Product Page Optimization makes A/B testing your screenshots free. The hard part is knowing what to test. "Make the screenshots better" is not a test. "A benefit headline in the first screenshot will convert better than a feature name, because people scan search results for outcomes" is.
Each idea below has three parts: the change, the hypothesis (why it might win), and what counts as a win. Pick one, build the variant, and let the data decide.
Before You Pick an Idea #
- Test one change per variant. A test can hold up to three variants, so use them to try three versions of the same idea, not three unrelated ideas.
- Go big if your traffic is small. A 5% lift needs roughly 15 times the impressions of a 20% lift. Check how long your test will take with the A/B test calculator before you commit.
- Write the hypothesis down first. It stops you from explaining a result after the fact.
The First Screenshot #
Your first screenshots show up right in search results, and most people judge your app by them without opening the page. Start here.
1. Lead with a different feature
- Change: put your second-strongest feature in slot one.
- Hypothesis: the feature you're proudest of isn't the one people search for.
- Win: a higher conversion rate than the original.
2. Benefit headline vs feature headline
- Change: "Track expenses" becomes "Know where your money goes".
- Hypothesis: people buy outcomes, not features.
- Win: conversion lift with the same image. See how to write captions that convert.
3. Show the result, not the input
- Change: show the finished report, photo or plan instead of the empty form that creates it.
- Hypothesis: the payoff is more persuasive than the process.
4. A panoramic first pair
- Change: make screenshots 1 and 2 one continuous image across both frames.
- Hypothesis: the continuous image pulls people into swiping to screenshot 2.
5. Social proof up front
- Change: add a real award, rating or download milestone to the first screenshot.
- Hypothesis: trust is what stops a download, not features.
- Note: only use claims you can back up, or App Review may reject them.
Captions #
6. Shorter captions
- Change: cut every headline to four words or fewer.
- Hypothesis: a headline is read in about two seconds on a small thumbnail, so fewer words land better.
7. Numbers in the headline
- Change: "Save time on invoices" becomes "Send an invoice in 30 seconds".
- Hypothesis: a specific number is more believable than a vague claim.
8. Question vs statement
- Change: "Can't sleep?" instead of "Sleep better tonight".
- Hypothesis: a question names the reader's problem and makes them stop.
9. Bigger type
- Change: increase caption size so it is readable in the search result thumbnail.
- Hypothesis: captions nobody can read don't sell anything.
10. Captions above vs below the device
- Change: move the caption from above the device to below it.
- Hypothesis: the first thing people see should be the app, not text. (Or the reverse. Test it.)
Visual Style #
11. Light vs dark screenshots
- Change: show the app in dark mode, or put light UI on a dark background.
- Hypothesis: standing out from the competing apps around you in search results matters more than the style itself.
12. Brand color vs contrast color
- Change: swap your brand-color background for the color that contrasts most with your category's usual look.
13. Gradient vs flat background
- Change: a solid color instead of a gradient, or the reverse. Our templates have both styles to start from.
14. Device frame vs no frame
- Change: remove the device frame and let the UI fill the screenshot.
- Hypothesis: the frame takes space the UI could use.
15. Tilted or 3D device
- Change: angle the device or give it depth instead of a flat front view.
- Hypothesis: a device that looks different catches the eye in a row of flat ones.
16. Zoom in on the key UI
- Change: magnify the one element that matters (a chart, a button, a result) instead of showing the whole screen.
- Hypothesis: the whole screen shrunk to thumbnail size shows nothing.
Order and Story #
17. Reverse the order
- Change: screenshot 5 goes first.
- Hypothesis: the feature you added last may be what people want most now.
18. Problem → solution story
- Change: first screenshot states the problem, the rest show how the app solves it.
19. One feature per screenshot vs overview first
- Change: start with a collage of the whole app, then go into single features.
20. Fewer screenshots
- Change: cut from 10 screenshots to 5, keeping only the strongest.
- Hypothesis: weak screenshots at the end make people doubt the strong ones.
- Win: conversion lift even though there's less to see.
Audience and Market #
21. Speak to a different user
- Change: rewrite captions for students instead of professionals, or the reverse.
- Hypothesis: one group converts much better than the other.
22. Localized visuals, not just localized text
- Change: for one market, use local content in the UI (names, currency, places).
- Hypothesis: translated text on American example data still looks foreign. See A/B testing localized screenshots.
23. Seasonal variant
- Change: a holiday or back-to-school version of the first screenshot.
- Note: results from a seasonal test don't tell you much about the rest of the year.
Beyond Screenshots #
24. App preview video vs no video
- Change: add or remove the app preview.
- Hypothesis: a weak video in first position hides your best screenshot.
- Our app preview specs guide covers the formats.
25. Alternate app icon
- Change: a different icon color or symbol.
- Note: Apple requires every icon you test to be in your app binary, so this needs an app update first.
How to Run These Quickly #
The slow part of any of these ideas is making the variant: every language, every device size, then uploading it into the test. In Screenshot Studio, a caption, background or order change applies to every language and size at once. Then upload it straight into a Product Page Optimization variant without dragging files into App Store Connect.
Before you start, read the Product Page Optimization mistakes that waste a test, and check the duration with the A/B test calculator.