App Store Optimization for iOS & Android

Store listing work starts with a number that is easy to misread. Your console shows impressions, product page views, and installs, with a conversion rate between the last two. Ten thousand impressions and 250 installs looks weaker than two thousand impressions and 200 installs, and the second app is usually in better shape, because the people who saw it wanted what it offered.

App Store Optimization is the work of getting the right people to your listing and making the decision easy once they arrive. That means the name and subtitle read in search results, the keyword fields that decide whether you appear at all, the first screenshots, the icon, the review prompts, and the words you choose for each country you sell in.

There is a ceiling on this work. A listing cannot make people keep an app they do not want. When retention is poor, better copy brings more people to the same disappointment, and the rating falls faster than it would otherwise.

Who it is for

This work suits an app that already has something worth installing and a store page that undersells it: a small team buried under larger competitors in search, a product expanding from one country into several, or an older app whose rating no longer reflects the current build. It is a poor fit for an app with no reason for users to return. When testers abandon the first screen, that is product and UX work before it is store copy.

What store listing work covers

The App Store and Google Play rank and display apps differently, so a product needs two versions of the same story. On the App Store you have a title, a subtitle, and a keyword field of 100 characters that users never see, while the visible description is not indexed. On Google Play the long description is indexed and the short description shows in the listing. Keyword research therefore differs per store, and terms belong in the fields that count.

Visual assets decide whether that interest becomes an install. The first two screenshots carry the value proposition, one idea per screen, legible at thumbnail size. The App Store autoplays preview video muted. The icon is where product and marketing overlap, and Calculator Lock shows why the obvious design is not always available: that app hides photos behind a calculator icon, so the icon must look like an ordinary utility rather than a branded privacy product.

Localisation is the least expensive growth available to an app that ships in English. Search behaviour differs by market and language, so translating keywords is not the same as translating a listing, and screenshot captions often need rewriting rather than lengthening.

Review prompts shape conversion more than most teams expect: asking after the user completes the action the app exists for behaves differently from asking on launch. Both stores control the prompt through their own APIs and rate limits, and store policy prohibits buying or incentivising reviews. Our UI and UX design team produces the store assets and the screens the prompt appears on.

How we read store analytics without fooling ourselves

Install count is the number everyone reports and the one that explains the least. A spike can come from a store feature, a seasonal shift, or a competitor outage, and none tells you whether the listing works.

We read the chain instead:

• Impressions to product page views, which shows whether the icon, title, and first screenshot earn a tap.
• Product page views to installs, the conversion rate copy and screenshots control most directly.
• Install to first meaningful session, then retention at day seven and day thirty.
• Crash free session rate and the rating trend, read next to the review text.

Two store changes in the same week turn attribution into guesswork, so we change one variable at a time where traffic supports it. Stores index metadata within a day or two, but conversion data needs a couple of weeks before a shift means anything. At a few hundred impressions a day, ordinary variation is larger than the effect of most listing edits.

Where the project starts

A plain idea has no app and no listing. Naming, category, and keyword research happen early, because the name has to be searchable, available, and free of trademark conflict, and the category decides which competitors you rank against.

An existing conventional codebase is a live app with reviews and console history. We audit the current title, keywords, screenshots, and reviews, find the markets where conversion is strongest, and improve from evidence.

An AI or vibe-coded prototype is an app assembled quickly, sometimes already published. The listing usually reads like a feature dump, the screenshots are device captures with no captions, and the assets sit at the wrong dimensions. The idea stays; we rebuild the presentation layer and align the store fields with what the app does.

What we deliver

The deliverable is an optimised listing plus the research behind it, so future changes rest on numbers. That covers keyword research per store and language, rewritten title, subtitle, and description fields, screenshots and a preview video script, icon direction, localised metadata, a review prompt plan that respects platform rules, and a measurement sheet for the conversion chain. Event and crash instrumentation comes from our mobile app analytics and crash reporting work.

How the engagement runs

1. Audit. We pull the console data and read the current listing on both stores.
2. Research. Keyword sets per store and market, competitor listings in the same category, and the review text that shows what users expected.
3. Creative. Icon direction, a screenshot sequence with captions, and preview video.
4. Publish. Metadata and assets in the sizes each store requires, submitted through our App Store and Google Play publishing workflow.
5. Measure. We track the chain for two to four weeks, then decide what to change next.

What it costs

Listing work is billed at our flat rate of $39 per hour, or through a monthly capacity plan from $2,699 when there is steady product and store work to run. The pricing page explains both models.

A focused refresh of one app on both stores is a short piece of work. Localisation multiplies the effort by the number of languages you support, and each locale needs its own keyword research rather than a copy of the English set. When installs and retention disagree, the problem sits in the product rather than the listing, and the fix is closer to the mobile MVP typical order of $4,800 to $8,500 across four to six weeks. We say so instead of selling listing work that cannot help.

Proof

Voice VPN is our own Flutter VPN product on the Play Market with a zero-account model and a credit economy built on ads and USDT referrals. Its listing has to explain a privacy tool with no sign-up and no subscription in the space a search result gives it.

Easy VPN Free is an open-source Android VPN app with no accounts, no logs, and no monetisation, published on GitHub. It became a reference point for Android VPN development, which shows what a clear description does for a free tool.

Calculator Lock is a Play Store privacy app whose icon must look like a plain calculator for the feature to function, and whose screenshots explain something a device capture cannot. It proves we design store assets around how the product works, including cases where standard marketing would break it.

Related services

• Mobile apps development for the full range of iOS, Android, and Flutter work we do
• App Store and Google Play publishing for submissions, review handling, and asset requirements
• Mobile app analytics and crash reporting to instrument retention, events, and stability properly
• UI and UX design for icon, screenshot, and in-app design work
• API development and integration when the app needs backend work before the value proposition holds up

Next step

Send us your store links, the console numbers you have, and the markets you care about. We will say which parts of the listing are holding the app back, and which are a product problem the listing cannot fix.

Contact us with your app details, or put your listing into the Vasilkoff.info estimator for a first scope and cost range.