Explainers / FRAUD
How can you tell if a business's reviews have been manipulated?
You usually cannot prove it from outside, but you can spot the patterns that regulators and platforms themselves describe: sudden surges, reviews that arrive too fast, reviewers with a shared trail, and text that does not match the business. Treat each as a reason to look closer.
Most advice on spotting fake reviews is folklore: too many exclamation marks, generic names, stock-photo avatars. Some of it may even be right. We have limited this page to signals that an enforcement agency, a platform or a published study has put in writing, and we say for each one what else could explain it.
Signals the FTC names
In its questions and answers on the review rule, FTC staff describe what should put a business on notice that reviews it obtained are fake. The same signs serve a reader:11 Federal Trade Commission, Consumer Reviews and Testimonials Rule: Questions and Answers
- Reviews appearing too quickly after purchase. A detailed review of a mattress posted the day it was delivered.
- A sudden surge in volume. Two reviews a month for three years, then thirty in a week.
- Reviews that refer to the wrong product. Praise for the "delicious food" at a tire shop.
The innocent explanations: a business started using a review-request tool (surge), a customer confused two businesses (wrong product), an eager customer (speed). One sign is noise. Two on the same profile in the same fortnight is a pattern.
Signals Google names
Google's descriptions of its own moderation list what its systems look for.22 Google, The Keyword, How reviews on Google Maps work (February 2022)33 Google, The Keyword, How machine learning keeps contributed content helpful
- "An unusually high number of 1 or 5-star reviews over a short period of time." Applies to attacks as much as to boosting.
- "A group of people leaving reviews on the same cluster of Business Profiles." Click three or four reviewer names. If the same handful of people have all reviewed the same locksmith in Phoenix, the same dentist in Tampa and the same moving company in Portland, you have found a ring.
- "A reviewer posting the same review on multiple businesses." Identical or near-identical text across profiles.
- Account history. A profile with one review ever is not proof of anything; most people review rarely. A page where nearly every five-star reviewer has exactly one review is different.
Signals Yelp names
Yelp's consumer alerts are tied to specific findings, and its 2025 report says which were most common. Two thirds of its compensation and suspicious-activity alerts involved many suspicious positive reviews "coming from the same IP address".44 Yelp Trust and Safety, 2025 Trust and Safety Report You cannot see IP addresses, but you can see the result when Yelp acts: a pop-up over the business page.55 Yelp Trust and Safety, Consumer Alerts If an alert is showing, read it. Yelp says it conducts an investigation before placing one.
Yelp's list of reasons a review might not be recommended also doubles as a description of what unreliable looks like to its software: less active users, rants, shared IP addresses and reviews "written by a friend of the business owner".66 Yelp Support Center, Why would a review not be recommended?
Signals from research
Extremity. In the Yelp data studied by Luca and Zervas, filtered reviews "tend to be more extreme (favorable or unfavorable) than other reviews".77 Luca and Zervas, Management Science 62(12), 2016, Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud Fakes cluster at one and five stars because a three-star fake buys nothing.
The caution here is strong. Honest reviews also pile up at the extremes; a 25-platform study found most review distributions are polarized because people with strong views are the ones who write.88 Schoenmueller, Netzer and Stahl, Journal of Marketing Research 57(5), 2020, The Polarity of Online Reviews: Prevalence, Drivers and Implications Extremity separates fake from real only weakly.
Weak reputation, strong motive. The same study found fraud more likely when a business "has few reviews or it has recently received bad reviews". Apply that to timing. A run of five-star reviews that begins days after a damaging one-star review is the classic repair job.
Competition. Businesses facing more local competition were more likely to receive unfavorable fakes.77 Luca and Zervas, Management Science 62(12), 2016, Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud A burst of one-star reviews from accounts with no history, hitting an established business just as a rival opens nearby, deserves the same skepticism as a burst of fives.
A ten-minute audit
This is the sequence we would run on a profile before a large purchase.
- Sort by newest. Scroll to find the review rate. Count reviews per month over the past year. Flag any month that is several times the usual rate.
- Open the reviewers in the flagged month. How many have a single review? How many share other reviewed businesses with each other?
- Read for specifics. Real customers mention what was done, when, by whom and what it cost. A writer who was never there has to stay vague or invent the details.
- Check the middle. A profile with 300 fives, 12 ones and nothing at two, three or four stars is unusual even by the polarized standards of review sites. It suggests either gating or fabrication.
- Look for the business's replies. Businesses that answer criticism in detail are harder to square with a fabricated profile, though this is our judgment and not a research finding.
- Cross-check one other platform. Fraud is expensive to run everywhere at once.
- On Yelp, open the not-recommended reviews. It is the only major platform that lets you see what its filter set aside.
What does not work
Reading tone. The platforms themselves do not rely on prose style, and fluent machine-written text gives readers even less to go on; Yelp says it filtered out nearly 500,000 suspected AI-generated reviews in 2025 using automated systems, not human readers.44 Yelp Trust and Safety, 2025 Trust and Safety Report If you find yourself judging a review by how it sounds, switch to judging the account and the timing.
Trusting the count alone. A thousand reviews are harder to fake than ten, but large counts can be built by gating, which uses real customers. Volume answers the small-sample problem and leaves the selection problem untouched.
If you think a profile is manipulated
Report it to the platform using the flag on the review; Google and Yelp both describe human review of flagged content.22 Google, The Keyword, How reviews on Google Maps work (February 2022) If you were offered payment for a review on Amazon, the company asks to be told.99 Amazon Customer Service, Understanding Customer Reviews and Ratings Then make your decision as if the rating did not exist: references, license lookup, a small first job.
Questions readers ask
Does a perfect 5.0 mean the reviews are fake?
No. With few reviews, a perfect score is statistically ordinary. See is a perfect 5.0 suspicious for the arithmetic.
Are one-review accounts always suspicious?
No. Many real people write one review in their lives, prompted by one memorable experience. The warning sign is concentration: a profile where the positive reviews come overwhelmingly from single-review accounts created around the same time.
Method The checklist is restricted to signals named in FTC staff guidance, platform moderation pages or peer-reviewed research. Folk rules we could not source are left out. See how we source and check.
Sources
Every source below was opened and read for this page. Links go to the publisher's own copy.
- Federal Trade Commission. Consumer Reviews and Testimonials Rule: Questions and Answers
- Google, The Keyword. How reviews on Google Maps work (February 2022)
- Google, The Keyword. How machine learning keeps contributed content helpful
- Yelp Trust and Safety. 2025 Trust and Safety Report
- Yelp Trust and Safety. Consumer Alerts
- Yelp Support Center. Why would a review not be recommended?
- Luca and Zervas, Management Science 62(12), 2016. Fake It Till You Make It: Reputation, Competition, and Yelp Review Fraud
- Schoenmueller, Netzer and Stahl, Journal of Marketing Research 57(5), 2020. The Polarity of Online Reviews: Prevalence, Drivers and Implications
- Amazon Customer Service. Understanding Customer Reviews and Ratings