Review schema: the complete 2026 guide for businesses
Review schema is a small block of JSON-LD structured data that describes a customer review to search engines and AI systems. The three types that matter most for a testimonial wall are Review, AggregateRating, and Quotation. Review schema does not guarantee a star rating in Google, and it does not guarantee a citation by ChatGPT.
What it does is make your reviews easier for machines to understand, which is the prerequisite for both.
A wall of testimonials with valid review schema is structurally a different kind of web page from one without it. Search engines can index it as a collection of reviews. AI assistants can read the same content with the right structure. The reader sees the same thing. The machines see something richer.
Whether to add it is a decision that depends on the size of your site, the type of business, and how much of your customer-proof strategy depends on machine-readable content. The decision framework below covers all three.
What review schema is, in a business sense
In a business context, review schema is structured data that describes a customer review using the Schema.org vocabulary, written in JSON-LD, and embedded in your page.
Think of it this way: schema is a language search engines and AI systems already understand. By writing a review in that language, you make the review machine-readable without changing how a human reader sees it.
The vocabulary is maintained by Schema.org, a collaboration between Google, Microsoft, Yahoo, and Yandex. The most common implementation format is JSON-LD, a JavaScript object embedded in a script tag in your page.
Why it matters: Google rich results and AI search
Review schema matters for two audiences, and the value is different for each.
For Google search. Valid review schema is eligible for rich results, the star ratings you see under a search snippet. Google has historically been cautious about which review types qualify, and the rules have changed over the years. Schema is necessary for a rich result, not sufficient.
For AI search. AI assistants like ChatGPT, Claude, Perplexity, and Gemini rely on a mix of web search and direct crawling. A page with valid schema gives the model a cleaner version of the content than it would otherwise have. The benefit is hard to measure, but it is real.
The honest summary: schema is a prerequisite for both surfaces, not a guarantee on either. Pages with valid schema can still appear without stars. Pages without schema never get them.
The Google side has been more conservative over time. The visible star rating in search results was once common. It is now rarer, partly because Google has tightened the rules about which pages qualify, partly because the rich-result surface itself has been redesigned. The valid use case for schema is structural, not display-oriented. Make the page legible to machines. Let the display follow.
The AI side is newer and less standardized. Different AI assistants consume structured data in different ways. Schema is one of several signals. Clean prose, semantic HTML, llms.txt, and robots.txt all play a role. Schema is a strong signal but not the only one.
The three types of review schema
Three types are in common business use. They describe different things, at different granularities.
| Type | What it describes | When to use it |
|---|---|---|
Review |
A single, named, third-party review of a specific product or service | You have one testimonial and want search engines to recognize it as a review |
AggregateRating |
The average of multiple reviews, with rating and count | You have many reviews and want to surface an average star rating |
Quotation |
A direct quote, not necessarily a review of a product | You have a quote from a customer that is not a product review, like a press mention or a podcast pull-quote |
Most testimonial walls use all three. Review for each individual testimonial, AggregateRating for the wall as a whole, and Quotation for any non-review quote on the page.
The three types have a hierarchy. Review is the atomic unit: one review, one product, one person. AggregateRating is a derived value: the average of multiple Reviews. Quotation is a different category entirely: a quote, not a review. Knowing the hierarchy helps when you write the markup. AggregateRating contains Reviews. Reviews do not contain AggregateRating. Quotation stands alone.
Hypothetical example: a page with and without schema
Two short scenarios. Same wall of testimonials. Different markup.
Hypothetical, page without schema
A testimonial page with five customer quotes rendered as HTML paragraphs. Each quote has a name and a company. A search engine crawler reads the page and sees text. It cannot tell which text is a review, which is a quote, which is a customer story. The content is invisible to machines as reviews.
Hypothetical, same page with valid schema
The same page, with JSON-LD blocks describing each review as a Review, the average rating as an AggregateRating, and one pull-quote as a Quotation. A search engine crawler reads the same page and recognizes five Review items, one aggregate rating of 4.8 from five reviewers, and one quotation. The content is visible to machines as reviews.
Same page. Same reader. Different machine legibility. The schema is what makes the difference.
When each type fits which situation
Different pages call for different schema. A short decision guide.
| Page | Schema type(s) | Why |
|---|---|---|
| Wall of testimonials, one customer | Review | Describe the testimonial as a machine-readable review |
| Wall of testimonials, many customers | Review for each + AggregateRating for the wall | Surface the average rating; structure the individual items |
| Product page with reviews | AggregateRating for the product | Show the average star rating next to the product |
| Press page with a quote | Quotation | The quote is not a product review |
| Landing page with one quote near CTA | Quotation, or no schema | Single non-review quote does not benefit much from schema |
The mistake most businesses make is using Review for everything. A pull-quote from your CEO’s podcast is a Quotation. A customer’s “switched from X to Y, saved 12 hours/week” is a Review. Mixing them costs you either the rich snippet or the data integrity.
When schema is worth adding and when it is not
The decision to add review schema is not always worth it. Three cases where it is not.
A page with a single testimonial, no rating, and no recurring business case for reviews. Schema on a single testimonial adds maintenance cost without producing a star rating. The reader sees the same content either way.
A page with mostly self-written or marketing content. Schema on self-promotional content is the most common manual action Google takes against review markup. It is not worth the risk.
A page where the page itself is not ranking or being cited. Schema helps a page that is already in the index get richer display. It does not help a page that is not indexed at all.
The cases where schema is worth adding: a testimonial wall with many reviews, a product page with structured review activity, a category page that lists multiple products each with their own reviews. The signal is high when the page is the natural review surface for the business.
What schema does and does not do
What matters more than what schema does is what it does not do. Schema is structural, not promotional.
- It does not guarantee rankings. Schema is one of many signals. A page with perfect schema and thin content will not rank. A page with no schema and excellent content can rank.
- It does not guarantee rich results. Google chooses when to show stars. Schema makes a page eligible, not chosen.
- It does not guarantee AI citations. AI assistants use many signals. Schema helps but does not decide.
- It does require maintenance. If the markup does not match the visible content on the page, Google treats the schema as misleading.
What schema does guarantee: the content is machine-readable in a specific, structured way. That is a meaningful prerequisite, not a magic trick.
Schema is a structural prerequisite, not a finishing touch. It is the difference between a page that search engines and AI assistants can read and a page they have to interpret. The interpretation is not impossible. The structure is just cleaner.
The journey after this page
If you have a working model of what review schema is, the next step is to see it in action. The schema markup for testimonials page is the implementation reference, with the full JSON-LD template and the seven mistakes that get the markup ignored.
If you are new to structured data in general, the natural first read is the customer testimonials pillar, which is the parent page for the whole Pillar 2 cluster, including the testimonial vs review distinction and the what makes a good testimonial quality checklist.
If your interest is the AI-search side, the llms.txt for testimonials page covers the companion format that helps AI assistants read your testimonial content cleanly.
If the decision is more about whether to add schema at all, the broader do testimonials help SEO page covers the visibility mechanics, including what schema can and cannot do.
If you are ready to implement, the next step is to look at how to collect the testimonials that will go on your schema-valid wall, with the how to ask and what to ask sides covered separately.
For the broader context, the what is social proof pillar is the parent reference, with the six-type taxonomy that includes review schema as one of the six.
What Testivo does with review schema
Testivo is a testimonial platform that emits valid review-related schema as part of its wall-of-love embed.
Each testimonial is emitted as a Review block, the wall as a whole is emitted as an AggregateRating, and any non-review quote is emitted as a Quotation. The free plan supports up to 15 testimonials. The Pro plan supports more.
Testivo is designed to make customer proof easier for search engines and AI systems to access and understand. It is not designed to guarantee a star rating, a rich result, or an AI citation. Those depend on the broader context of the page, the site, and the search query.
If you want to put review schema into practice without writing JSON-LD by hand, Testivo is one tool that emits the markup automatically. Other tools support the schema layer too: review widgets that emit AggregateRating, schema-generator plugins, and hand-written JSON-LD for developers. The testimonial software buyer’s guide walks through the full evaluation.
Review schema applies to the review type, which is one of eight. The types of social proof page covers the full typology.
Review schema applies to the review output. The customer advocacy pillar is the system context.
Review schema applies to the written format. The types of testimonials page covers all seven formats.
Review schema applies to displayed testimonials. The displaying testimonials page is the placement reference.
The most common schema question is FAQPage vs Review. The FAQ page schema vs review schema page is the new Pillar 8 child that addresses it directly.
The review schema applies to the testimonial output. The customer referrals page is the sibling.
The Review schema type is the most relevant schema for AI citability. The testimonials and AI search page is the AI impact reference, the review schema page is the schema type reference. For the conversion-impact framework, the social proof for conversion page is the dedicated reference.
