How Review and AggregateRating schema work
These are two related types that do different jobs. Review describes one individual review: who wrote it, what they said, what score they gave and the scale that score sits on. AggregateRating describes the summary of many reviews at once: the average score, the scale, and how many reviews or ratings it was calculated from.
Neither stands alone. Both are attached to the thing being reviewed — a product, a book, a course, a piece of software — so the markup says this item has these ratings rather than floating free on the page. When search engines choose to use it, the result is the star rating you sometimes see beneath a listing.
Why Review and AggregateRating schema matter
Stars in a search result are one of the few visual signals that survive the transition from your page to a page you do not control. They change how a listing reads before anyone has clicked it, which is why the temptation to fake them is so strong, and why the rules around them are stricter than for almost any other markup.
The rule that catches most businesses is the ban on self-serving reviews. Ratings that a business collects about itself, published on its own site and marked up on its own page, are not eligible for review stars in search. Testimonials on your home page do not qualify, however genuine they are. Ratings about a specific product you sell, or the other item types search engines support for review snippets, are a different matter.
Where Review and AggregateRating schema go wrong
Marking up ratings the visitor cannot see is the clearest breach. The markup must reflect reviews that are actually on the page, available to the reader, and about the thing the page is about. Pulling an average from a third-party platform and asserting it in your own markup is not the same as displaying those reviews.
Two subtler errors matter too. A rating value on the wrong scale — a score out of one hundred declared against a five-point scale — publishes a number that is simply wrong. And review gating, where only happy customers are invited to leave feedback, produces an average that is technically real and practically dishonest, which breaches the review platforms’ own policies before search ever gets involved.
How to act on it
Start by asking whether the page qualifies at all. If the reviews are about your business rather than a product or another supported item type, put your effort into your public review profiles instead, where the ratings do show, and keep the on-site testimonials as trust content for human readers.
Where markup is appropriate, display the reviews on the page, generate the aggregate from the same data, declare the correct scale, and let the count be whatever it is. Validate with the Rich Results Test and check for stray rating markup a theme or plugin may already be emitting. Collecting genuine reviews steadily, through a proper review generation process, is the part that actually moves the needle; the markup only reports what you have earned.