App Store Optimisation Checklist: 18 Checks Before You Ship

Almost every team we talk to has some version of the same story. The app ships, the listing goes up with a title somebody wrote in five minutes, a few weeks pass, and installs land under plan. Nobody can say exactly why. Someone suggests changing the icon. Someone else suggests running ads. The listing itself — the thing that decides whether you get found at all — usually gets looked at last.

The short version

Most underperforming listings aren't a mystery. They're a handful of metadata and creative decisions that were made quickly and never revisited. Here's the thing: those decisions sit in front of everything else you do. Paid traffic, press, word of mouth — all of it eventually lands people on a store page, and that page is what converts them.

In our experience, the order most teams get backwards is this: they look at creative first because it's visible, and metadata last because it feels like admin work. Metadata is cheaper to change and far easier to measure. Start there.

Start with the fields the stores actually index

On iOS, the title, subtitle and keyword field are indexed. On Google Play there's no keyword field at all — the title, short description and full description carry the load. That difference alone explains a lot of cross-platform confusion we see in audits.

Things worth checking:

• Are your highest-intent keywords in the title and subtitle, or buried in the keyword field where they carry less weight?
• Is the keyword field wasting space repeating words already in the title? Comma-separated, no spaces, and no duplicates — that's the whole trick.
• Are you stuffing singular, plural and misspelled variants of the same root? It makes the field longer without making you reachable on more queries.

We'd rather see eight well-chosen terms than thirty padded ones. Padding reads as padding to a reviewer and does nothing for a search query.

Creatives: what people judge before they read a word

Most users never swipe past the second screenshot. That's not a design opinion, it's just where the drop-off is. So the reason to install has to land in frame one, and the caption has to be readable on the smallest device you support.

Two things we end up flagging over and over:

• Screenshot captions that were checked on a desktop mock and are unreadable on a real phone.
• Preview videos that only make sense with sound on. Autoplay behaviour varies by store and region, and a large share of users will never turn audio on.

And shrink your icon to roughly notification size before you approve it. If the shape stops reading at that size, it won't help you in a search result either.

Localisation, where listings quietly break

This is the section that costs teams the most and gets the least attention. Translating English keywords word for word almost never matches what people in that market actually type. We've seen listings where the localised keyword set was technically correct and practically useless.

What I'd do instead: treat each locale as its own small research job. Pull store autocomplete suggestions in that language, look at what the local competitors rank for, and only then write the metadata.

Also translate the screenshot text, not just the store copy. Untranslated creative is the clearest signal that a listing was localised mechanically, and users notice it immediately.

Tracking: the part everyone agrees to and then skips

You cannot tell whether a metadata change helped unless you recorded where you stood before making it. Yet a surprising number of teams change their title on a Tuesday and ask us on Friday whether it worked.

The minimum that makes changes measurable:

• Rank per target keyword, captured on a fixed schedule, not ad hoc.
• Install sources separated so paid and organic don't blend into one number. Blended numbers make organic work look better or worse than it is.
• Conversion rate split by locale and traffic source. A global average hides the specific locales that are dragging.

One caution: a single rank reading 24 hours after a change is noise. Store rankings move for reasons that have nothing to do with your metadata. Give it a week, then compare.

Compliance, the boring section that saves you later

Overstated claims in a description drive refunds and one-star reviews. Both feed back into ranking. Screenshots that show an interface the app doesn't have will get flagged eventually, and a removal costs far more than the installs it bought.

Rating prompts are the other one. Fire them according to the current platform guidance rather than on launch. We know it's tempting to ask immediately — the guidance exists because immediate asks annoy people.

What to do next

• Run your live listing against the five sections above and write down every item that fails, in the order they appear.
• Fix the metadata items first. They're the cheapest to change and the easiest to attribute.
• Set a recurring review date tied to your release cadence, so the listing gets looked at after each update instead of once a year.