App Store Optimisation That Compounds

App Store Optimisation gets treated as a launch checklist item. It is closer to a continuous experiment, and the teams that treat it that way compound an advantage over the ones who write a listing once.

What the stores actually index

They do not work the same way, and the difference changes how you write.

On the App Store, your name, subtitle and keyword field carry the indexing weight. The description does not contribute to keyword matching — it converts, it does not rank.

On Google Play, the title, short description and full description are all indexed, so the long description does double duty.

Writing one listing and pasting it into both wastes the strengths of each.

Keywords: specific beats broad

Competing for a single high-volume term is usually a losing position against incumbents with years of install velocity. Several specific terms that describe what the app actually does will deliver more installs and better-retaining ones.

Do not repeat words across the App Store name, subtitle and keyword field — repetition is wasted space, not emphasis. Drop spaces in the keyword field and use commas; the field is small and every character counts.

Your category and your competitors' listings are the cheapest keyword research available.

Screenshots do the converting

Most people decide from the first two screenshots without scrolling. That is the real estate that matters.

Raw screenshots of your UI underperform. Captioned screenshots — a short line of value text above a framed screen — consistently do better, because the screenshot alone does not explain what the app is for.

Lead with the single most compelling thing the app does, not with onboarding or a settings screen.

Ratings are a ranking factor and a conversion factor

They affect placement, and they affect whether someone installs after seeing the listing. Both matter.

Ask for the rating at a moment of success — after the user completed something, not on launch. Asking immediately on first open is the most common mistake and produces exactly the ratings you would expect.

Reply to reviews, particularly negative ones. It is visible to future readers, and it occasionally converts a one-star into a revision.

Localisation is the underrated lever

Translating the listing — not the whole app — into the languages of your top markets is a small amount of work with disproportionate return. Each localised listing indexes in that language.

Machine translation is better than nothing and worse than a native speaker. For your top two or three markets, pay for the human.

Test one thing at a time

Both stores support experiments on listing assets. Run them properly: one variable, enough traffic to be meaningful, long enough to cover weekday and weekend behaviour.

Changing the icon, screenshots and subtitle together tells you something moved and nothing about why.

Measure installs that stay

Install volume is a vanity number if those users leave within a week. Track conversion from impression to install, and retention by acquisition keyword where your analytics allow it. A term that converts slightly worse but retains far better is the better term.