How to Do App Store Keyword Research: A 4-Phase Workflow

The workflow overview

App Store keyword research is most effective when it follows a consistent, repeatable process rather than an ad hoc collection of hunches. A good workflow has four phases: Ideate (build a long list of candidates), Score (evaluate each candidate for popularity and difficulty), Select (choose the best terms for the space available), and Update (write the metadata and submit).

Each phase has different outputs and different tools. Ideation is a creative, expansive phase — you want quantity. Scoring is analytical — you apply objective criteria. Selection is a constrained optimization — you're fitting the best terms into 100 + 30 + 30 characters. Update is an execution step with a defined output: a metadata submission.

Running the full workflow before every metadata update (roughly every 4-8 weeks for an actively managed app) is more effective than occasional large overhauls. Small, data-driven changes accumulate to large ranking improvements over time.

Phase 1 — Ideate: Build your candidate list

Start with your app's core value proposition stated in 10 different ways. What problem does it solve? What category does it belong to? What are users trying to accomplish? Each framing generates different keyword candidates.

Then work through structured sources: your top competitors' names and subtitles (copy the exact words they've publicly chosen), App Store search suggestions for your 3-5 core terms, user review language (what phrases do satisfied users use to describe what the app did for them?), and your own feature vocabulary (every feature name is a potential keyword).

Add terms from Apple Search Ads keyword suggestions if you have access — when starting a campaign, Apple suggests related terms that reveal adjacent search demand. You don't need to run the campaign to harvest the list.

Target 80-150 candidates. More is fine; the scoring pass will cull them efficiently.

Phase 2 — Score: Get popularity and difficulty for every candidate

Run your candidate list through a keyword scoring tool. For each term in your primary storefront, you need: popularity (0–100, how much search demand exists), difficulty (0–100, how competitive the current ranking set is), and competing apps count (how many apps appear in results for this term).

Score all candidates in your primary storefront first. Then for your top 20-30 candidates, check scores in 2-3 secondary storefronts — typically your next-largest markets. Secondary storefront scores often reveal that a term blocked in the US is accessible in the UK, Canada, or Australia.

As you score, mark each term with a simple tier: Green (popularity 30+, difficulty 55 or below), Yellow (borderline — either popularity is lower or difficulty is higher, worth considering), Red (popularity below 25 or difficulty above 70). The scoring pass typically reduces 100+ candidates to 20-30 viable terms.

Phase 3 — Select: Fit the best terms into your metadata

Selection starts with a constraint: you have 100 characters for the keyword field, 30 for the subtitle, and 30 for the name (most of which is taken by the brand name). Work backwards from the highest-weight fields.

Name: Can you work a high-value keyword into your app name without damaging the brand? Many successful apps use the format 'App Name: Category Keyword'. The name's 30 characters are the most valuable in ASO — a keyword here outranks the same keyword buried in the keyword field.

Subtitle: Choose a second-tier keyword or a phrase combining two related terms. The subtitle is both algorithm-visible and user-visible (it shows in search results), so it needs to work for both readers.

Keyword field: Enter your remaining selected terms as comma-separated words. Each word is indexed individually — you don't need to enter full phrases. 'budget,planner,weekly' covers 'budget planner', 'weekly planner', and 'budget weekly' without wasting characters on the spaces.

Check for cannibalization: any word already in your name or subtitle doesn't need to be in the keyword field. Remove duplicates and use the freed characters for additional terms.

Phase 4 — Update: Submit and track

Write the final metadata, review it once for readability (the name and subtitle especially — they're visible to users), and submit with your next app update. The metadata update can go in an otherwise minor update; it doesn't need to be bundled with a major feature release.

After submission, track results: App Store Connect's app analytics shows search impressions and installs by search term with a lag of a few days. Note which terms are generating impressions and which are converting to installs. Terms that generate impressions but few installs may indicate a conversion problem (screenshots, description), not a ranking problem.

Plan your next iteration: the workflow repeats. Terms that underperformed get replaced in the next update; terms that generated strong results get supplemented with related low-competition terms. The cumulative effect of iterative, data-driven updates compounds over time.