What to know

App Store search visibility depends on more than one keyword field. Apple uses metadata and app information to help people find relevant apps, but the exact ranking system is not fully public. No tool can promise a fixed position.

Use a three-part model. First, visible metadata tells a person what the app does. Second, the keyword field adds relevant language within Apple’s current rules. Third, the product page and product experience determine whether the result is a good match. A ranking check can show a position, but it cannot explain every cause behind that position.

Match user intent, place important terms where they read naturally, and avoid wasting limited fields on duplicates. The name and subtitle are visible. The keyword field gives space for additional relevant terms within Apple limit.

Keep storefronts separate. Record the exact term, country, language, locale, date, field, and source signal. A phrase can mean something different in another market, and an observed position in one storefront does not prove the same behavior elsewhere.

Track changes by storefront and date. If a term brings views but not downloads, the page promise may not match the search intent. If a term is not observed, record the method and date before changing the whole field. Absence from one check is not proof that the app can never appear.

Use a field map before making an edit. For each phrase, write its user task, the field where it appears, the evidence source, and the product screen that satisfies it. This prevents the keyword field from becoming a dumping ground and makes it easier to remove a term when the product changes.

When a ranking check changes, compare like with like: same storefront, locale, query spelling, device or browser method, date window, and app version. Record neighboring results and the exact query, but avoid turning one observation into a claim about Apple’s private ranking system.

If indexing appears delayed or inconsistent, check the basics first: the app is available in that storefront, the metadata is actually submitted, the term matches the product, and the observation was recorded correctly. Give the team a clear issue log instead of repeatedly rewriting every field.

Use a controlled validation sequence after a metadata change: save the approved old and new fields, confirm the submitted version and storefront, wait for the observation window your team uses, then repeat the same query with the same method. Change one main field or phrase at a time when the goal is to learn what the edit changed.

Keep an indexing issue log with the query, locale, app version, field change, check method, date, observed result, and next action. Mark “not observed” separately from “not indexed” because a single check cannot establish how Apple processes the term or whether the result is simply outside the captured view.

Review the page promise before adding more terms. If the query is relevant but the screenshots, description, or onboarding do not satisfy its task, improving field coverage alone may create a poor match. Remove a phrase when the product no longer supports the intent, even if the phrase once appeared useful.

Before changing a field, capture a baseline row with storefront, locale, query, method, date, submitted version, observed result, and page promise. After the change, repeat the same check and compare the rows side by side. If the method, audience, or query changes, start a new comparison instead of merging the observations.

Use a decision rule that allows no change. If the term is relevant but the product evidence is weak, keep the current metadata and send the gap to product or listing review. Indexing evidence cannot make an unsupported promise safe to publish.

Reopen the issue when the app, locale, availability, field rules, or observation method changes. Preserve the baseline, new result, and explanation so “not observed” does not become an invented claim about Apple’s private indexing system.

For a shared-expenses app, a phrase such as “split bills” should be checked against the product flow, the visible metadata, and the screenshot that proves the task. If the build only records personal spending, the phrase may be readable but still be a poor indexing and product-fit decision.

ASOgenic can help organize the evidence, compare field options, and preserve a review record. It should not invent indexing behavior or promise a position. A person should confirm the term matches the current app before the listing changes.