Review and AggregateRating schema: Google's real required fields

Most guidance on review schema repeats the same short list, add reviewRating, add an author, done. Google’s own review-snippet documentation is longer than that list and stricter than it, because it does not only name required properties, it also names behavior it will disqualify a page for even when every property is technically present. The gap between “markup that validates” and “markup Google will actually show a star for” is exactly where most implementations fail.

The real required and recommended properties#

Google’s documentation splits into two eligible cases, and each has its own required set rather than one shared list.

A single Review needs author (a Person or Organization, with what Google calls a valid name shorter than 100 characters), itemReviewed when the review is not nested inside the item it reviews, and reviewRating carrying a numeric ratingValue. The itemReviewed type has to be one Google actually recognizes for this feature: its documented list names, among others, Book, Course, Event, LocalBusiness, MediaObject, Movie, Organization, Product, Recipe and SoftwareApplication. A review of anything outside that list does not qualify for the star snippet regardless of how the rest of the markup is built.

An AggregateRating without individual reviews attached needs itemReviewed, a numeric ratingValue, and, in Google’s own words, at least one of ratingCount or reviewCount. Neither one alone with the other missing satisfies the rule; the documentation is explicit that one of the two specific properties must be present, not that a count of any kind is enough.

Google’s recommended, non-required additions are datePublished on a Review, and bestRating and worstRating on either a Review’s reviewRating or an AggregateRating. Those three are worth adding but do not block eligibility on their own if a page is otherwise correctly built.

The real documented abuse and validation rules#

The properties above are structural. Google separately disqualifies pages on behavior, and this is the part implementations miss because a validator that only checks property presence has nothing to say about it.

The rule with the most practical bite is the self-serving one. Google’s guidelines state that when the entity being reviewed controls the reviews about itself, a page using LocalBusiness or any other Organization type is ineligible for the star review feature, and that this applies whether the markup is direct or served through an embedded third-party widget. A business publishing testimonials about its own services on its own site, marked up as LocalBusiness reviews, is the exact shape this rule targets. It does not reach Product review markup on a retailer’s own product pages, since the rule is scoped to LocalBusiness and Organization types specifically, which is why product review widgets remain common while self-hosted business-testimonial schema does not qualify.

Separately, Google’s technical guidelines ban fake or undisclosed incentivized reviews outright, defining the category to include reviews that are not based on a genuine experience of the product or service, and reviews written in exchange for money, discounts, vouchers, or free products that do not clearly and prominently disclose that incentive. A page can satisfy every required property and still fail this rule if the reviews behind the markup were bought without disclosure.

The visibility rule closes the remaining gap: Google’s guidelines require that the review content being marked up is readily available to users from the same marked-up page, stating it must be immediately obvious to users that the page has review content. Markup describing reviews with no matching visible text on the page is a documented violation on its own, independent of whether ratingValue, author, and itemReviewed are all correctly filled in.

On the numeric scale, Google’s documentation states the default is a 5-point scale, 1 lowest to 5 highest, and that a missing bestRating is assumed to be 5 while a missing worstRating is assumed to be 1. Neither property is required, but a site using a different scale, a 10-point or a 100-point rating, has to declare both explicitly or Google will read the value against the wrong default range.

Why we care#

Every one of these rules is checkable against a live page in minutes: does itemReviewed carry an eligible type, is ratingCount or reviewCount present, does the reviewed entity control the reviews on a LocalBusiness or Organization page, is the review text actually visible where the markup claims it is, and does the declared scale match bestRating/worstRating if it is not 1 to 5. None of that requires guessing at Google’s intent, all of it is stated on the documentation page itself. What Google’s own reported numbers say about how often sites actually violate these rules at scale, rather than what the rules are, is covered on this site by Google’s review-fraud enforcement scale report, which traces Google’s 2025 Maps enforcement totals separately from the rule text covered here.