How to Ask for an App Store Review

A receipt scanner that asks for a review on first launch is begging a stranger who has not scanned anything. The same app, asking after the third receipt that actually saved, is talking to someone who just finished the job. Apple’s sample does the second thing. It waits until a multi-step process is done, then calls the system prompt. The number of completions in the sample is arbitrary. Three successful scans is a reasonable line for this app because a scan is the job.

The rules live on Requesting App Store reviews . The system displays the review prompt a maximum of three times within a 365-day period. You pick the moment. The system can still decline to show the sheet, and people can turn requests off on the device. What the stars do to search is a different question, covered in do ratings change App Store search rank. Public replies are in how to reply to App Store reviews.

Use the system prompt, not a homemade star dialog

SwiftUI reads the requestReview environment value and calls it as a function. That is the current way to present the system sheet. Older code calls the StoreKit review request by name. Either way, you are asking the system to present Apple’s sheet. You do not draw the stars yourself, you do not skin the card, and you do not decide that this user is “a promoter” before the sheet exists.

Apple’s sample delays the call on purpose. The prompt appears from a “process completed” scene, not from a button whose only job is Rate. The sample also skips the prompt if this bundle version already asked, and it waits until the person has finished the process several times. Their sample uses four completions and says the count is arbitrary. Copy the shape, not the integer: a finished job, a version you have not already prompted, and a short pause so the sheet does not land on top of the save animation.

The receipt scanner: third successful scan

Count successful saves, not app opens. A successful scan means the camera returned a total, the person confirmed it, and the receipt landed in the list. Failed captures, blurry retakes, and the empty-state tutorial do not count. On the third success, if this version has not already asked, wait a couple of seconds on the saved-receipt screen and then call requestReview.

The prompt belongs after the third saved receipt on a version you have not already asked about. Launch, the first scan, a Rate button, and a failed purchase are the wrong moments.

Store two values on device, the way the sample stores them: how many successful scans have completed, and the last bundle version that prompted. If the count is at least three and the version string is different, present the request, then write the current version down so the same build does not ask again. Days can pass between scans. That delay is a feature. Someone who has filed three real receipts can say whether the total was right.

Why the enjoyment gate is the risky part

The pattern looks polite. A card asks “Are you enjoying the app?” Yes opens the system sheet. No opens a mail compose to you. You never “bother” an unhappy person, and your average only hears from fans. Review teams already know this shape. It is a known rejection pattern: you showed the system prompt only on the happy path and hid the rest. It is also a bad product move. The people who would have tapped one star are the people whose notes name the bug.

Apple’s guidance says to avoid requesting a review as the result of a user action, and to make the request at the end of a sequence they successfully complete. A pre-question is a user action whose only purpose is to filter the prompt. Google’s in-app review guide is explicit about the same trick: the app should not ask whether they like the app, or whether they would give five stars, before or while presenting the card. You do not need a loophole. You need the third scan.

Incentives sit in the same bin. A free month, a coffee card, or “rate us and we will turn export on” is not a request. It is a payment for a rating. Do not do it. If someone wants to write on their own terms, put a row in settings that opens the product page with the write-review action. Apple’s sample uses https://apps.apple.com/app/idYOURAPPSTOREID?action=write-review with your real App Store ID. That link is for a person who chose to write. It is not a second copy of the system quota.

Google Play has its own in-app review API

Android is not “show a stars dialog you drew.” Play’s In-App Review API lets you trigger a Play card without leaving the app. Google enforces a time-bound quota and does not publish the number, and the number can change. Calling the flow often, including more than once in a short period such as less than a month, might not show a dialog. Do not hang a button on that call. If the quota hides the card, the button looks broken. Send the person to the store listing instead.

Play’s guide also says to trigger the flow after the person has used the app enough to have an opinion, and not to ask a predictive question first. The receipt scanner can use the same third successful scan on Android. Reply tactics are different on Play, because the console reply, the public notification, and partial rollouts are their own mess. Those live in the Google Play review reply post, not in this prompt.

What a prompt cannot fix

Three clean scans will not rescue a 1-star cluster about a wrong total. If the parser drops the tax line, the people who rate you from the system sheet will say so, and they should. Fix the parser, ship the version, and reply on the reviews that named it. Leave the prompt out of the paywall too. The job was the scan. A person who just closed a purchase screen is in a different scene.

A settings row is allowed to be a link

Some people look for a place to review and get angry when they cannot find it. Give them one. A settings row labeled Review on the App Store opens the write-review URL. It does not call requestReview. You will not burn the three-in-365 cap on someone who already decided to write, and you will not show a dead button when the system suppresses the sheet. Keep the copy plain. “Review” is enough. “Rate 5 stars” is a nudge you do not need.