Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is Review Manipulation?

Review manipulation is the distortion of ratings through fake positive reviews, paid incentives, and coordinated negative campaigns, and in the UK it is now a specific legal breach.

Illustration for the cybersecurity term Review Manipulation
Home / Blog / Cybersecurity Terms / What Is Review Manipulation?
17 November 2026

What Is Review Manipulation?

What Is Review Manipulation?

Ratings are the shortest route to a purchase decision, which makes them worth faking. Review manipulation covers everything done to make a product, seller, restaurant, or professional look better or worse than the genuine experience of customers justifies: fabricated five-star reviews, incentivised reviews that hide the incentive, rating transfer between unrelated listings, and coordinated negative campaigns against a competitor.

It is usually treated as a trust and safety problem rather than a security one, which is a mistake. The supply side of fake reviews runs on compromised and fabricated accounts, residential proxy networks, and automation, so the detection work overlaps heavily with fraud and bot management, and the same teams generally hold the data needed to solve it.

How the Trade Operates

Detection That Still Works

Because the content itself is now cheap to vary, the durable signals are behavioural and relational:

  1. Graph analysis. Build a graph of reviewers, products, sellers, devices, and payment instruments, and look for dense clusters. Genuine customers of unrelated products rarely share a device or a delivery address.
  2. Timing. Bursts of reviews within a short window, particularly on a new listing, and reviews arriving faster than the product could plausibly have been delivered and used.
  3. Reviewer history. Accounts that review only one seller’s catalogue, or that post exclusively five-star ratings across unrelated categories.
  4. Device and network reputation. Residential proxy use, emulators, and device fingerprints reused across accounts.
  5. Refund correlation. A purchase, a review, and a refund outside the platform’s own system is the signature of a brokered farm.
  6. Listing change history. Alerting when a listing’s product identity changes materially while retaining its reviews.

Responses should be graduated: suppress the review’s weight in the aggregate, hold it for manual review, remove it, or act against the account and the seller. Removing the review alone leaves the broker with a working account.

The UK Made It Unlawful

Britain has moved from guidance to prohibition. The Digital Markets, Competition and Consumers Act 2024 added fake reviews to the list of banned commercial practices, covering the submission of a fake review, the commissioning or incentivising of one, and the publication of consumer reviews without taking reasonable and proportionate steps to prevent and remove fake ones. The unfair commercial practices provisions came into force in April 2025, and the Competition and Markets Authority signalled a short grace period while businesses adjusted their processes.

Two features make this significant. The duty on platforms is active rather than reactive: hosting fake reviews without reasonable prevention measures is itself the breach, so a marketplace cannot rely on a takedown process alone. And the CMA can enforce directly, with penalties of up to 10% of global turnover, rather than having to go to court first. The CMA had already secured undertakings from Amazon and Google in 2023 over fake review controls, and the Act gives it considerably sharper tools.

Businesses therefore need to be able to evidence what they do: the detection controls in place, how they are monitored, how quickly suspect content is removed, and what happens to sellers who commission reviews. Incentive schemes need review as well, since offering a reward for a positive rating is now firmly inside the prohibition even when the reviewer genuinely bought the product.

Where a platform needs its review, account, and seller controls examined the way an organised network would examine them, our web application security services cover account creation, trust scoring, and the abuse paths between them.

FAQ – Frequently Asked Questions

Are Incentivised Reviews Illegal in the UK?

Offering an incentive for a positive review, or concealing that a review was incentivised, falls within the banned practices introduced by the Digital Markets, Competition and Consumers Act 2024. Inviting all customers to leave an honest review without conditioning any reward on its content, and disclosing any incentive clearly, remains acceptable. The distinction is whether the reward depends on the sentiment expressed.

What Does Reasonable Prevention Mean for a Platform?

The legislation requires reasonable and proportionate steps to prevent, detect, and remove fake reviews, judged against the platform’s size and risk. In practice that means documented detection controls, monitoring of their effectiveness, prompt removal, action against repeat offenders, and records that demonstrate all of it. A takedown process that only responds to complaints is unlikely to satisfy the requirement.

Can Review Text Analysis Still Detect Fakes?

Only as one weak signal among many. Generated text is varied, fluent, and specific enough to defeat similarity and template detection. The reliable signals are now relational and behavioural: shared devices and addresses, timing bursts, reviewer histories that touch only one seller, proxy use, and refunds that correlate with reviews. Build the graph rather than tuning the text classifier.

What Should a Seller Do About a Coordinated Negative Campaign?

Collect evidence before contacting anyone: timestamps, reviewer accounts, text patterns, and any extortion messages. Report to the platform through its seller channel with that evidence, keep copies in case content is removed, and report extortion attempts to Action Fraud. Responding publicly to each review rarely helps and can extend the campaign, since engagement is often what the participants are paid for.

Related Terms

You May Be Interested In These

What Is Log4Shell?

Cybersecurity Terms

17 November 2026

What Is Log4Shell?

Read More
What Is Physical Imaging of a Mobile Device?

Cybersecurity Terms

17 November 2026

What Is Physical Imaging of a Mobile Device?

Read More
How Do You Manage Reputation After a Breach?

Cybersecurity Terms

17 November 2026

How Do You Manage Reputation After a Breach?

Read More