Transparent research methods

How the index works

App Market Index is an independent discovery product. It organizes public marketplace information and adds clearly labeled automated pattern analysis. It is not affiliated with or endorsed by Wix.com Ltd.

Independence and fairness

Integrity labels describe detected patterns, not intent and not proof of wrongdoing. A high rating alone is never evidence. Automatic results are published with a disclaimer, can be manually reviewed, and retain an aggregate explanation of the signals used.

Data sources and retention

The canonical catalog is stored in Wix CMS. Marketplace listing details, review summaries, company profiles, and verified taxonomy assignments are synchronized by replaceable server adapters with timeouts, bounded concurrency, validation, and retries. When a source fails, its data is marked stale and the last valid CMS value remains available.

Developer profiles store public company description, website, public support email, privacy policy, ratings, the raw company response, and synchronization timestamps. Billing emails, urgent emails, and phone numbers are not modeled. Review text is never persisted by the integrity system.

Integrity analysis: `integrity-v1`

Each scan considers up to 200 recent published reviews, deduplicates reviewers, checks a bounded set of reviewer histories, and maps reviewed apps to their stored developer IDs. The scanner saves aggregate counts and reason codes only.

A suspicious reviewer has reviewed at least two distinct apps from one developer, with at least 60% of evaluable history involving that developer.
A strong reviewer has reviewed at least three distinct apps from one developer, with at least 75% of evaluable history involving that developer.
Near-duplicate wording and concentrated timing may strengthen a score but cannot create a verdict alone.
Fewer than five evaluable reviewers or under 50% history coverage produces an insufficient-data result.

Verdict thresholds

Suspected self-posting (yellow) requires one strong reviewer or at least two suspicious reviewers spanning multiple developer apps.

Likely self-posting (red) requires at least three strong reviewers, or at least five suspicious reviewers representing at least 25% of evaluable reviewers, spanning at least three apps from the developer.

Other states are: unchecked, no self-posting signal, insufficient data, and scan error. A manual status override, when present, becomes the effective public status.

Limitations and corrections