App Store keyword research

Most guides tell you to find keywords with high volume and low competition. Apple publishes neither number. Here is what the App Store actually gives you, and how to read it into a decision you can act on.

What App Store keyword research actually is

App Store keyword research is the work of deciding which search phrases your app should try to rank for, using evidence from the store itself rather than instinct. It sits underneath every other App Store Optimization decision: your title, your subtitle and your 100-character keyword field are finite, and every word you spend on one phrase is a word you cannot spend on another.

It is not the same job as SEO for the web, and the differences matter more than the similarities. Web pages compete on hundreds of ranking signals against billions of documents. An App Store result page has a few million apps behind it, a much smaller set of signals, and a metadata budget you could write on a napkin. That makes the problem more tractable — and it makes wasting your budget more expensive, because you have so little of it.

The other difference is where the data comes from. In web SEO you can buy a search volume figure that originates, however indirectly, from measured queries. Apple does not publish search volume at all. Every number you have ever seen in an ASO tool is inferred from something else, and knowing what it was inferred from is the difference between using it well and being misled by it.

Where the data actually comes from

There are only a few public sources, and every honest tool is built on the same ones. Knowing them lets you sanity-check anything a tool tells you.

Apple’s autocomplete

Start typing in the App Store search bar and Apple completes the phrase. This is the closest thing to a published demand signal that exists: Apple has no reason to suggest a phrase nobody types. If you type a partial phrase and Apple completes it into something specific — “quit vaping” into “quit vaping tracker” — that longer phrase is being searched by real people. If Apple suggests nothing at all for your phrase, you are guessing, and you should treat every other number about it with suspicion.

Autocomplete is also ordered, and localized. The suggestions for the German store are different from the US store, and not translations of each other. This is one of the few genuinely free, genuinely reliable signals available, and it is badly underused.

The search results themselves

Search a phrase and the store tells you a great deal without being asked: how many apps it considers relevant, which ones it ranks first, how many ratings each has, when each was released, when each was last updated, and what each has chosen to put in its title. Almost every metric in an ASO tool — difficulty, competition, saturation — is a formula over exactly this.

Apple Search Ads popularity

If you have an Apple Search Ads account, Apple shows a popularity score from 5 to 100 for keywords in the ad tool. It is a real Apple number, and it is the closest thing to volume you can get. It is also logarithmic, relative rather than absolute, and available only for the terms Apple chooses to show you. Useful, but not the complete answer people assume it is.

Why ten tools give ten different numbers for the same keyword. Because they are all estimating from the sources above using different formulas, and none of them are measuring. Put “nicotine tracker” into several ASO tools and you will get several different popularity figures, none of which are wrong exactly — they are different guesses presented with equal confidence.

The practical response is not to find the one true tool. It is to ask what evidence sits behind any number before you act on it, and to treat a confident figure with nothing behind it as worse than no figure at all.

The five signals worth reading

A keyword is not one number. These five answer different questions, and a term can look excellent on one and hopeless on another.

SignalThe question it answersHow to read it
Demand evidenceDo people actually search this?Apple autocomplete is the strongest tell. A phrase the store suggests is one people type. No suggestion means you are hypothesising.
DifficultyHow strong is this result page overall?Driven by the rating counts and title optimization of the apps already ranking. Describes the page, not your chances.
RankabilityCould an app with no ratings get in?A different question entirely. A hard page with three soft slots is more winnable than an easy page held by ten mid-sized incumbents.
RelevanceCan your app honestly claim this phrase?If your app is not what the searcher wants, ranking for it produces installs that churn and reviews that hurt.
Intent matchDo the results answer the search?When people search something specific and the store returns something adjacent, that gap is an opening — and it is not the same thing as low competition.

The one that gets skipped is rankability, and it is the one a new app most needs. Difficulty and rankability are different measurements, and confusing them is the single most common reason a carefully researched keyword set produces nothing.

A method you can actually follow

This works by hand with a spreadsheet, and it is what a tool should be automating rather than replacing.

  1. Write down the problem, not the product. People search for the outcome they want, not the category you built in. A vaping app competes for “quit vaping”, “nicotine cravings” and “money saved quitting” before it competes for “habit tracker”. Start with five to ten plain-language phrases describing what your user is trying to do.
  2. Expand each one through autocomplete. Type each phrase, and partial versions of it, into the store search bar. Write down every suggestion. This is where the specific, high-intent long tails come from, and they are the terms a new app can actually win.
  3. Search each candidate and record what you see. For every phrase: how many results, the top ten apps, their rating counts, and whether any of them carry the exact phrase in the title. This is tedious and it is the part worth automating.
  4. Look past position ten. Check who sits at 11 through 30. Entering the top twenty is a realistic first goal; entering the top ten on launch day usually is not. A page with a soft tail is an entry route even when the head looks brutal.
  5. Sort by what you can win now, not by what you want. Split your list into what a new listing can take on metadata alone, what opens up once you have ratings, and the head terms worth naming as a destination. Spend your launch metadata entirely on the first group.
  6. Write the metadata backwards from the list. The strongest phrase goes in the title. The next cluster goes in the subtitle. Everything else becomes comma-separated single words in the keyword field, with nothing repeated.

The mistake that costs the most time

Almost everyone starts by targeting the biggest term they can find that has a tolerable difficulty score, and almost everyone loses months to it.

Worked examplehabit tracker
Difficulty
84
Demand evidence
Strong
New-app rankability
25

Enormous demand, and eight of the top ten apps carry the exact phrase in their title with tens of thousands of ratings behind them. A new app cannot enter this page with metadata. It is a real destination and a terrible launch target, and no amount of keyword-field cleverness changes that.

Worked examplenicotine craving tracker
Difficulty
54
Demand evidence
Suggested
New-app rankability
70

Lower demand, but Apple suggests the phrase, six of the top ten have fewer than 2,000 ratings, and no incumbent uses the exact phrase in a title. Nobody is defending it. This is what a launch target looks like — smaller, specific, and actually reachable.

The second app will be ranking and collecting installs while the first is still invisible. Nothing stops you targeting the head term later, from a position of having users.

Doing it country by country

The App Store is not one store, it is a set of storefronts with separate result pages, separate autocomplete, and separate competition. A phrase that is hopeless in the United States is regularly wide open in the United Kingdom, Canada, Australia, India or Brazil, and the search language rarely translates cleanly.

Two practical consequences. First, research each market you intend to publish in separately, rather than translating a US keyword list — the phrasing people actually type is a local question. Second, if you are a small developer, a smaller storefront is often the best place to get your first rankings, prove the concept, and accumulate the ratings that let you compete in a larger one.

Checking whether it worked

Ship the metadata, then capture your positions for every target phrase on a schedule and compare across time. One check tells you almost nothing, for a reason most guides skip: result pages move on their own. Some are held by the same apps for months; others shuffle every week. Until you know which kind you are looking at, you cannot tell a real gain from ordinary noise.

The same history answers a question worth more than your own rankings: what your competitors changed and what happened afterwards. A competitor rewriting their subtitle and climbing on a set of terms shortly after is a correlation, not proof — but it tells you which terms they are actively fighting for, which is intelligence you cannot get any other way.

If you would rather not build the spreadsheet, this is what GetRevuu does end to end: it expands your idea into a keyword universe from live storefront data, scores each phrase on all five signals, sorts them by when you can realistically win them, and captures the history that makes movement readable.

Common questions

How many keywords should I target for a new app?

Fewer than you think. Apple gives you a 30-character title, a 30-character subtitle and a 100-character keyword field, which is roughly 20 to 25 distinct words in total. A new app with no ratings realistically competes for a handful of specific phrases, not a category. Pick three to five you can genuinely win at launch, and keep the rest on a list for when you have traction.

Should I use single keywords or phrases in the keyword field?

Single words, comma separated, with no spaces after the commas. Apple combines words across your title, subtitle and keyword field to form phrases automatically, so writing "quit vaping tracker" as a phrase wastes characters on words the store would have combined for you anyway. Never repeat a word that already appears in your title or subtitle.

How long does it take to see ranking changes after updating keywords?

Usually a few days to a couple of weeks after the release is live, and the movement is rarely instant or clean. This is why a single before-and-after check is misleading: some result pages shuffle constantly on their own. You need a few captures spaced out over time before you can tell your change apart from the page just moving.

Does the app description affect App Store keyword rankings?

On the Apple App Store, no — the description is not indexed for search, which surprises people coming from Google Play, where it is. Your description still matters enormously for conversion once someone lands on the page, but it is not where your keywords should go.

Can I do App Store keyword research without a paid tool?

Yes, and it is worth doing by hand at least once so you understand what the numbers mean. Type partial phrases into the App Store search bar and note what Apple autocompletes, then look at who ranks for each phrase and how many ratings those apps have. It is slow and hard to keep consistent across many phrases and countries, which is the problem tools exist to solve, but the method is the same.

Stop guessing which keywords you can win.

GetRevuu scores every phrase against the live App Store and tells you whether a new app can realistically rank for it.

Start free