App keyword research splits into two jobs that people constantly conflate. One is app store optimization — picking the terms in your title, subtitle, and keyword field that surface you inside the App Store and Google Play. The other is web search — ranking pages on Google for the queries people type before they ever open a store. They use different data, different tools, and different tactics, and most install growth eventually comes from the second one.
Here is how to do both properly, and where the honest limits of each data source sit.
The two search surfaces and why both matter
Inside the stores, search is the dominant discovery channel — roughly two-thirds of installs start with a store search. Those queries are short, generic, and heavily branded: “budget app”, “photo editor”, “notion”. Competition is winner-take-most, because the top three results collect the overwhelming majority of taps on a small screen.
On the web, the query set is far richer. People search “how to split rent with roommates”, “best free invoice app for freelancers”, “convert heic to jpg on iphone”. Those queries have intent you can answer with a page, and they feed installs through a landing page or a comparison article. They also compound — an article ranking for a long-tail term keeps delivering installs for years, whereas store ranking evaporates the moment a competitor outranks you.
Where store keyword data comes from
Neither Apple nor Google publishes app search volume the way Google publishes web search data. What ASO tools report is inferred, and you should treat every number as directional.
- Apple Search Ads — the closest thing to ground truth on iOS. Its popularity scores and impression forecasts come from Apple’s own auction data. Even a tiny test campaign buys you real numbers.
- Store autocomplete — free and genuinely useful. Type a seed and record every suggestion; the order reflects popularity. Do it on a clean device, since suggestions personalize.
- Google Play Console — its acquisition report shows the actual search terms that led to your installs. Underused, and the only first-party Play data you get.
- Third-party ASO platforms — Sensor Tower, AppTweak, and similar. They model volume from panels and ranking observations. Good for competitive comparison, unreliable as absolute numbers.
How to build a store keyword set
Work through this in order. It takes an afternoon and you should redo it every quarter.
- Seed from the problem, not the product. Write ten phrases a user would type if they did not know your app existed. “Track spending”, not “personal finance management platform”.
- Mine competitors. List the top ten apps in your category and record every word in their titles and subtitles. Repeated words across several competitors are the category’s high-volume core.
- Expand with autocomplete. Run each seed through store search suggestions and collect everything. This is where the long tail lives.
- Score by relevance first, volume second. Ranking first for a term nobody searches is worthless, but ranking twelfth for a huge term is worse. Target terms where you can plausibly reach the top three.
- Allocate the real estate. On iOS: title carries the most weight, then subtitle, then the 100-character keyword field. Never repeat a word across those fields — you waste characters. On Play: the title, short description, and long description are all indexed, so natural repetition across the description helps.
Then wait. Store ranking changes take one to three weeks to settle after a metadata update, and install velocity and retention influence position as much as keyword placement does. Change one variable per release or you will never know what worked.
Web keyword research for app companies
This is where standard SEO tooling applies directly, and where app keyword research becomes ordinary keyword research with an install-focused funnel. You are hunting three query families:
- Category terms — “best expense tracker app”. High intent, high competition, usually dominated by listicles from review sites. You compete by being on those lists as much as by ranking yourself.
- Alternative and comparison terms — “[competitor] alternative”, “[competitor] vs [competitor]”. Small volume, exceptional conversion. Cheapest installs you will ever earn.
- Problem and how-to terms — the largest pool by far, and the one most app teams ignore. Someone searching how to solve the problem your app solves is a warm install if the article ends with an honest recommendation.
Build this the same way you would for any site: multi-seed keyword exploration, filter by difficulty against the weakest page-one competitor rather than the average, and save the survivors into a project keyword pool you can write against. SEO Rocket returns up to 150 ideas per search with volume, difficulty, CPC, and SERP features, and the export is free after one query — enough to build the full topic map for a category in a single session.
Do not skip the SERP check
Before committing to a term, look at what is actually ranking. Two patterns kill app-company articles.
The first is app store pages ranking in Google. For branded and near-branded queries, the store listings themselves often occupy the top slots, and you will not outrank Apple’s own domain. Optimize the listing instead and move on.
The second is entrenched review sites. If positions one through ten are all affiliate listicles with strong domains, your best play is outreach to get included, not a competing listicle from a brand with an obvious conflict of interest. Readers discount vendor-written “best of” lists, and so does Google.
Measuring what the research produced
Attribution between web search and installs is genuinely hard, and anyone who tells you otherwise is selling something. The practical setup:
Use Search Console for query-level truth on the web side — impressions, clicks, and average position for every page. Pair it with rank tracking so you see position trends across the whole keyword set rather than only terms that already get clicks. Remember that daily movement of two or three positions is noise; judge on four-week trends. On the store side, watch category rank and the Play Console search-term report, and treat third-party ASO volume estimates as comparative rather than absolute.
Then connect the two with a landing page per campaign and store attribution parameters, so a click from an article is traceable to an install rather than disappearing into “organic store search”.
A sequence that works for a new app
Month one: fix the store listing. Title, subtitle, keyword field, screenshots. This is the highest-leverage hour of work available to you and it is free.
Month two: build the web keyword pool and publish comparison and alternative pages. Small volume, fastest returns, and they rank quickly because almost nobody writes them well.
Months three onward: work the problem and how-to tail, one article per target cluster, each ending in a genuine recommendation rather than a hard pitch. Track everything, expect the first meaningful movement around week eight, and resist changing your store metadata every time a position wobbles.
