Review schema: who can earn star ratings, and who cannot

What review schema is — and who is allowed stars

Review schema is the structured data that attaches a rating to a named item — one branch of schema markup. Google reads it and may print a review snippet: the stars, the score and the count beneath your search result.

Here is the restriction almost nobody writes up.

Google's review snippet reference (last updated 8 September 2026) supports Local business and Organization reviews only for sites that capture reviews about other businesses or organisations. If the reviewed entity controls the reviews about itself, its pages using those types are ineligible for the star review feature.

Read that twice if you run a brand site. The twelve testimonials on your homepage cannot earn stars — not because your code is wrong, but because you collected them about yourself.

Eligibility is Google's decision, not a function of whether your markup validates. These are the primary item types it names:

Google also supports reviews on a further set of schema.org types and their subtypes:

Article, Service and Job posting are on neither list. Review markup there is valid schema.org code that will never draw a star.

So is it still worth adding? On a supported type, with genuine reviews you did not collect about yourself, yes. Anywhere else, no. You gain machine-readable context nobody has promised to use, and you take on guideline risk for a snippet that cannot appear.

Review vs AggregateRating: two tags, two jobs

Review marks up one individual review — one author, one score, one date. AggregateRating marks up the summary — the average score across many ratings, plus how many there were. Almost every star rating you see in search is AggregateRating, because Google prints the average, not one person's opinion.

• Review needs a reviewer and the item. Use it for the individual reviews on the page.
• AggregateRating needs a value and a count — the "4.3 out of 5 from 412 ratings" line.
• Both can sit on one item: the aggregate as the summary, nested Review objects underneath.
• Either way, the ratings must be visible to a visitor on that page. Describing ratings a user cannot see is the most common reason review snippets quietly stop appearing.

Review properties: required, recommended, and what breaks

Four things are required: what was reviewed, who reviewed it, the rating object, and the score inside it. Everything else is recommended — the snippet can render without it, but you lose accuracy.

AggregateRating properties: ratingCount is not reviewCount

AggregateRating needs the item, the average value, and at least one count. The two count properties are not interchangeable.

Worked example. A copper water bottle on a D2C store collects 412 star ratings, of which 58 people also wrote something. That is ratingCount 412 and reviewCount 58 — not 412 twice, and not 58 in both fields. Supply both when you have both.

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Copper Water Bottle 1L",
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.3",
    "bestRating": "5",
    "worstRating": "1",
    "ratingCount": "412",
    "reviewCount": "58"
  }
}

Google's review guidelines — and what changed on 24 July 2026

On 24 July 2026, per Google's documentation updates log, a new review snippet guideline was added covering fake and undisclosed incentivised reviews. In plain English: the reviews have to be real, and anything you gave in exchange has to be disclosed.

The full set of rules you have to clear before marking anything up:

• No fake or undisclosed incentivised reviews. Google's wording: "Don't include fake or undisclosed incentivized reviews on your page or in your structured data markup." Its examples: reviews not based on a genuine experience, and reviews given in exchange for an undisclosed benefit.
• Disclose any benefit given. A discount, a free sample or a prize draw entry counts, and the disclosure has to be clear and prominent.
• No reviews you control about yourself. Local business and Organization are supported only for sites capturing reviews about other businesses or organisations. Where the reviewed entity controls the reviews, those pages are ineligible for the star review feature.
• Reviews must be on the page, visible to a user, and about the exact item you marked up.
• Only mark up ratings your own site hosts. Pulling an average in from a marketplace or a maps profile and declaring it as yours is not allowed.

One more honest note, because the opposite gets sold a lot. Structured data is not an AI visibility tactic. Google's guidance on its AI features (last updated 10 July 2026) states there are no additional technical requirements and no special structured data needed to appear in AI Overviews or AI Mode. Review markup governs rich-result eligibility in Search. That is the whole of its job.

Why we decline testimonial markup — and what we do instead

From our own client work: this request reaches us every few months. "We have twelve glowing testimonials on the homepage. Can you make the stars show up in Google?"

We say no, and we give the same two reasons each time.

First, it would not work. Google supports Organization and Local business reviews only for sites that review other organisations, and testimonials you gathered about yourself are the exact case its documentation rules out. The page is ineligible before the markup is even written.

Second, the downside is not zero. Markup that misrepresents a page is exactly what a manual action for spammy structured data exists to catch. Trading site-wide credibility for a snippet that cannot render is a bad bet.

So we do not put Review or AggregateRating markup on our own testimonials, or on a client's. Here is what we do instead:

1. Send rating demand to Google Business Profile. Real customer reviews there already carry stars in Maps and local results — see how to optimise the profile so they get seen.
2. Fix the local fundamentals first. A profile nobody surfaces collects no reviews. Work the local SEO checklist before chasing stars.
3. On ecommerce, collect genuine reviews on product pages and mark those up. Product is an eligible type, which makes it the one place stars are properly winnable — the ecommerce SEO guide covers the page structure.
4. Treat review volume and sentiment as its own workstream, not a markup job. In regulated sectors it needs a defined process, as in our take on healthcare reputation management.

Testimonials still belong on the page. They convert. They just do not get stars.

Three patterns that risk a manual action

Almost every review snippet problem we audit is one of these three.

Where to add the markup, and how to check it

One WordPress trap first. Do not paste a JSON-LD block into the post editor. Many sites — ours included — strip in-body JSON-LD to stop duplicate schema, so it silently disappears on save. It belongs in your SEO plugin's custom schema field, your theme head, or the plugin that already holds your real review data.

1. Confirm the item type is supported and that you do not control the reviews about the item. If either fails, stop here.
2. Check the reviews are real, yours and visible. If any of the three is false, the markup is a liability.
3. Build the block. Our schema markup generator produces the object with the right property names.
4. Paste it into AIOSEO > Schema > Custom Schema (or your own plugin's equivalent), never into the body.
5. Validate the live URL in Google's Rich Results Test, then watch the matching Search Console enhancement report for errors.
6. Keep the numbers in sync. A hardcoded ratingValue drifts within a month. Generate it from live review data.

Where to go from here

Stars are winnable. They are just not winnable on reviews you own.

For most Indian businesses that leaves two real places: product pages carrying genuine customer reviews, and the Google Business Profile that already shows ratings. Both run on review collection, not markup. Get that right and the snippet follows, with nothing on the page you would have to defend.