What the law actually says
The Digital Markets, Competition and Consumers Act 2024 lists fake and misleading reviews among its "blacklisted" commercial practices, automatically unfair with no need to prove harm, at Schedule 20, paragraph 13. Three separate things are banned: submitting or paying for a fake review, publishing a review without disclosing that it was incentivised, and publishing reviews in a way that's misleading, hiding negative reviews or giving positive ones more prominence than they'd otherwise get.
That third point puts a duty on the platform itself, not just the reviewer. A site publishing reviews at all now has to take reasonable, proportionate steps to stop fake or undisclosed-incentive reviews appearing in the first place, which is the legal basis behind Amazon's own undertakings below.
How big are the fines, and why
The same Act gave the CMA a power it didn't have before: fining a consumer-law breach directly, without taking the case to court first. Its own guidance on the new regime puts the ceiling at 10% of a business's global annual turnover, not just its UK revenue.
The reason it can go that high comes back to the blacklist itself. Because fake reviews are automatically unfair, the CMA doesn't have to show that any particular shopper was actually misled or changed their mind, only that the practice happened. That's a materially lower bar than most consumer-protection cases, and it's why the exposure is measured against turnover rather than a fixed penalty.
Real cases: who the CMA is actually investigating
On 27 March 2026 the CMA opened its first formal DMCC Act cases over fake or misleading reviews, against five named businesses. All five remain open investigations, not proven findings.
| Business | What's being investigated |
|---|---|
| Autotrader & Feefo | Whether 1-star reviews were excluded from Autotrader's platform ratings |
| Dignity | Whether staff were asked to write positive reviews of its crematoria services |
| Just Eat | Whether its ratings system inflated restaurant and grocer star ratings |
| Pasta Evangelists | Whether customers were offered undisclosed discounts for 5-star reviews |
Amazon's own record on this
Amazon isn't a bystander here. In June 2025, it gave the CMA formal undertakings covering incentivised reviews, reviews from fake or compromised accounts, and catalogue abuse, sellers merging a product listing with another one to inherit its reviews and star rating. Google gave a similar undertaking in January 2025. Neither case ended in a fine; both were resolved through the company committing to change its practices, and the CMA followed up in July 2025 by sending 54 advisory letters to other businesses about fake-review compliance.
Worth being plain about what that means: this isn't a hypothetical risk for one platform, it's an active enforcement programme that has already reached the biggest marketplace in the UK.
How Amazon tries to catch fake reviews
Amazon's own account of its detection systems, published on its .co.uk policy pages, describes machine-learning models trained on seller ad spend and abuse reports, language-pattern detection for incentivised reviews, and graph-based models built to spot networks of coordinated fake accounts rather than single bad reviews in isolation. Amazon states it proactively blocked more than 250 million suspected fake reviews worldwide in 2023, its own figure, not independently audited.
The Verified Purchase badge is a narrower check than people often assume. It confirms the reviewer bought the item through Amazon's own marketplace at close to full price, which rules out some fake-review patterns but not others. A real buyer can still be paid or given a product to leave a glowing review without disclosing it, exactly the practice the DMCC Act now bans.
Real signs a review might be fake
- A sudden cluster of 5-star reviews landing the same day, especially on a product with little other review history.
- Language that discloses an incentive without saying so plainly, phrasing like "received this for testing" without stating whether it was free or discounted, which is the exact grey area the DMCC Act's disclosure rule targets.
- A reviewer's history full of unrelated products, all 5 stars, all posted in a short window, a pattern more consistent with paid reviewing than genuine, varied shopping.
- Review text that doesn't actually match the product listed, mentioning a different colour, size or model, a sign of catalogue abuse rather than a fake reviewer.
- Treating a Verified Purchase badge as proof of an honest opinion, rather than proof of a real transaction, is the most common mistake; the two aren't the same guarantee.
What to do if a review misled you
Report the listing or the specific review to Amazon; that's the direct route and the one that feeds Amazon's own enforcement systems described above.
There's no separate legal right to compensation just because a review turned out to be fake. But if the item itself genuinely doesn't match what was advertised, the general protections in the Consumer Rights Act 2015, refund, repair or replacement, apply on their own terms, regardless of what any review said.
About this page
LightningBrands doesn't verify reviews or investigate sellers ourselves, this page exists because the legal side of fake reviews is real, currently under-reported, and worth setting out plainly. Every specific fact above traces to a primary source: the DMCC Act's own text on legislation.gov.uk, the CMA's own case announcements on gov.uk, and Amazon's own policy statements on its detection systems. Where a figure couldn't be independently confirmed, it's left off rather than repeated as fact.
See how Amazon's own discount storefronts actually work or today's hand-checked deals on this site.