Review fraud
Detect fake reviews posted from the same device across many accounts or generated by bots.
Fingerprint fit Supporting signal
Overview
Review fraud is the posting of fake reviews to manipulate the reputation of a product, service, or company. It comes in two forms: fake positive reviews, where someone buys or generates five-star ratings to inflate their own rankings and sales, and fake negative reviews, where low ratings are used to damage a competitor, called review bombing when done at volume. The scale is large enough that major platforms remove reviews by the hundreds of millions each year, and fake reviews are estimated to influence well over a hundred billion dollars in global spending. Beyond misleading shoppers, review fraud drives returns, customer complaints, regulatory action, and a hard-to-measure loss of trust when products do not match their ratings.
Review fraud reaches a platform through several channels, and they are not equally visible to any one tool. Some fake reviews are written by real people paid a commission to buy an item and leave a five-star rating, and AI now generates realistic review text at scale, both of which are best caught by content-level methods like sentiment and language analysis, network analysis of reviewer-product clusters, and manual moderation. Other fake reviews are produced by scripts that create an account, post a review, and repeat, often hiding behind VPNs, cleared cookies, browser tampering, or CAPTCHA-solving services. This automated, repeat-device side is where device intelligence contributes.
Fingerprint supports a layered review-fraud defense by adding device-level signal to the content and network analysis a platform already runs. It assigns each browser and device a stable, unique Visitor ID that persists even when cookies are cleared, IPs change, or browsers are tampered with, so a platform can recognize suspicious patterns like one device creating many accounts in a short window or posting reviews across multiple accounts, and block a known offender at the device level regardless of which account they switch to. Bot Detection flags reviews submitted by automated tools, while Suspect Score add risk context to gate suspicious submissions with step-up verification while leaving genuine reviewers untouched. Review fraud most affects e-commerce marketplaces and any platform whose ratings and reviews drive purchasing decisions, including app stores, travel and hospitality, and local business directories.
How to implement it
-
Count reviews and accounts linked to each Visitor ID so you can recognize one device posting reviews across many accounts.
-
Block known offenders at the device level by storing the Visitor ID of flagged reviewers, so a blocked device cannot post again from a new account.
-
Use Bot Detection to flag review submissions from automated tools and scripts running review fraud at scale.
-
Use Velocity Signals with linked ID to flag a burst of reviews or accounts tied to one Visitor ID in a short window, a sign of scripted or coordinated review posting.
-
Use Suspect Score to gate suspicious submissions with step-up verification while letting genuine reviewers post without friction.
-
Use Proximity Detection to surface coordinated review operations, where many devices clustered in one location each post reviews that look like separate, distinct customers.
Which signals to check
Also known as
review fraud · fake reviews · review manipulation · rating fraud · review bombing · fake ratings · opinion spam · review spam
Related use cases
-
Survey fraud
Detect duplicate survey submissions from the same device and automated responses from bots that corrupts data and drains incentives.
- Account Integrity
- Automation Risk
- +2
11 signals Marketing, SaaS -
Multi-accounting
Detect when one device operates many accounts to farm rewards, evade bans, or gain unfair advantages, even across cleared cookies, new IPs, and different identities.
- Account Integrity
- Incentive Abuse
- +1
16 signals E-Commerce, Fintech, Gambling, Gaming, Media -
Fraud farms
Surface clusters of co-located devices that signal a device farm or organized fraud ring, even when each device looks unique on its own.
- Automation Risk
- Device Integrity
9 signals E-Commerce, Fintech, Gambling, Gaming