How Product Page Optimization works
In App Store Connect you create a test with up to three treatments next to your original product page. Each treatment can change:
- the app icon,
- the screenshots,
- the app preview videos.
You decide what share of App Store visitors sees the test and in which localizations it runs. Visitors are split randomly between the original and the treatments, and App Store Connect reports the conversion rate of each variant together with a confidence level. A test can run for up to 90 days, and you can apply a winning treatment as your new default page.
What you can't test with PPO: the app name, subtitle, description or price. Those stay the same for all variants.
How to set up an App Store product page test
- Pick one hypothesis. For example: "screenshots that show the result beat screenshots that show the interface".
- Prepare the assets. Alternative icons must already be included in the app binary, so an icon test needs an app update first. Screenshots and previews can be uploaded directly.
- Create the test in App Store Connect under Product Page Optimization: name it, choose the number of treatments, the traffic share and the localizations.
- Submit the treatments for review. Apple reviews treatment assets before the test can start.
- Let it run until App Store Connect shows a result with enough confidence, ideally over at least one or two full weeks, so weekday and weekend visitors are both included.
- Apply the winner or keep the original, and write down what you learned for the next test.
What to test
| Element | Test ideas |
|---|---|
| First screenshot | Benefit headline vs. feature headline; real photo vs. UI; social proof ("4.8 stars from 20,000 users") |
| Screenshot order | Lead with the most-requested feature from your app reviews |
| Icon | Colour, simplified symbol vs. detailed illustration |
| App preview | With vs. without video; first 3 seconds showing the result |
The first two or three screenshots matter most, because they are what users see in search results without scrolling. See app store screenshots for sizes and limits.
How to read the results
- Look at the confidence level, not just which variant is ahead on day three. Early leads often disappear.
- Watch for seasonality. A test that runs over a holiday or a featuring week may not repeat.
- Check your biggest storefronts separately if the test runs in several localizations; a winner in the US may lose in Japan.
- A test without a winner is still a result: it tells you that element is not your bottleneck.
PPO vs. custom product pages vs. store listing experiments
| Product Page Optimization | Custom product pages | Store listing experiments | |
|---|---|---|---|
| Store | App Store | App Store | Google Play |
| Purpose | A/B test the default page | Separate pages for specific traffic (ads, campaigns) | A/B test the store listing |
| Who sees it | Random share of all visitors | Only people with the page's link | Random share of visitors |
| What can change | Icon, screenshots, previews | Screenshots, previews, promotional text | Icon, graphics, screenshots, descriptions |
Common mistakes
- Testing three unrelated changes at once, then not knowing which one made the difference.
- Stopping too early because one variant looks ahead after a few hundred visitors.
- Testing details nobody sees, such as the eighth screenshot, instead of the first impression.
Example
A fitness app tests three screenshot sets for 30 days. The variant that shows a workout in progress beats the original by 8% in conversion rate with high confidence and becomes the new default. The team's next test uses the same idea for the app preview video.