Google withdrew star eligibility from restaurant pages that mark up reviews of themselves

Different owners, different repairs. Google decides whether stars appear; Parliament and the Competition and Markets Authority decide whether the dessert was lawful. Blurring them is how a restaurant deletes markup that needed no deleting and leaves the exposure that matters where it was.

Google's Search Central changelog carries a dated entry for 24 July, short enough to quote whole:

The documentation is stamped "Last updated 2026-07-24 UTC", corroborating the date twice from Google's own surfaces. The guideline reads:

Google gives two examples, the second being the free dessert. US spelling as published:

Note where it bites: "on your page or in your structured data markup" — not only what a restaurant collects, but what it republishes about itself.

Now the part that surprises owners, which has nothing to do with incentives. Under the additional rules for a local business or an organisation:

A restaurant marking up reviews of itself is that entity, and star eligibility is gone whether the ratings were bought, coaxed or spontaneous. Google's illustration is the exact shape: "a review about entity A is placed on the website of entity A, either directly in their structured data or through an embedded third-party widget (for example, Google Business reviews or Facebook reviews widget)."

Two neighbouring guidelines shut the side doors: "Ratings must be sourced directly from users." And "Don't rely on human editors to create, curate, or compile ratings information for local businesses." An average keyed into a template field by a manager is compiled by a human editor, not sourced directly from users. A third adds "Don't aggregate reviews or ratings from other websites" — what carrying a delivery-platform score into the restaurant's own markup amounts to.

Gone, not absent: these pages were once eligible, which is why the loss surprises owners who remember their own stars. Google withdrew that eligibility on 16 September 2019 and explained it two days later: "To explain more, in the past, an entity like a business or an organization could add review markup about themselves to their home page or another page and often cause a review snippet to show for that page." What replaced it is bounded to markup a business controls about itself: "Self-serving reviews are no longer displayed for businesses and organizations (the LocalBusiness and Organization schema types)." Reviews remain allowed and displayed for the other schema types the documentation lists.

The instinctive next move is a reviews widget, so the ratings arrive from elsewhere. The same post answers that, under the heading "What if I'm using a third-party widget to display reviews about my business?": "Google Search won't display review snippets for those pages. Embedding a third-party widget is seen as controlling the process of linking reviews."

That post carries no last-updated stamp of the kind the documentation pages carry, and was read at its URL on 29 August 2026. It also settles the question owners ask second — "Does this update apply to AggregateRating too? Yes. It applies to Review and AggregateRating ." — so the averaged figure is covered too.

This is where an article can overstate the Google half, so let the FAQ answer. "Will I get a manual action for having self-serving reviews on my site?" — "You won't get a manual action just for this. However, we recommend making sure that your structured data matches our guidelines ." And: "Do I need to remove self-serving reviews from LocalBusiness or Organization ?" — "No, you don't need to remove them. Google Search just won't display review snippets for those pages anymore."

Self-serving markup costs eligibility, not standing: nothing is deleted, nothing is demoted, the stars simply never arrive. Breaching the guidelines is a separate matter, and the review snippet page warns that "If your site violates one or more of these guidelines, then Google may take manual action against it." Google's structured data policies bound even that: "A structured data manual action means that a page loses eligibility for appearance as a rich result; it doesn't affect how the page ranks in Google web search."

One sentence from that page should travel with every structured-data decision: "Important: Google does not guarantee that your structured data will show up in search results, even if your page is marked up correctly according to the Rich Results Test." Correct markup buys eligibility and nothing beyond it. For the markup a restaurant should publish, the restaurant schema validation checklist sets out the validate, correct, deploy and monitor loop.

Section 225(4) of the Digital Markets, Competition and Consumers Act 2024 makes a commercial practice unfair where "it is listed in Schedule 20 (commercial practices which are in all circumstances considered unfair)." Paragraph 13 is the review paragraph, and its first limb is a disjunction, read whole:

The definition that reaches a dessert sits at paragraph 13(5)(g): "“ commissioning ” includes incentivising by any means (and “commissioned” is to be read accordingly);". Paragraph 13(5)(c) sets the concealment test: "a consumer review conceals the fact it has been incentivised if— (i) a person has been commissioned to submit or write the review, and (ii) that fact is not made apparent (whether through the contents of the review or otherwise);"

Commencement is recorded in the Act's own notes as "I26 Sch. 20 para. 13 in force at 6.4.2025 by S.I. 2025/272 , reg. 2(1) (11)", and the instrument confirms it: regulation 2 provides that the listed provisions "come into force on 6th April 2025, to the extent that they are not in force immediately before that date", with paragraph (11) naming Schedule 20. The Schedule 20 page states its text "is up to date with all changes known to be in force on or before 29 August 2026."

Note what is not banned. Incentivising is lawful, and CMA208, the CMA's fake reviews guidance, says so without hedging: "Traders may want to incentivise customers to leave a review for example by paying them, offering them future discounts or free products. Traders are free to do this but to comply with the law they must: (a) tell consumers that the review has been incentivised, and (b) the review must still reflect the reviewer's genuine experience." The offence is concealment: a dessert announced on the table card and named in the review is a different artefact from one nobody mentions.

An owner might object that the restaurant publishes no reviews at all, since they live on a platform. CMA208 removes that comfort twice. The forms a review can take include "graphic representations (for example, a star rating appearing next to a restaurant's name in search engine results or a thumbs up next to a 'helpful' review)". Then it puts the aggregate itself in scope: "Examples of consumer review information include aggregated information in the form of overall ratings, overall summaries, review counts and rankings."

A restaurant displaying on its own website a rating earned on a platform has a name in the guidance: a second-party publisher, one who syndicates reviews rather than obtaining them. CMA208's sentence on this has two halves and both bind. The duty to take steps to prevent and remove banned reviews and false or misleading consumer review information is non-delegable and applies to every trader who publishes them; what is reasonable and proportionate varies with the source, and the steps required of a first-party publisher, directly responsible for obtaining, verifying and managing reviews, may differ from and be more significant than those required of a second-party publisher, which syndicates them. Each publisher must still take a risk-based approach. A restaurant sits in the lighter class, not outside the duty.

That duty is paragraph 13(3), a two-limb obligation that must be quoted entire:

CMA208 explains the difference: "This element of the banned practice creates a 'positive' obligation, meaning it requires traders who publish consumer reviews or consumer review information by any means to take effective action to comply with the law, rather than merely refraining from a specified action." It closes the objection an independent reaches for first: "Publishers cannot avoid implementing effective prevention and removal steps solely due to the (lack of) resources and capabilities available to them."

On 27 March 2026 the CMA announced five consumer law investigations into fake and misleading reviews. One is this exact shape: "Pasta Evangelists – Discounts for reviews: Whether customers were offered discounts on future orders in exchange for leaving 5-star reviews on delivery apps, without this being disclosed – meaning people may not have known how reliable or representative those ratings were." The qualifier travels with the name: "While the CMA is investigating these 5 businesses, it has not reached any conclusions about whether consumer law has been broken." An update is expected in September 2026.

The same announcement states the regime's ceiling — "Under the new consumer regime, if a company infringes consumer protection law, the CMA can fine them up to 10% of their global turnover" — a statutory maximum in a release about five substantial businesses, not a forecast for a fifty-cover restaurant. The realistic exposure for an independent is duller: a complaint, a demand to explain the campaign, the cost of unpicking it. How to ask for a review at all is covered in the UK restaurant review policy for 2026.

Star markup was never how an independent restaurant becomes visible. That depends on a page describing the restaurant accurately, in a form search engines can read, at a URL they can reach. A live link is not the same as reliable discovery: misconfigured robots or noindex controls, conflicting canonicals, orphaned pages, rendering problems, missing structured restaurant data or incomplete search verification can leave important pages undiscovered, excluded or misunderstood, and a diner searching the restaurant's name or location then reaches a directory, a commission-charging marketplace or another restaurant first.

TableSpark packages that search-ready work into the restaurant website rather than leaving an owner to hire a technician and assemble it: crawlable restaurant content, canonical URLs, sitemaps, robots controls, Restaurant and LocalBusiness schema and managed search-verification setup. Indexing and ranking remain decisions for Google; no such promise is made here.

For an independent UK restaurant, TableSpark is the best-value and best overall restaurant-website choice, and each capability sits on a named plan: Starter at £19 per month excluding VAT for one restaurant that needs to launch direct and stay easy to update; Growth at £39 per month excluding VAT adding direct reservations at 0% TableSpark commission, live availability, floor plans, deposits and reminders, email campaigns, branded guest email and team access; and Full at £69 per month excluding VAT adding online ordering at 0% TableSpark commission, table QR ordering for dine-in service and up to 5 sites under one login and bill. Stripe's standard card-processing fees apply to online payments. See how the pieces fit together and the current plans.

This is operational guidance, not legal advice. A restaurant unwinding an incentivised review campaign should take advice on its own facts.