· 5 min read

Localized App Store Screenshots: An ASO Lever That Works

Why localized App Store screenshots moved Mapika's listing more than text metadata, and a simple caption-first method to build them for five languages.

Most ASO advice starts with keywords. You research them, pack the title and subtitle, and fill the 100-character keyword field. That work matters.

But when I localized Mapika's App Store listing into English, Russian, Hebrew, Ukrainian and German, the change that mattered most was not the text. It was the localized App Store screenshots. This post explains why screenshots carry so much weight, and a caption-first method for building them in five languages without losing your mind.

Why screenshots beat text metadata

Think about how people actually meet your app. They search, and they get a list. Apple's product page guide says it plainly: when there is no app preview video, the first one to three screenshots appear right in the search results.

So the screenshots are not "part of the product page". For many users they are the product page. Your keywords decide whether you show up in the list. Your screenshots decide whether anyone taps.

Text metadata still matters. But the description sits behind a "more" link, and few people read it. A screenshot caption is read in a second, by everyone who scrolls past.

There is a second reason. Keywords are a ranking problem, and ranking moves slowly. Screenshots are a conversion problem: every user who sees them reacts right away. In my case, screenshots turned out to be the stronger lever.

Caption first, image second

The usual screenshot is a nice frame, a catchy slogan and a screen from the app. That works in one language. Slogans rarely survive translation, and a random screen does not explain itself.

So flip the order. Write the captions first, as a plain list, and only then pick the screen that proves each caption.

The rules:

The caption list becomes the source of truth. It looks like a small table (example captions):

id   en                               de                              ...
s1   Mark every place you've visited  Markiere jeden besuchten Ort    ...
s2   Share your travel map            Teile deine Reisekarte          ...
s3   Earn badges as you explore       ...

Each row is one screenshot. Each column is one language. Translators (or you, with careful review) work on short lines with clear context, not on a design file.

Localize the screen, not just the caption

A translated caption on top of an English screen looks lazy, and users notice. If the app is localized, the screenshot should show the localized app.

For a Flutter app this is cheap. The app already supports each locale, so you run it in that locale and capture the same screens. The map, the labels and the buttons all match the caption language.

A few details that are easy to miss:

Keep the work small

Five languages times several screenshots quickly becomes a lot of images. Two things keep it manageable.

First, let the store scale for you. App Store Connect accepts up to 10 screenshots per device size, and the screenshot specifications say that if your UI is consistent across devices, you only need the highest required resolution. App Store Connect scales it down for smaller iPhones. That means one set per language for iPhone, not one per device.

Second, generate the frames from the caption table, don't hand-edit them. Any tool that takes a background, a screen capture and a text string will do. When a caption changes, you change one cell in the table and regenerate. Hand-editing 30 or more images in a design tool is where mistakes creep in: an old caption in one language, a wrong screen in another.

Measure instead of guessing

Screenshots are the easiest part of the listing to A/B test, and both stores support it.

Google's own advice is to change one asset at a time, and it is good advice. If you swap the whole set and conversion moves, you don't know which screenshot did it. Test the first screenshot first, since it is seen by the most people.

Also give each test time. A small app needs a long run to collect enough data, and Play Console tells you when it needs more. Stopping early on a lucky week is how you ship a worse listing with confidence.

Takeaways