ASO Keyword Research

Find keywords the app can realistically win, backed by aso data: never gut feel.

Preflight: see the aso router skill (CLI installed, session status). Popularity needs a session; difficulty/rank/opportunity work without one.

Iron law

Never shortlist a keyword without running it through aso keywords. Popularity and difficulty come from the CLI, not intuition.

Step 1: Seed keywords

Brainstorm 10-20 seeds across categories, not just the obvious core term:

Step 2: Expand each seed

aso suggest "<seed>" --limit 50 -c US

Returns Apple Ads recommendations (with popularity) plus App Store autocomplete ideas. Run this for your top 3-5 seeds; pool and dedupe the results.

Step 3: Score in batches

aso keywords <term1> <term2> ... <termN> --app <appId> -c US --min-popularity 10 --max-difficulty 60

Max 100 terms per call: chunk the pooled candidate list into batches of ≤100. Passing --app also returns your current rank per keyword, so you see baseline position, not just opportunity. Popularity is cached 24h; add --fresh if you need a live re-pull (e.g. right after a metadata change).

--min-popularity <n> and --max-difficulty <n> pre-filter the batch; terms cut this way show up under the response's filteredOut with a reason, skim them once in case a relevant term got cut for the wrong reason.

Each result gives popularity (5-100, 5 = Apple's floor), difficulty (0-100, the ASOManiac model, see aso for what it weighs), opportunity = popularity × (100 − difficulty) / 100, appCount, topApps, your rank, brand (true = someone else's brand term, exclude it), and confidence (high/medium/low, trust in that term's numbers). Drop every brand: true term from the candidate pool immediately, it isn't a keyword opportunity.

Step 4: Shortlist (relevance first, then opportunity)

1. Cut anything irrelevant to the app, no matter how good the score: an off-topic keyword that ranks well converts nobody and can hurt review signals.
2. Within the relevant set, sort by opportunity descending.
3. Watch the popularity/difficulty shape, not just the composite score:

Long-tail strategy for new / small apps

An app with few ratings can't out-authority a competitor with 50k+ reviews on a difficulty 80 keyword, no matter the popularity. Calibrate the difficulty ceiling to the app's own ratings count:

Bias new/small apps toward more, narrower keywords (3+ word phrases) over fewer broad ones: you win by being unambiguously the best answer to a specific query before you can compete on a generic one.

Keyword tiers

Tag every shortlisted term with one tier, don't just leave a flat opportunity-sorted list:

Step 5: Output

Ranked table, sorted by opportunity, relevant-only:

| # | Keyword          | Pop | Diff | Opportunity | Your rank | Tier |
|---|-------------------|-----|------|-------------|-----------|------|
| 1 | baby cam          | 44  | 18   | 36.1        | -         | Primary |
| 2 | nanny cam         | 35  | 15   | 29.8        | 34        | Secondary |
| 3 | discreet camera   | 5*  | 22   | 3.9         | -         | Long-tail |

(* = popularity floor, treat as "no signal" not "dead")

Close with:

1. Top 10-15 by opportunity: the candidate pool.
2. Recommended token set: which terms are strong enough for title (1-2), subtitle (2-4), and keywords field (remaining, packed toward 100 chars), token-level, not a finished draft. Hand off to aso-metadata to actually assemble and lint the fields.
3. Watchlist: relevant but currently too competitive for this app's authority, revisit after ratings grow.