Screenshot opinions are cheap; screenshot data is not. Both stores now give you native A/B testing for your listing images, but they work differently enough that a result from one says little about the other. Here is how to test screenshots properly on each store in 2026, and what to test first.
Apple: Product Page Optimization
Product Page Optimization (PPO) lives in App Store Connect and lets you run up to 3 treatments against your default page, testing screenshots, app previews and icon. Tests can run up to 90 days, traffic is split among variants, and each treatment goes through App Review before it can start, so plan a day of buffer. A quirk worth knowing: PPO variants apply per storefront, which pairs naturally with localized screenshot sets: what wins in Japan is not what wins in Brazil.
Related but distinct: Custom Product Pages, which have grown from 35 to 70 per app and are now linkable to search keywords. They are targeting tools, not A/B tests, but they reuse the same asset discipline: more variants means you need a workflow that produces screenshot sets quickly.
Google: Store Listing Experiments
Play's experiments are older and looser: you test listing variants (including screenshots and the feature graphic) directly from the Play Console, with traffic percentage under your control and no review gate for each variant. Results come with confidence intervals; resist the urge to stop a test the first day a variant is ahead.
The rule everyone breaks
Never transpose a result across stores. Audiences, layouts, and rendering differ: Play shows your graphic and screenshots differently than the App Store does, and the user populations behave differently. A winning first screenshot on iOS is a hypothesis on Android, not a conclusion.
What to test first
Slide one, always: the first screenshot carries a disproportionate share of conversion. Test one variable at a time: caption phrasing, background color, benefit order. Swapping the entire set tells you that "something" worked, which teaches you nothing reusable. Keep a change log next to your test results; three months later you will not remember what treatment B was.
Producing variants without pain
The practical bottleneck of screenshot testing is producing clean variants. If each variant costs a design afternoon, you will not iterate. With a preset-based workflow (how the process fits together), a variant is a caption edit and a re-export, which is what makes weekly iteration realistic.
The short version
PPO on Apple (3 treatments, 90 days, review required), Store Listing Experiments on Play (looser, no gate). Test slide one first, one variable at a time, never copy conclusions between stores, and make variant production cheap enough that you actually iterate.
