Google just banned the reviews Britain outlawed last year

The line Google added on 24 July#

A fox notices small changes in a familiar field, and this was a small change in a very familiar field. On 24 July 2026 a new guideline slipped into Google's review snippet documentation, logged plainly in the Google Search Central documentation updates as a new guideline covering fake and undisclosed incentivized reviews.

The wording in the Google review snippet documentation is direct: do not include fake or undisclosed incentivized reviews on your page or in your structured data markup. Two examples follow it. Reviews not based on a genuine experience of a product or service, and reviews written in exchange for a benefit such as money, discounts, vouchers or free products that do not clearly and prominently disclose the incentivisation.

Read that second example twice, because it is the one that catches honest businesses running an otherwise clean Google review policy. The guideline does not ban incentivised reviews. It bans undisclosed ones. A review earned with a discount code is perfectly acceptable if the page says so, clearly and prominently. The quarry here is concealment, not generosity.

This is a markup rule with a ranking consequence#

Most teams searching for fake reviews Google has acted on will never find this line, because it does not sit in the spam policies. The mechanism matters. This sits in structured data documentation, which means the immediate penalty is not a manual action but a lost rich result. Your star ratings stop rendering, your click-through rate falls, and nothing in your analytics tells you why. That is the quiet part of a guideline change: the damage arrives before the diagnosis does.

Three of July's seven changes touched structured data, per the Search Central changelog, and on our reading the review guideline was the only one of the seven that added a prohibition rather than a capability. That asymmetry is the tell. Google is not expanding what reviews can do, it is narrowing who is allowed to show them.

Two trade reports pinned the same detail down within days of each other. Search Engine Land reported on 24 July 2026 that the added line covers reviews not based on genuine experience and reviews obtained in exchange for money, discounts, vouchers or free products without transparent disclosure. Three days later Search Engine Journal noted that the same documentation warns that if your site violates one or more of these guidelines, then Google may take manual action against it, and that the changelog gives the reason for the addition as to improve user review transparency.

So the rich result is the first thing you lose, not the only thing you can lose. A manual action is a heavier instrument, and the documentation now reserves the right to reach for it over incentivised reviews. That single sentence turns a markup housekeeping job into a site-level risk.

Britain banned incentivised reviews first#

Here is the part that reframes the whole story. Google is not leading on this. It is catching up, and it is catching up by well over a year.

The Digital Markets, Competition and Consumers Act brought new rules into force on 6 April 2025, and as CMS Law sets out, Schedule 20 paragraph 13 prohibits submitting or commissioning a review that purports to be, but is not, based on a person's genuine experience. It equally prohibits concealed incentivised reviews, where a reward such as payment, a discount or a free product is hidden from the reader.

The enforcement is the difference. Google can withdraw a rich result. The Competition and Markets Authority can fine up to 10% of global turnover, and for a banned practice it does not need to prove that a single consumer was misled. The conduct alone is the breach.

That middle figure is not a criticism of Google so much as a warning about where teams take their compliance cues. If your review programme was built to satisfy search guidelines, it has been out of step with British law for four hundred and seventy-four days. The guideline update did not create your exposure. It revealed it.

The reassuring line in both regimes is the same one. Incentivised reviews stay legal and stay markup-eligible when they are labelled prominently and unambiguously, and when they genuinely reflect the reviewer's experience. Nobody is asking you to stop asking for reviews.

What enforcement has actually done since 2025#

Guidelines get attention. Enforcement gets budget. The trail the regulators have left since the start of 2025 is longer and louder than most marketing teams realise, and a good stretch of it runs straight through incentivised reviews.

The Competition and Markets Authority published dedicated fake reviews guidance, reference CMA208, on 4 April 2025, written to help businesses that publish reviews meet their legal obligations and understand what is prohibited by law. The statutory teeth sit one layer below it: paragraph 13 of Schedule 20 makes it a banned practice to submit, or commission another person to submit or write, a fake consumer review or a review that conceals the fact it has been incentivised, per legislation.gov.uk.

The regulator went to the platforms first, and it went quietly. On 24 January 2025 Google signed legally binding undertakings with the Competition and Markets Authority, agreeing that UK businesses caught boosting star ratings with fake reviews get prominent warning alerts on their Google profiles, may have their review function deactivated so they cannot receive new reviews, and for repeat offenders will have all reviews deleted for six months or longer, with Google reporting on compliance to the CMA for three years. The same announcement carries the figure that explains the appetite: the CMA estimates that as much as £23 billion of UK consumer spending a year is potentially influenced by online reviews across six broad sectors.

Amazon followed on 6 June 2025 with undertakings covering fake reviews and catalogue abuse, which the Competition and Markets Authority describes as sellers hijacking the reviews of well-performing products and adding them to an entirely separate and different product, in order to falsely boost its star rating and mislead consumers. Users who post fake reviews risk being banned from writing further reviews, and all their previous reviews being deleted.

Then the patient prowl became a pounce. On 27 March 2026 the CMA opened investigations into fake and misleading reviews at five businesses under its DMCC Act consumer powers, bringing the total number of businesses under review using the new consumer powers to 14, per the Competition and Markets Authority. Chief Executive Sarah Cardell put the stakes in one sentence: "Fake reviews strike at the heart of consumer trust - with many of us worrying about misleading content when looking at reviews online."

One of those five is our subject exactly. The CMA is examining whether Pasta Evangelists offered customers discounts for leaving 5-star reviews without disclosure, which is the textbook shape of undisclosed incentivised reviews, per the Competition and Markets Authority. The CMA expects to make an update on these investigations in September 2026, and where it finds an infringement of the law it can require businesses to change their practices and impose fines of up to 10% of global turnover.

The academic evidence explains both why the temptation persists and why it collapses. He, Hollenbeck and Proserpio, writing in Marketing Science, studied a market for fake Amazon reviews operating through Facebook and found that buying them is associated with a significant but short-term increase in average rating and number of reviews, and that rating manipulation has a large causal effect on sales. After firms stop buying, their average ratings fall and the share of one-star reviews increases significantly, particularly for young products, which the authors read as evidence that rating manipulation is mostly used by low-quality products. A bought rating is a loan, not an asset.

What counts as a real disclosure#

Most review programmes that fail do not fail on honesty, they fail on placement. The incentive was real, the review was real, and the disclosure was three clicks away in a methodology page nobody opens.

Both regimes use the same two adverbs. Google asks for incentivisation disclosed clearly and prominently. The CMA guidance summarised by CMS Law asks for labelling that is prominent and unambiguous. Neither is satisfied by a footnote.

The structural rule is simple enough to hand to a developer. The disclosure travels with the review, not with the site. If a review can be syndicated, excerpted or marked up on its own, the label has to survive every one of those journeys. Track the label the way you would track a scent across wet ground: if it disappears at any point, the trail is broken.

The aggregate rating inherits the problem#

Aggregate ratings are where this gets expensive, because an average built partly from undisclosed incentivised reviews is itself a misleading presentation. The CMS Law summary is explicit that publishing consumer review information in a misleading way is caught too, including suppressing negative reviews, cherry-picking positive ones and presenting outdated material.

So the audit cannot stop at individual reviews. Any aggregate rating you mark up carries the compliance of every review inside it, which means one undisclosed incentive can taint the star rating on every product page that draws from the same pool.

The audit, done in a week#

None of this needs a new platform. It needs somebody to walk the trail from incentive to rendered star and check that the label survives the journey. That is a week of work for most sites and an afternoon for a small one.

One more guideline is worth folding into the same sweep, because it catches sites that never ran an incentive at all. Per Google's review snippet documentation, Google also asks that reviews be marked up for specific products or services rather than categories or lists, that review content be readily visible on the page, and that reviews for a local business come directly from users rather than editors or self-serving sources.

Search Engine Roundtable lays out the exact one-sentence guideline addition and quotes Barry Schwartz dating it precisely.

Running incentivised reviews without breaking anything#

An audit clears the undergrowth. A review management programme keeps it clear. Incentivised reviews are not a project with an end date. They are a hedgerow that grows back, which is why the work belongs on the SEO and GEO roadmap rather than filed in a legal folder and forgotten.

Start at the offer, because that is where the whole thing is won or lost. Venable LLP reports that the FTC's warning letters singled out compensation, such as a discount, refund or free product, that is expressly or implicitly contingent on the review expressing positive sentiment. Contingency is the hinge that turns incentivised reviews from lawful into prohibited, so ask for a review, not for a good one.

Detection is now part of the day job#

The platforms have industrialised this, and their numbers set the baseline you are measured against. In 2025 Google blocked or removed over 292 million policy-violating reviews while publishing more than 1 billion helpful reviews, removed over 13 million fake Business Profiles, placed posting restrictions on more than 782,000 policy-violating accounts and blocked 79 million inaccurate or unverified edits. The same nets that catch outright fakes also catch badly labelled incentivised reviews, so a labelling slip reads to a platform as an integrity signal rather than an oversight.

There is a newer scent on this trail, and it is not one you laid. The Transparency Company, a review-fraud detection firm whose own interest here is worth weighing, examined roughly 73 million reviews across 127 business subcategories in home services, legal and medical across 100 US cities, and found 13.7% were actually or highly likely to be inauthentic, with annual consumer harm in those sectors put at approximately $300 billion. It also found 3.1% of the reviews in its dataset were AI-generated, growing 80% month over month since June 2023, which makes moderation a content marketing discipline as much as a legal one.

Why this bites harder in 2026#

Reviews used to be a conversion asset with a search bonus attached. They have quietly become something more load-bearing, because generative systems lean on them heavily when summarising what a product is like. A star rating is a compact, quotable, machine-readable judgement, which is exactly the shape an answer engine reaches for, and it quietly makes reputation management an organic search discipline.

That is the reframe worth carrying into the budget conversation. Cleaning up incentivised reviews is not a compliance chore, it is reputation management with a search dividend attached, and it protects the single most quotable asset on your product pages. Teams treating it as SEO and GEO work rather than legal housekeeping tend to fix it faster, because they can see what it earns rather than only what it avoids.

The regulated categories carry the most exposure, as usual. A discount-driven review programme in healthcare marketing stacks consumer law on top of sector rules about testimonials, and in FinTech marketing the same review may also count as a financial promotion. Two regimes on one sentence is manageable. Three is where teams start missing things.

There is a content angle too, and it is the cheerful one. Labelled incentives read as confident rather than compromised, and a review programme that says plainly how it works usually converts better than one that hides its methodology. That is a content marketing win hiding inside a compliance task.

Fold the sweep into whoever already owns review management and be done with it. Count incentive routes, count labelled reviews, count the gap, and keep the aggregate ratings clean. The fox that checks the same hedgerow every season is never surprised by what has moved into it.

Frequently asked questions#