The ASO Keyword Research Framework: Find Your 50 Priority Terms
Most app teams approach ASO keywords the wrong way. They open an ASO tool, sort by search volume, and pour the top hits into the title and subtitle. Six weeks later the rankings haven't moved, the indexed terms don't match the intent of installers, and the team blames the algorithm. The problem was upstream: there was never a framework for choosing which keywords to chase.
This article is the keyword research framework Semnexus uses on every ASO engagement. It produces a short-list of 50 priority terms — 30 for the App Store, 20 for Google Play — that earn placement in metadata, screenshots, and reviews. The framework runs in four phases over about a week.
Why 50 and not 500
Most ASO tools surface several thousand candidate keywords per app. The temptation is to maintain a giant tracked list and try to optimize for everything. That is not what the algorithms reward.
- The App Store gives you 30 characters in the title, 30 in the subtitle, and 100 in the keyword field. That is roughly 17 to 22 indexable terms.
- Google Play indexes the entire long description, so the surface is bigger, but Play favors repetition and contextual density. Beyond 25 to 35 prioritized terms, density drops below the level Play rewards.
- A focused list of 50 is enough to fill both stores' indexable surfaces with room for seasonal swaps.
The first job of the framework is to cut the list down, not grow it.
Phase 1: Seed the universe (Day 1)
Build the raw candidate list from four sources:
- Your own metadata and reviews. Pull the title, subtitle, current keyword field, recent reviews, and support tickets. Extract every noun phrase that describes what users say the app does. This is the most underused source of high-converting terms.
- The top 10 ranked competitors in your category. For each, capture their title, subtitle, and Play long description. Note any term that appears in two or more competitors.
- App Store and Play autosuggest. Type your category seed terms (for example, "budget app" or "language learning") and capture every autosuggest completion. Repeat for the top 5–10 root terms.
- One ASO tool, used for completion only. Run your app through one tool (AppTweak, ASOMobile, or Sensor Tower work). Pull the recommended keywords filtered to relevance score 50 and above. Do not use the volume sort yet.
A correctly seeded universe is 400–800 terms. If you have fewer than 200, you have under-sourced. If you have more than 1,000, you have padded with weak relevance and you will waste time scoring noise.
Phase 2: Score on three axes (Days 2–3)
Every candidate gets three numeric scores from 1 to 5. The product of the three is the priority score.
Relevance (1–5). Would a user searching this term reasonably want your app? Be strict. "Budget" is a 5 for a budgeting app. "Money" is a 3 because too many adjacent apps satisfy it. "Finance" is a 2 because the intent is too broad.
Difficulty (1–5, inverted). How hard is it to rank in the top 10? Use the difficulty score from your ASO tool, but invert it so that 5 means easy and 1 means impossible. For most early apps, anything with difficulty above 60 (a 2 or 1) is off-list until the app has more reviews and download velocity.
Volume (1–5). Search volume relative to your category. Use your tool's volume score. A 5 is one of the top 20 volume terms in your category; a 1 is fewer than 100 monthly searches.
The priority score is Relevance × Difficulty × Volume. The maximum is 125. In practice, terms scoring 36 and above are the ones worth chasing.
Phase 3: Map terms to surface (Day 4)
Sort by priority score and assign each top-scoring term to a surface. The App Store and Play have different ranking surfaces, and a term lives where it matters most.
App Store surface map:
| Surface | Character budget | Best for |
|---|---|---|
| App title | 30 chars | One head term, brand-adjacent |
| Subtitle | 30 chars | 2–3 high-intent terms that explain the value |
| Keyword field | 100 chars | Long tail; do not repeat title or subtitle |
| In-app purchase names | 30 chars each | Secondary head terms tied to features |
| Promotional text | 170 chars | Seasonal terms, not indexed but lifts CTR |
Google Play surface map:
| Surface | Best for |
|---|---|
| App title | 1 head term plus one feature |
| Short description | 80 chars; 2 head terms with intent verbs |
| Long description | Density of 8–12 priority terms, mentioned 3–5 times each in context |
| Developer name | One head term if naturally branded |
The exercise is allocation, not stuffing. Each term gets one primary surface. If a term cannot find a natural home, it is the wrong term for now.
Phase 4: Lock the 50 and ship (Day 5)
The final short-list has three buckets:
- Tier 1 (10 terms): head terms in title and subtitle. These are the terms you commit to. You do not change Tier 1 for at least 90 days.
- Tier 2 (25 terms): long-tail in keyword field and Play long description. These are the ones you can swap quarterly based on performance.
- Tier 3 (15 terms): branded plus seasonal. These live in promotional text, screenshots, and review prompts.
Ship the metadata update once. Resist the urge to keep tweaking. ASO algorithms need 14 to 28 days to recompute rankings against new metadata. If you change every week, you never get a clean signal.
What to track after ship
The right rankings dashboard is small:
- Tier 1 rank per term. Top 10 is the success threshold. Top 3 is the goal.
- Indexed-term count. Total terms where the app appears in the top 50. Should grow week over week.
- Product page conversion rate. New keywords should not depress conversion. If conversion drops, the keyword may be drawing the wrong intent.
- Branded vs unbranded install split. Branded should stay flat; unbranded should rise. If unbranded does not move within 30 days, the keyword choice is the suspect.
Frequently asked questions
Should I use the same 50 terms for App Store and Play? No. About 60% will overlap, but Play rewards repetition and longer phrases, while the App Store rewards distinct terms across fields. Build two short-lists from the same scored universe.
How often should I refresh the keyword set? Tier 1 every 90 days, Tier 2 every 30 days, Tier 3 monthly. Refreshing more often makes the ranking signal noisy.
Do I still need a paid ASO tool if I follow this framework? You need one for difficulty and volume scoring. Beyond that, the framework is the tool. Most teams over-spend on multiple ASO platforms and under-spend on time inside them.
What is the role of reviews in keyword ranking? Reviews are indexed on Google Play and weakly indexed on the App Store. Encouraging reviews that naturally use Tier 1 terms (through a contextual prompt after a value moment) gives a meaningful ranking lift on Play.
Where do app preview videos and screenshots fit? Not in keyword ranking, but in conversion. Once the keyword brings the user to the product page, the screenshot and preview decide whether the install happens. Both efforts have to be coordinated.
If you want a second set of eyes on your keyword scoring or you would rather hand the framework off, the ASO team at Semnexus runs this exact process as part of every paid-acquisition engagement. The AEO and content team handles the equivalent exercise for AI-search visibility once the app store side is locked.
