ASO & keyword research

Keyword Explorer helps you organize ASO research. Its scores are heuristics for prioritization—not Apple Search Ads popularity data and not a guarantee that a listing will rank.

Open Keyword Explorer.

Start with search intent, not a word list

Write the job a user expects the app to do, then collect seed phrases from four sources:

1. the problem in the user’s own language;
2. titles and subtitles of direct competitors;
3. repeated phrases in reviews;
4. category terms that describe the product, not just the audience.

For a sleep product, sleep, white noise, smart alarm, and sleep tracker describe different intents. Research them separately before combining variants.

Read the scores

Score	Use it for	Limitation
Difficulty	Relative estimate of how hard the current result set may be to compete with	It is heuristic, not an Apple-provided metric.
Popularity	Proprietary relative-demand score	Compare terms inside one locale; do not treat it as search volume.
Traffic	Proprietary prioritization score	It is directional, not an install forecast.
Opportunity	Composite opportunity score	A high score still needs product relevance and conversion potential.

The best keyword is not simply the highest opportunity score. It must accurately describe the app and lead to a result page where your listing can satisfy the intent better than current competitors.

Build a shortlist

1. Select the target country and language before scoring.
2. Enter one seed phrase and inspect related suggestions.
3. Remove phrases that are broad, ambiguous, trademark-dependent, or irrelevant to the core job.
4. Track 10–30 candidates, including a mix of head terms and specific long-tail phrases.
5. Set a target app so competitor-based recommendations have context.
6. Export the shortlist for your ASO experiment log.

For every retained keyword, record:

• the intended audience and problem;
• why the app is relevant to the phrase;
• current difficulty/opportunity and locale;
• the title, subtitle, or keyword-field change you plan to test;
• the observation date and expected conversion impact.

Use competitor recommendations carefully

Recommendations extract terms from the competitor context available to GetAppNiche. They are idea generation, not proof that a competitor receives installs from that query. Open the competitor, check the result context, and score the phrase in your target locale.

Locale research

Do not translate a successful US keyword list word-for-word. Research each storefront because language, intent, competitors, and category conventions change. A localized phrase can have lower competition but also a different meaning or weaker product fit.

Use Screenshot concepts & localization to carry the chosen intent into the visual listing, then measure the result in App Store Connect.

A simple prioritization rule

Ship a keyword test when all three are true:

1. Relevant: the app genuinely satisfies the intent.
2. Reachable: the current competitors and difficulty are plausible for your stage.
3. Valuable: ranking for the term would bring users likely to activate or pay.

If one is missing, keep the term as research—not as a metadata recommendation. For the underlying ASO concepts, read What is App Store Optimization?.

Worked shortlist example

Suppose a focused sleep-timer app is preparing its US English listing. The numbers below are illustrative; the decision logic is what matters.

Candidate	Intent fit	Difficulty	Opportunity	Decision
sleep	Too broad: meditation, tracking, sounds, and alarms	92	28	Research only
sleep timer	Exact core job	54	71	Test in subtitle
white noise timer	Relevant only if the feature exists	38	67	Test if product fit is true
smart alarm	Different job from stopping audio	61	60	Reject

The winning term is not necessarily the one with the lowest difficulty or highest opportunity. Here, sleep timer wins because its intent matches the product and the listing can make that promise honestly. Record the pre-test scores, metadata change, launch date, and App Store Connect conversion result so the next decision uses evidence rather than memory.