App Store and Google Play Ratings: How Theyre Calculated (and Reset)

How Apple calculates the rating shown on the App Store

Apple's own "Ratings, reviews, and responses" page is direct about the mechanics: users rate an app from one to five stars, and those individual ratings "inform your app's summary rating, which is displayed on your product page and in search results." That summary rating is a straight average — every rating counts the same regardless of which version of the app it was left on or how long ago it was submitted. Apple doesn't quietly decay old ratings the way some platforms do; a five-star rating from three major versions ago still carries exactly as much weight as one from yesterday, unless a developer actively resets it.

The summary rating is also calculated per territory, not globally — an app can show 4.7 stars in the US App Store and a noticeably different number in another country's storefront, because each territory's rating is built from ratings submitted in that territory. Written reviews work alongside, not inside, that number: Apple's page notes that "written reviews allow users to share more detail about their experience with an app," but the star average itself is driven purely by the rating value, not by anything in the review text.

Resetting an app's overview rating

Apple gives developers a genuine reset button, but it's narrower than it sounds. Per App Store Connect's own help page on resetting an app's overview rating, a developer can opt in to a reset from the version page in App Store Connect — selecting "Reset rating when this version is released" before submitting a new version for review. Once that version ships, the overview rating resets on the App Store and a note appears on the product page indicating the rating was recently reset, until enough new ratings come in to produce a fresh number.

A few specifics matter more than people assume going into this:

• Written reviews don't reset. Apple's documentation states plainly that the reset "won't apply to written customer reviews, which will continue to display on the App Store" — so a reset clears the star average but not the text sitting underneath it.
• It's all-territory, and it's permanent. A reset "affects all regions or countries simultaneously," and once the new version ships, there's no undo — the previous rating can't be restored.
• Universal apps and bundles have carve-outs. For a universal app, the reset only applies to the platform version being reset, and app bundles can't have their rating reset at all.
• It requires the right role. Only an Admin or App Manager in App Store Connect can trigger the reset.

Apple's own guidance is also explicit that this isn't something to reach for on every release: the ratings-and-reviews page recommends "using this feature sparingly," since a newly reset app shows a thin or nonexistent rating for a while, and "having few ratings may discourage potential users from downloading your app." The honest use case is a version that fixes something that was genuinely dragging the old rating down — not a routine reset to erase a bad month.

How Google Play calculates the rating shown on the Play Store

Google Play's approach is structurally different, and Google says so directly. Per the Play Console help page on viewing and analyzing ratings and reviews, Play Console actually tracks two separate numbers: the "Google Play rating," defined as the rating "shown to users on Google Play," which "is calculated based on your most recent ratings," and a separate "lifetime average rating," which is the simple average going back to launch. Those two numbers can diverge noticeably — an app that fixed a major bug six months ago can show a current Google Play rating well above its lifetime average, because the visible number is, in Google's own words, "weighted towards more recent ratings to reflect changes and updates that you make to your app."

Google doesn't publish the exact weighting formula or the precise recency window, so treat any specific number of days you see quoted elsewhere as an estimate, not documented fact. What Play Console does confirm is a ratings breakdown that segments by country/region, Android version, and device type, so a developer can see where a rating is strong or weak without guessing. That segmentation is also the closest thing to an official acknowledgment that Play Store ratings aren't a single flat number behind the scenes — different cohorts of users can be rating a meaningfully different experience, especially on a lower-end device or an older OS version.

Why these mechanics matter beyond trivia

The practical consequence of Apple's all-time, equal-weight average is that a bad launch stays visible for a long time unless a team deliberately resets it — which is exactly why the reset option exists, but also why product and marketing need to coordinate before using it. The "recently reset" note Apple displays on a freshly reset product page is visible to every potential downloader, not just internal dashboards, so flipping that switch the same week as a big marketing push can work against the launch rather than for it. The right sequence is usually to reset on the version that actually fixes the underlying problem, let a reasonable number of new ratings accumulate, and only then lean into marketing that references the rating.

Google's recency weighting creates a different kind of planning problem: because the visible Google Play rating already discounts old ratings automatically, a team doesn't need — and can't trigger — a manual reset the way Apple allows. But it also means a short-lived wave of one-star ratings tied to a bad release can drag the visible number down faster than the lifetime average would suggest, and recover faster too, once a fix ships and new ratings start outweighing the old ones. Reading the Play Console ratings breakdown by country and device, rather than just the single headline number, is usually the faster way to tell whether a dip is a global regression or a problem isolated to one OS version or region.

The native prompts: SKStoreReviewController and the Play In-App Review API

Both platforms give developers a sanctioned way to ask for a rating without sending the user out to the store, and both platforms wrap that prompt in rules designed to stop it from being abused.

Apple: SKStoreReviewController

Apple's StoreKit framework exposes a review request through SKStoreReviewController, and Apple's documentation is specific about the limit: an app "can prompt for ratings up to three times in a 365-day period." Calling the API doesn't guarantee the prompt appears — Apple's system decides in the moment whether conditions are met, and a developer gets no callback telling them whether the user actually submitted a rating or dismissed it. Apple's guidance to developers is to call it at a natural pause in the experience, after something has gone well for the user, rather than immediately on launch or mid-task.

Google Play: the In-App Review API

Google's In-App Review API documentation describes a similar flow through the Play Core library (version 1.8.0 or higher), available on Android 5.0 and up with the Play Store installed. Google is explicit about the restrictions around it:

• A quota applies: "Google Play enforces a time-bound quota on how often a user can be shown the review dialog," and calling the launch method again too soon "might not always display a dialog" — the exact quota value is, per Google, "an implementation detail" that "can be changed by Google Play without any notice."
• No pre-prompt question is allowed: the app "shouldn't ask the user any questions before or while presenting the rating button or card," including anything like "Do you like the app?" that would filter who sees the prompt.
• No custom call-to-action button should trigger it, since a user who's already hit their quota would tap a button that silently does nothing — Google's guidance is to send that user to the Play Store listing directly instead.
• The review card itself can't be visually altered — no overlay, no resizing, no changing its design once it's on screen.

Neither platform allows incentivizing the prompt

Both app stores treat a paid or incentivized rating as a policy violation, not a gray area. Apple's App Review Guidelines warn that attempting "to manipulate reviews, inflate your chart rankings with paid, incentivised, filtered or fake feedback, or engage with third-party services to do so" can lead to removal from the Apple Developer Program entirely. Google's Play Console help page on user ratings, reviews, and installs is equally direct that offering money, goods, or any other incentive in exchange for a rating, review, or install is against policy — including something as specific as a discount offered in exchange for a high rating. Practically, that rules out anything that routes only happy users toward the prompt (a form of gating) or offers a reward for leaving one — both platforms can act on it, and it undermines the one number a team is trying to protect in the first place.

How Tailwind Score treats app store ratings

For a product team tracking reputation across more than one storefront, the version and territory quirks above aren't just trivia — they're the reason a flat average across App Store and Google Play numbers can be misleading. Tailwind Score treats App Store and Google Play ratings as first-class inputs into its pooled 0–100 reputation number, and because Apple's rating is per-territory and Google's is recency-weighted by version, Tailwind Score weights by region and version rather than averaging the headline numbers blindly. That matters most right after a release: a version that regresses in one region shouldn't get buried by strong numbers everywhere else, and a reset App Store rating shouldn't get read as a real reputation drop when it's actually a deliberate, documented developer action.

If your team already tracks App Store and Google Play ratings separately from everything else — website reviews, social mentions, support tickets — pooling them into one score makes the version-over-version trend visible instead of buried in two different dashboards. The same pooling logic applies to the other platforms Tailwind Score tracks, from Apple Maps' own thumbs-up ratings to Facebook's recommendation system — each with its own quirks worth understanding on its own terms before you fold it into a single number. See how Tailwind Score pools app ratings with the rest of your reputation data, region and version differences included rather than averaged away. For more on how a pooled score is built in the first place, see our guide to online reputation scores.