Are your reviews and star ratings attached to the right items for AI to read?
Confirms reviews and aggregate ratings are correctly attached to products, businesses, books, or movies.
What this signal tests
We check whether your reviews and aggregate ratings are emitted as structured data and attached to a supported parent type (Product, LocalBusiness, Book, Movie, Recipe, SoftwareApplication, etc.). For aggregate ratings, we look for ratingValue and either ratingCount or reviewCount. Individual reviews need reviewRating, an author, and reviewBody. The itemReviewed field must point at the supported parent.
Why it matters for your visibility in AI
Reviews are the dominant social-proof signal in AI shopping and local discovery. When an AI tool recommends a product, restaurant, or app, it heavily weights aggregateRating. Items without ratings often fail an implicit quality threshold and are filtered out. Items with ratings appear with their star values directly in AI summaries, which dramatically improves click-through and trust. Google's policy here is strict and worth understanding: reviews must be attached to specific items (products, businesses, books, movies), never to a generic Organization. Self-reviews of your own company are explicitly disallowed and result in markup being ignored or penalised. AggregateRating belongs on the thing being reviewed, not on the brand selling it. Sites that get this wrong often have their reviews silently dropped from AI presentations, even though the markup technically parses.
Pass criteria at a glance
| Criterion | Passes when |
|---|---|
| Valid Review or AggregateRating on supported parent. |
How we test it
We look for Review or AggregateRating blocks in your structured data and check three things: they are attached to a Schema.org type that Google permits reviews on (Product, LocalBusiness, Book, Movie, Recipe, SoftwareApplication, Event, Course, CreativeWork); the parent is identified via itemReviewed when the review is standalone; and the rating values are within the declared scale (default 1-5). Self-reviews on Organization or improperly-nested ratings fail.
Show technical detection method
Valid Review/AggregateRating with required fields attached to allowed parent type.
If your site fails: how to fix it
- Identify which item types on your site can validly carry reviews: products, local business locations, books, movies, software apps, recipes, events, or courses.
- On each such page, add aggregateRating with ratingValue (average), ratingCount (total ratings), and ideally reviewCount (subset of ratings that have a written review). bestRating defaults to 5; declare it explicitly to be safe.
- Where individual reviews are visible on the page, emit them as Review objects with author (a Person with name), datePublished, reviewBody, and reviewRating with ratingValue.
- Do NOT attach reviews or aggregateRating to a sitewide Organization block. This is explicitly disallowed by Google and causes the markup to be ignored.
- Pull ratings from your real review platform (your own review database, G2, Trustpilot, etc.). Inventing ratings is detected by cross-checking with third-party platforms and damages trust.
- Validate review-bearing URLs in Google's Rich Results Test under the relevant parent type's test (Product, Local Business, etc.).
Quick facts
| Maturity | ESTABLISHED |
|---|---|
| Weight | medium |
| Category | Structured Data |
Primary sources
Related signals
Frequently asked questions
Why can I not put aggregateRating on my Organization?
Google specifically disallows self-reviews on Organization and LocalBusiness sitewide blocks because they were heavily abused. Reviews must be of specific products, services, or locations, not of a company in the abstract. If your business is a single location, attach reviews to a LocalBusiness for that location. For multi-product companies, attach reviews per product.
Can I show fewer than the real number of ratings?
No. Inflated or deflated ratingCount values are flagged by Google's structured data validators and by AI tools that cross-check with third-party platforms. Always emit the real, current count. If you have very few reviews and prefer not to display the markup, omit it entirely rather than fabricating a higher number.
Do I need individual Review objects or just AggregateRating?
AggregateRating alone is sufficient to pass the signal. Individual Review objects on top of AggregateRating add depth and let AI tools surface specific review quotes in answers. If you have the data, emit both; if you only have the average and count, AggregateRating alone is fine.
What if my reviews come from a third-party platform like G2 or Trustpilot?
You can still emit them in structured data on your own site, but you must accurately attribute the source. Some platforms provide widgets that emit the markup for you. If you populate the markup yourself, keep the values in sync with the source. AI tools and validators may cross-check, and discrepancies reduce trust.
Run your own scan
Run a free scan and see how your site grades across all 155 AI-readiness signals.