App store keyword research
App store keyword research is the process of finding the queries real users type into the App Store and Google Play, scoring them by volume, difficulty and relevance, then placing the winners in the indexed metadata fields. A workable process has five steps: build a seed list from your own features, competitors and store autocomplete; expand it with long-tail variants; score each term for search volume, difficulty and how well your app actually satisfies the query; select a target set that fits the character budget of each store; and place terms so no word is duplicated across fields. Re-run it every release, because store demand shifts far faster than web search demand.
Key takeaways
- Relevance beats volume: ranking third for a term you genuinely satisfy converts better than ranking tenth for a bigger one.
- Store autocomplete is the highest-signal free data source — it reflects real in-store demand, not web search.
- Score every candidate on volume, difficulty and satisfaction; a term you cannot fulfil will rank then bleed installs.
- Duplicate words across Apple's name, subtitle and keyword field are wasted characters, not reinforcement.
- Track ranked positions per term after each release, otherwise you cannot tell which placement change caused which move.
Step 1: build the seed list
- Your own features, written the way a user would describe them rather than the way your roadmap does.
- The problem language from your reviews and support tickets — users describe the job in words your marketing never uses.
- Competitor titles and subtitles, which are effectively their bet on the highest-value terms.
- Store autocomplete for each seed letter-by-letter; suggestions are ordered by real in-store demand.
- Category top charts, to see which terms the highest-converting apps in your space claim.
Step 2: score the candidates
| Score | What it measures | Rule of thumb |
|---|---|---|
| Volume | How many people search the term monthly in-store | Ignore anything below the store's floor — it will never move installs |
| Difficulty | How entrenched the incumbents are | New apps should target difficulty under 25 for their first wins |
| Satisfaction | How well your app answers the query | If a reviewer would call it a mismatch, drop it regardless of volume |
Step 3: place the winners
- Apple: strongest term in the app name, secondary cluster in the subtitle, the remainder as single comma-separated words in the 100-character keyword field. No word twice.
- Google Play: title carries the strongest term, short description repeats it once naturally, full description repeats each priority term three to five times across 4,000 characters.
- Keep a per-market placement sheet — the right term in Germany is rarely the direct translation of the right term in the US.
Step 4: measure and iterate
Record ranked position per target term, impressions, product page views and conversion rate before each metadata change. Without a baseline you cannot attribute a move to a placement decision, a creative change or a seasonal swing.
Give a metadata change at least two weeks before judging it. Store indexes re-crawl on their own schedule and early movement is noise as often as signal.
Worked example: finding the term a listing was missing
A budgeting app had done keyword research once, at launch, from a competitor list. Two years later it ranked well for 'budget app' and 'expense tracker' and had never noticed that a third of its own support tickets used a phrase absent from every field of the listing: 'envelope budgeting'. The method that missed it was the common one — seed from competitors, score, place — which can only ever find terms someone else already targets.
Three sources fixed the blind spot, none of them a keyword tool. The app's own review text, its support tickets, and the search terms from its paid campaigns together produced eleven phrases the competitor-derived list did not contain. Two had meaningful volume and almost no competition, because the competitive set had made the same mistake.
'Envelope budgeting' went into the subtitle, displacing a generic benefit line, and the app ranked top three within a fortnight — a term with modest volume and unusually high intent, since nobody searches it casually. The broad head terms were left where they were: they were already won, and defending them costs less than the effort of moving from position four to three.
Research habits that produce weak lists
The recurring failures in keyword work:
- Seeding only from competitors, which guarantees you find nothing they have missed.
- Sorting by volume and ignoring difficulty and intent, then targeting terms you cannot win.
- Ignoring long-tail phrases because each looks small, when in aggregate they carry qualified traffic.
- Researching once at launch and treating the list as permanent while the category moves.
- Using web SEO volumes as a proxy for store demand, which behaves differently.
- Building one list for both stores, when Apple and Google index different fields in different ways.
Frequently asked questions
How many keywords should an app target?
Realistically 15 to 30 terms per storefront. Apple's fields hold roughly that many once duplicates and stop words are removed, and a focused set outranks a scattered one.
Where does app store keyword data come from?
Store autocomplete, your own Search Ads impression data, ranked-position tracking over time, and third-party panels that model volume. No source is exact; treat volume as a relative ranking, not an absolute number.
How often should keyword research be re-run?
Every release, and additionally whenever a major competitor changes its title or a seasonal demand spike is approaching.
appXL Research
App Store Optimization Research Team
The appXL research team analyzes App Store and Google Play ranking data across the apps our agent manages, and publishes what it finds.