How to Set Up a Product Page Optimization Test

Product page optimization compares alternate app icons, screenshots and app previews against your original page. Results appear in App Analytics, and you can set the winning version to display to everyone. It applies to iOS and iPadOS apps on iOS or iPadOS 15 or later, needs the app to be Ready for Distribution, and is not available for custom product pages, Apple Watch or iMessage product pages. It does not test text such as the description or subtitle. For ideas on what to test, see App Store A/B testing ideas.

1. Write a hypothesis first. "Showing the streak chart in screenshot 1 will raise conversion over the current checklist image." A test without a hypothesis produces numbers without a decision.
2. Change one thing. If you change icon and screenshots together, you cannot tell which helped.
3. Name the test. The reference name can be up to 64 characters. Make it findable in App Analytics, for example "Streakly Nov 2026 screenshot 1 caption test (US)", which is 48 characters.
4. Choose the number of treatments. Up to three. More treatments divide the traffic and slow every one of them down.
5. Set the traffic proportion. This is the share of visitors randomly shown a treatment. People see the same treatment on each visit during the test.
6. Pick localizations. Each treatment can be localized in all the languages the app supports or only those you choose. A test may take longer to reach meaningful results depending on the localizations you select, so start with the storefronts where you have the most traffic.
7. Prepare assets. Screenshot and preview treatments can be submitted without a new app version. Alternate icons are different: they must already be part of the app binary for the current App Store version, built with Xcode 13 or later with alternate icon support, and sized 1024 by 1024.
8. Submit and start. The test runs for 90 days or until you stop it, and a stopped test cannot be restarted.
9. Wait for results. Apple shows results after five first-time downloads associated with the test, but five downloads is not a verdict. Wait for confidence.
10. Apply or discard. Apple recommends waiting until at least one treatment is declared better or worse than the baseline with at least 90% confidence. Applying a treatment while the test runs ends the test. If the winner uses an alternate icon, include it in your next app version.

A hypothetical habit tracker, Streakly, tests three screenshot treatments and allocates 30% of traffic. That leaves 30 / 3 = 10% of visitors per treatment and 70% for the original. If the app receives 1,000 product page views a week in the tested storefront, that is about 100 views for each treatment, 300 in total, and 700 for the original. Apple's own example shows the same arithmetic: 40% of traffic across two treatments gives 20% each and leaves 60% for the original.

Fewer views per treatment means a longer test. Streakly can instead allocate 50% to a single treatment, 500 views to each side. The trade-off is risk: a bad treatment costs half the traffic while it runs. For a small app, one treatment at a time is usually the better design.

Not every test produces a winner. Apple's own developer stories on this page include a 12-day test of an app preview where the original product page outperformed the treatment, contrary to the developer's hypothesis. A loss is a result: it tells you not to ship the change.

Record what you tested, the dates, the storefront and the outcome. Then change the next element. Do not run a new test on top of a change you applied the same week.

Read screenshots that convert for Apple's guidance on what to put in them, and the localized screenshots guide for multi-language tests. LocalASO covers the text side of the listing; see the Optimize docs.