Review schema and rich results: the 2026 UK compliance guide

What review schema markup does, and why 2024 changed it

Review schema markup is structured data (usually JSON-LD) that tells search engines a page contains reviews of a specific product, service, or local business. When Google trusts both the schema and the underlying reviews, it may render star ratings alongside the search result.

The 2024 clarifications were narrow but consequential. Google confirmed that reviews rendered via a third-party widget qualify for rich results only if the reviews are visible to the visitor on the page (not hidden behind a click), and only if the schema emitted matches those visible reviews exactly. Emitting schema for hidden or aspirational reviews is now the single most common trigger for a manual action in this area.

Google also narrowed the entity types eligible for the star snippet in 2024. Generic Organization, Article and Blog types are no longer eligible; only Product, LocalBusiness (and its subtypes), SoftwareApplication, Recipe, Book, Movie and Event are.

Third-party reviews on service pages remain eligible for LocalBusiness, provided the reviews are for the local business itself and not for a specific unrelated product. This is the pattern most UK service businesses need.

The canonical LocalBusiness JSON-LD for a UK service page

The pattern below is what our team recommends for a typical UK service business page. It is minimal, valid against Schema.org, and passes the current Google Rich Results test as of March 2026.

Emit one LocalBusiness object per page (or a subtype such as Dentist, Plumber, LegalService for tighter categorisation). Include aggregateRating summarising all visible reviews, and one review object per visible review — no more, no fewer.

The aggregateRating.reviewCount must equal the number of review objects emitted and the number of reviews visible on the page. Any mismatch is a common cause of silent snippet loss.

Reviews must include author (Person with name), reviewBody, reviewRating (with ratingValue), and datePublished. Missing any of these fields does not fail validation but reliably suppresses the snippet in production.

For multi-location operators, emit one LocalBusiness per location page, each with its own aggregateRating scoped to that location's reviews only. Do not roll up group-level reviews onto individual location pages — that is a policy breach.

The diagnostic checklist when stars disappear

The most common support question we receive on review schema markup is 'we used to have stars in search and now we don't'. Nine times out of ten it is one of the six issues below, in this rough order of frequency.

1. 1The schema emits reviews the visitor cannot see (hidden pagination, tabbed content, 'load more' behind a click). Fix: render all reviews in the initial DOM or paginate on the server side.
2. 2The reviewCount in aggregateRating does not equal the number of review objects. Fix: recompute at render time; never hardcode.
3. 3The entity type is Organization or Article. Fix: switch to LocalBusiness or a specific subtype.
4. 4The reviews are self-authored by the business. Fix: use only third-party reviews and mark them as such.
5. 5The reviews are duplicates from another domain (scraped or syndicated without permission). Fix: display only reviews collected on your own instance of the platform.
6. 6A manual action is in place. Fix: request review in Search Console once the underlying issue is remediated.

Widgets, platforms and the safe pattern in 2026

Most UK operators do not hand-write review schema markup. They use a widget from a review platform that emits schema on their behalf. The safe pattern in 2026 has three requirements a platform must meet.

First, the widget must render reviews in the initial HTML DOM, not inject them via JavaScript only after user interaction. Google's rendering pipeline has become more tolerant of client-side rendering since 2024, but for review rich results the safest posture remains server-side or initial-DOM rendering.

Second, the schema the widget emits must match exactly what the visitor sees. If the widget shows 'load more' after 5 reviews, the schema must emit only those 5. As soon as the visitor clicks and 20 more reviews appear, either the schema must update (SPA) or the initial state must include all 25 (SSR).

Third, the widget must expose per-review verification tier data in the review object as a plain-language qualifier — this is a soft signal and does not affect eligibility, but Google has been observed to weight verified reviews more highly in the snippet ranking since mid-2025.

ScoreReview UK's widgets meet all three requirements by default. If you use a different platform, ask the vendor for its Rich Results test output on a sample page — if they cannot produce it, that is your answer.

“Google's 2024 tightening did not kill review rich results — it killed the shortcut. Businesses that display real reviews with matching schema still see stars; the ones that were faking it lost them, correctly.”

Summary: what to check on your own site this week

Run the Rich Results test on your top three landing pages, confirm each emits LocalBusiness or Product schema with matching aggregateRating and visible review objects, and note any mismatches. Then check Search Console for manual actions.

If everything validates and you still do not see stars in production search, the cause is almost always a soft signal — small volume of reviews, low review recency, or an entity that Google is not yet confident enough to feature. All of these resolve with a sustained review programme.

Rich results are the compounding return on an honest review programme. There is no shortcut and, since 2024, there is no longer a workaround. Do the work once and the snippet compounds for as long as the reviews keep flowing.