Keyword research that actually moves installs

Keyword research is the foundation of ASO. Screenshots convert; keywords decide whether anyone sees the screenshots at all. Over half of installs still start with someone typing a phrase into the store search box.

The three numbers that matter

Every serious keyword shortlist balances the same trio:

• Relevance — does this phrase describe what your app actually does? Irrelevant traffic bounces and can hurt conversion signals.
• Volume / popularity — do people search it? On iOS, Apple exposes a relative Search Popularity score; tools map that into something comparable across terms. On Play, volume is inferred from public signals and store behavior.
• Difficulty — how hard is it to break into the top results, given who already ranks there?

The sweet spot for most indie apps is high relevance + meaningful volume + moderate difficulty. Chasing only mega-head terms (“game”, “photo”) is how small apps burn months. Ignoring volume entirely is how you “win” keywords nobody searches.

A simple workflow that scales

1. List the jobs your app does in plain user language.
2. Expand with autocomplete suggestions — the phrases the store itself completes as people type.
3. Pull live search results for each candidate and note who owns the top five.
4. Score volume and difficulty; keep a mix of mid-tail terms you can win in 30–90 days and a few ambitious head terms.
5. Ship metadata, then track rank weekly for the shortlist — not for 500 vanity keywords.

iOS and Google Play are not the same list

The same English phrase can be easy on one store and brutal on the other. Always research per platform. Title and subtitle weight differ. Play leans harder on description text and broader machine matching. Your keyword sheet should have two columns, not one.

What to pull from an API

For this workflow you repeatedly need: autocomplete suggestions, live search results, keyword popularity, and your (or a competitor’s) chart position. Knockplum exposes those as credit-metered endpoints so a script or agent can run the loop every Monday without copy-pasting from a UI. See /docs for request shapes, credit costs, and error formats.