Ratings and reviews: how to ask, reply and read them on both stores

What the rating people see is made of#

Apple shows the summary rating on the product page and in search results, and says ratings and reviews can influence how your app ranks in App Store search. Google ranks apps on a combination of ratings, reviews, downloads and other factors. Neither store publishes the weights.

So read the rating per country: a bug that hits one country shows mostly there. A bad version fades from the Play rating as new ratings arrive; on the App Store, your lever is the reset. The ratings page has the short version.

Asking for a rating on the App Store#

Use the system prompt. In SwiftUI, read the requestReview environment value to get a RequestReviewAction and call it. In UIKit, call AppStore.requestReview(in:) with the window scene; StoreKit lists both under requesting reviews. Both need iOS 16 or later, so an app that supports older versions still calls SKStoreReviewController on them. That class is deprecated from iOS 18, though Apple's own ratings page still names it.

Calling it is a request: App Store policy governs whether the prompt appears. StoreKit's rules:

• A person who hasn't rated or reviewed the app on this device sees the prompt at most three times in a 365-day period.
• A person who has rated or reviewed on this device sees it again only for a new app version, and only once more than 365 days have passed since their last review.
• People can also opt out of these prompts for every app they have installed.

Because the prompt may not appear, Apple says not to call it from a button tap or other user action. For a "Rate this app" item in Settings, link to your product page with the query parameter action=write-review, which opens the review page.

Apple's Human Interface Guidelines recommend: ask only after people have shown engagement, such as finishing a level or a significant task; not on first launch or during onboarding; at a natural break, not mid-task; and consider allowing at least a week or two between requests. Apple's sample code asks only once per app version, after a task has been completed at least four times, and after a pause on the completion screen. Apple calls the four arbitrary; choose your own.

NoteA development build always shows the prompt; in a TestFlight build the call does nothing. Neither tells you how often real users see it.

Asking for a rating on Google Play#

Google's In-App Review API shows a card inside your app with 1 to 5 stars and an optional comment. Google enforces a time-bound quota on how often a user can see it, doesn't publish the value, and says it can change without any notice. One hint: calling launchReviewFlow more than once in less than a month might not always show a dialog. Google's rules:

• Ask after a user has experienced enough of the app or game to give useful feedback, and don't prompt excessively.
• Show the card as it is: no change to its size, opacity or shape, no overlay on or around it, on the topmost layer, and never removed by your code.
• No button that triggers the API: the user may have hit the quota and see nothing. Point buttons at your Play Store listing.
• No questions before the card, as below.

The API doesn't report the outcome. Google's integration guide says it does not indicate whether the user reviewed, or even whether the dialog was shown, so your app carries on either way, error or not. The ReviewInfo object it needs is valid only for a limited time: request it ahead, but only once you're sure you will show the card.

To test it, use the internal test track, where the quota isn't enforced. The tester's account must be the primary account in the Play Store, have the app in its library and not have reviewed it yet; a protected account, such as an enterprise one, can't review. With internal app sharing the submit button is disabled. FakeReviewManager fakes the API's results for automated tests and shows no card. The rest of the Android listing is in Google Play ASO.

What both stores forbid#

That is guideline 5.6.1, so a star screen of your own is out on the App Store. Guideline 3.2.2(x) adds that an app must not force people to rate or review it to reach its features or content. And the guidelines' section on business says:

Note "filtered". Apple doesn't define it, but a "Do you like the app?" screen that sends only the yes answers to the rating prompt is the plainest case of it, and Google's API guidance rules out the same question in so many words.

Google Play's ratings policy prohibits inflating ratings, reviews or installs by illegitimate means, including incentivized ratings and reviews. Its first example is a notification offering a discount in exchange for a high rating. Its don'ts add deceptive pop-ups, forced ratings and automated services.

What stays allowed:

• A Rate or Review link in Settings that opens the store page.
• A support link that is always there. Apple asks you to make support contact easy to find in the app and on the product page, so people with a problem can reach you before they write a review. A link anyone can use is not a filter; a question that decides who sees the prompt is.
• Your rating in marketing: Apple allows it if it matches your current rating. Quoting a review needs the reviewer's permission.

Resetting the App Store rating#

You can reset the App Store overview rating when you release a new version. In App Store Connect, open the version and, in the Reset Overview Rating section, select "Reset rating when this version is released". It needs the Admin or App Manager role. Apple's help page holds the details that trip people up:

• It resets every country and region at once.
• Once the version is released, the old rating can't be restored.
• Written reviews stay on the page.
• Until enough people rate the new version, the page says the rating was recently reset. Apple doesn't say how many ratings "enough" is.
• For a universal app, it resets only the platform you chose. App bundles can't be reset.

Apple recommends using it sparingly: the reset rating reflects the current version, but having few ratings may discourage people from downloading. Google Play has no reset.

A decision rule: reset only when three things hold. The low ratings are about a problem this version fixes. That problem shows in the rating in your main countries, not just one, since the reset clears all of them. And new ratings arrive often enough to rebuild a summary within weeks. If any one is missing, keep the rating and fix the problem.

Replying to reviews#

If you can't answer everything, Apple suggests starting with the lowest star ratings and reviews that mention technical issues with the current version. Name the specific point and what you're doing about it. When an update fixes a problem older reviews named, put it in the release notes and reply to those reviews to say so.

Two exceptions on the App Store: in your reply, point reviewers with download or billing problems to Apple Support, and use Report a Concern for offensive or spam reviews instead of replying (the reviewer isn't told). Apple can also email you when someone edits a review you answered. On Play, Google discourages automated replies you plan to rewrite later: through the API only the first reply notifies, so a placeholder spends it.

Does a reply move the rating? Play Console's Updated ratings section shows how returning reviewers changed their rating, split by whether they got a reply. Google's Play business page says responding to a negative review "can increase that rating by an average of +0.7". It doesn't say how, when or across which apps that was measured (the line has been on the page since at least 2021), or whether the average counts only reviewers who came back to edit: treat it as Google's figure, not a forecast.

Reading reviews: by version, by country, by complaint#

In App Store Connect, Ratings and Reviews lets you pick a country or region from the menu at the top left, and filters reviews by version, by star rating, and by whether a review was edited or answered. The App Store Connect API lists reviews per app or per version, filtered by territory and rating.

Play Console's Reviews page filters by date, language, reply state, star rating, app version and device, has a search box, and translates reviews into your Console language. The Ratings page splits the rating by country, language, app version, Android version, device type, device model and operator, against your peers' median. Reviews analysis adds theme summaries and benchmark topics such as stability and speed, for English reviews only.

A rating doesn't need a written review. On Play, the Ratings page counts stars by version and country, but the Reply to Reviews API returns only reviews with a comment. On the App Store, tvOS apps can be rated but not reviewed. So on Play, when a version's stars drop and its reviews are few, read the rating breakdown before the text.

1. Pick the current version and a fixed window, such as the 14 days since its release.
2. Filter to 1- and 2-star reviews, one store at a time, in your main countries.
3. Give each review one label for its main complaint: a crash, sign-in, the paywall, a missing feature. Reuse the labels week to week.
4. Count the labels. One that tops the count on this version and not on the last is your candidate.
5. On Play, check it against the stars. If the version's share of 1- and 2-star ratings rose while that label grew, both point the same way.

A weekly routine, with a worked example#

Once a week per app, as part of the loop in how to do ASO:

1. Write down the ratings with dates: on Google Play, per main country and for the current version; on the App Store, the overview rating per main country.
2. Read and label the new 1- and 2-star reviews.
3. Reply to the low-star reviews about the current version first.
4. Log each time your code asks for a rating, so you know at least how often you asked.
5. When a fix ships, note the day it went live, name the fix in the release notes, and reply to the reviews that named the problem.

The reminders complaint repeats on both stores and is new with 3.4, so it is what to fix. Version 3.4.1 ships with "Fixed: reminders that stopped firing in 3.4" in the release notes, and the 17 reviews that named it get a reply saying which version fixes it. No App Store reset: the problem lasted two weeks, and a reset would clear the rating in every country.

Fourteen days later, the note reads: "3.4.1 averaged 4.2 from 40 Play ratings in its first 14 days, against 3.2 from 50 for 3.4. Before and after, not a controlled test." At 40 ratings the average is fragile: four more 1-star ratings would take it to about 3.9.

Where RocketASO fits#

RocketASO reads your App Store and Google Play reports, names the fix estimated to be worth most this week, and compares installs before and after it, at 7 and 14 days. Each read is a before-and-after read, not a controlled test. It covers reviews and ratings too. It starts you with a draft of the copy or the keyword field, and nothing reaches the store until you approve it. To see where a listing stands first, run the free ASO audit.