product page, install — is now explicit in Analytics with conversion rates at each step. You can split it by source: search, browse, referrer, App Store Ads.
This tells you which part of your store listing is the bottleneck. Low impression-to-view conversion means your icon, title, or subtitle aren't compelling enough on the search results page. Low view-to-install conversion means your screenshots, description, or ratings aren't closing the deal once the user is on your page. These are different problems with different fixes and the aggregate number used to obscure which one you had.
5. offer redemption and trial-to-paid#
If you run promotional offers or free trials, the new trial-to-paid cohort view is the metric that matters most. You can see trial-to-paid conversion for offer code redemptions separately from organic trial starts, which is important because the two populations behave very differently and lumping them together was misleading for years.
how to actually set up a useful cohort analysis#
The UI for cohorts is decent but not obvious. The rough pattern that works:
Pick a meaningful attribute — almost always "download date" or "download source" to start. Define the window: a month is usually the right granularity for indie apps, because week-over-week noise is too high to act on.
Pick your outcome metric. Retention at day 7, day 30, and day 90 is a reasonable default. For subscription apps, also track paid conversion at day 7 and day 30 — most trial conversions happen in the first week, and the rest are a slow tail.
Compare sequential cohorts. You're looking for two things: absolute numbers that tell you whether the business is healthy, and deltas between cohorts that tell you whether your recent changes are helping or hurting.
Don't try to analyze 10 cohorts at once. Pick three — last month, last quarter, a year ago — and compare those. More than that is visual noise you can't reason about.
the peer benchmark — how to read it without freaking out#
Apple's benchmark numbers are differential-private aggregations over your peer category. That means two things. First, the number you're comparing against is statistically close to the true peer average, but not identical to it. Second, the number is slow to update — Apple is sampling across thousands of apps and applying privacy noise, so your benchmark won't bounce around daily.
The right way to use the benchmark is as a floor and a ceiling, not a target. If your conversion is below the 25th-percentile number, you have work to do. If it's above the 75th, the top-of-funnel problem is the leverage. If it's in the middle, the marginal gains are smaller than the gains from other parts of your stack — and a different metric is probably where you should be focusing.
Do not obsess over chasing the benchmark up a percentile. An app with unusually high retention can be funding a conversion rate that's below the peer benchmark, and that's a perfectly fine business.
the api — what indie devs should actually export#
The new subscription reports via the Analytics Reports API are the piece that unlocks real workflow integration. Two reports are exportable today:
Subscription retention events, granular enough to pipe into your warehouse and join against your own event data. If you track in-app events in a separate system (Firebase, RevenueCat, Amplitude, a homegrown pipeline), this is how you reconcile "did the user who churned do something specific in the app beforehand."
Subscription offer performance, split by offer code and redemption status. If you run any kind of promotional pricing — seasonal discounts, winback offers, intro pricing — this report tells you which offers actually paid for themselves.
Expect the rest of the subscription data in Sales and Trends to migrate into exportable API form over the next 12 months. Build your pipeline against the new API, not against Sales and Trends.
a practical weekly review workflow#
For a solo or small-team developer, an hour a week is enough to get real value out of the new analytics. The rough outline:
First ten minutes: scan your download-to-paid and proceeds-per-download against their peer benchmarks. If either has moved meaningfully — up or down — drill in. If both are stable, skip to the funnel.
Next ten: look at the full funnel — impressions, product page views, downloads, paid conversion. Note which step moved most. Correlate with anything you changed in the last two weeks (screenshots, description, pricing, offers).
Next twenty: pull up cohort retention. Compare this month's cohort to last month's at day 7. If there's a delta, figure out whether it's a product change (did you ship something that might have degraded onboarding) or an acquisition-mix change (did you shift more spend to lower-quality sources).
Final twenty: write two sentences. "This week's read is X, which suggests Y, which means my highest-leverage action for next week is Z." Save those. Over three months, the pattern in those notes is worth more than any single data pull.
where stora fits#
App Store Connect analytics tell you what's happening; the hard part is acting on what they tell you. That's where the submission pipeline matters. If your funnel data is pointing at a low view-to-install conversion, the next step is a screenshot refresh or a description rewrite — and if that takes two weeks of localization work, the data is stale by the time the change ships.
Stora's store-listing automation closes that loop. The screenshot generator regenerates your full set across every device size in one pipeline run. The AI store listing tool rewrites descriptions in every locale at once. The one-click publish flow pushes the update without you hand-managing metadata. The data tells you what to change; the platform makes the change fast enough that the data is still true when the change goes live.
the takeaway#
App Store Connect Analytics went from a flat dashboard to a real analytics tool in one release. The companies that integrate it into their weekly workflow in the next 60 days will have a data advantage over the ones that don't — and the peer benchmarks make that advantage quantifiable. Open the new dashboard. Build a cohort. Write a weekly note. The tools are here now; the only variable left is whether you use them.