Knowing how to spot fake reviews has become a basic consumer skill, much like checking a contractor's license. A star rating is often the first thing you see about a plumber, dentist or repair shop, and it can decide who gets your call. Yet that number comes from strangers, and some may never have been customers.
The volume is large. Google says it blocked or removed more than 292 million policy-violating reviews in 2025, up from more than 240 million in 2024. Those figures cover every policy violation, so they are not a count of fake reviews alone, and they say nothing about what got through.
Here is the uncomfortable part. Research shows people are poor at judging a single review by reading it, and the Federal Trade Commission warns that it "can often be nearly impossible to tell for sure." So the common advice to look for suspicious wording is weaker than most guides suggest.
Our position is simple. Stop judging individual reviews and start judging patterns. Timing, reviewer history, the spread of ratings and a second source tell you more than any one sentence. This guide explains what each check can and cannot show, what the law says, and how to report a review that looks wrong.
Why Fake Reviews Keep Appearing on Local Listings
Fake reviews exist because ratings move money. A study of Washington State restaurants by Michael Luca of Harvard Business School found that a one-star increase on Yelp led to a 5 to 9 percent increase in revenue for independent restaurants (2011). A separate study of San Francisco Bay Area restaurants found that an extra half star made them sell out 19 percentage points more often (Anderson and Magruder, 2012). Those numbers come from restaurants, but the incentive is easy to see for any local business, particularly one that depends on first-time customers who have never heard of it.
The same logic explains who cheats. Luca and Georgios Zervas used Yelp's filter as a stand-in for fraud and found that roughly 16 percent of restaurant reviews were filtered as suspicious (Management Science, 2016). Restaurants with few reviews or recent bad reviews were more likely to commit fraud, and those facing more competition were more likely to receive unfavorable fake reviews. A filter is an imperfect proxy, so treat 16 percent as an estimate for one site and one industry, not a national rate.
It also helps to separate four problems that often get lumped together.
- Fake reviews describe experiences that never happened, or were written by someone who does not exist.
- Incentivized reviews come from real customers who received something in return.
- Insider reviews come from owners, employees or relatives who hide that connection.
- Suppressed reviews are honest complaints that a business has blocked, buried or scared away.
Each one distorts a rating in a different way, and only the first is a pure invention. A rating can therefore mislead you even when every review on the page came from a real person. That is why this guide speaks of suspect reviews rather than fake ones. From the outside, you can rarely prove which problem you are looking at, and you rarely need to.
What US Law and Google Prohibit
Since October 21, 2024, a federal rule has directly addressed review fraud. The FTC's rule on consumer reviews and testimonials, 16 CFR Part 465, covers fake reviews, including AI-generated reviews attributed to someone who does not exist. In December 2025, FTC staff warned 10 companies about possible violations, noting civil penalties of up to $53,088 per violation. Those letters were warnings, not findings that anyone broke the rule.
"Fake or false consumer reviews are detrimental to consumers' ability to make accurate and informed choices"
That line comes from Christopher Mufarrige, Director of the FTC's Bureau of Consumer Protection, in the same December 2025 announcement.
The rule also reaches beyond the reviews themselves. A business may not suppress criticism through intimidation or baseless legal threats, and it may not pose a site it controls as an independent review source. It also may not suggest that the reviews on its own site represent most or all of those submitted while hiding the negative ones based on rating.
Google's rules go further than federal law in one respect. The FTC rule allows incentives as long as they do not require a particular sentiment. Google's Maps policy bars any reward for posting a review, along with reviews based on a conflict of interest and selective requests aimed only at happy customers.
| Practice | FTC rule | Google Maps policy |
|---|---|---|
| Buying or writing fake reviews | Prohibited | Prohibited |
| Reward tied to a positive review | Prohibited | Prohibited |
| Reward for any honest review | Allowed if no sentiment is required | Prohibited |
| Employee review with no disclosure | Prohibited for covered insiders | Prohibited as a conflict of interest |
| Asking only happy customers | Not named in the rule, though the FTC Act may apply | Prohibited as selective solicitation |
| Threatening a reviewer into silence | Prohibited | Prohibited as discouraging negative reviews |
Why Reading a Single Review Rarely Works
Most advice on this topic tells you to watch for gushing language, and the evidence for that is thin. In a 2011 Cornell study, Myle Ott and colleagues built a test set of 800 hotel reviews, half real and half deliberately fake. Three human judges scored between about 53 and 62 percent on a sample of them, which the authors described as roughly at chance. A trained computer model reached about 90 percent, which shows that the differences exist but sit below what people can notice.
Generative AI makes this harder. Balázs Kovács of Yale ran two experiments published in Marketing Letters in 2024 and found that people could not reliably tell AI-written reviews from human ones, even when paid for accuracy. A detection tool he tested was also fooled.
Vague praise, then, is a reason to keep reading rather than a verdict. Plenty of honest customers write "great service" and move on, while a skilled faker can invent a technician's name and a specific repair. Specific details still make a review more useful to you, because they tell you what actually happened. They just cannot prove it happened.
Patterns That Separate Normal Reviews From Suspect Ones
A single review can fool you. A pattern is harder to fake, because it takes coordination across many accounts over time.
Timing
Sort reviews by newest and scan the dates. The FTC advises consumers to watch for a burst of reviews over a short period, and Google treats unusual volumes or patterns as rating manipulation. A burst is stronger evidence when the older reviews describe a very different business.
Reviewer history
Open a few reviewer profiles. The FTC notes that an account created just to write one review may be fake. Look for repetition across the group, such as many accounts with one review each, or accounts that praise businesses in far-apart cities within days.
Rating mix
Luca and Zervas found that filtered reviews were more extreme, in both directions, than other reviews. Be careful with this one, because honest reviews are polarized too. A study of over 280 million reviews on 25 platforms found most platforms lean toward extremes, largely because people with strong opinions are the ones who bother to write (Schoenmueller, Netzer and Stahl, 2020).
Repetition across reviews
Compare several reviews side by side. Identical phrasing, the same unusual spelling or the same selling points in the same order are hard to explain as coincidence.
Fit with the business
Read a few reviews against what the business actually does. Praise for a service it does not offer, or for a location it does not have, suggests the author never went there. Testimonials on a company's own website deserve a lighter weight in general, because the business chooses what appears there.
| Signal | Why it raises concern | Innocent explanation |
|---|---|---|
| Sudden burst of five-star reviews | The FTC lists bursts as a warning sign | The owner just started asking customers |
| Reviewer with one review ever | Accounts may be created for a single job | Many people only review once |
| Nearly identical wording | Suggests one author or a shared script | Rare, though short phrases do repeat |
| Reviews split between 5 and 1 stars | Fraud tends to be extreme | Honest reviews are polarized too |
| Perfect 5.0 on many reviews | Spotless records are unusual | A small, careful business can earn it |
Fake Negative Reviews Point the Other Way
Fake reviews can hurt as well as flatter. The FTC notes that a company might post fake negative reviews to harm a competitor, and the Luca and Zervas findings suggest competition raises that risk. Google has also described attempts to demand payment in exchange for removing fake one-star reviews, a form of extortion aimed at business owners.
So a cluster of angry reviews deserves the same skepticism as a cluster of glowing ones. Check whether they mention anything specific, whether they arrived together, and whether the reviewers have any history. A calm, factual reply from the owner can also help you judge. An owner who says the person was never a customer, and explains how that was checked, gives you something to weigh. An owner who attacks the reviewer gives you something else to weigh.
What Platforms Catch and What Gets Through
Platforms remove a great deal, but timing matters. Researchers Sherry He, Brett Hollenbeck and Davide Proserpio tracked about 1,500 Amazon products that solicited fake reviews and found that roughly half of the reviews were eventually deleted, with an average lag of over 100 days (Marketing Science, 2022). The rating boost faded after about a month, and ratings fell once sellers stopped buying. That is one marketplace, but it shows why a clean listing today is no guarantee.
Google says it now revisits reviews to find new abuse patterns months after posting, and during spam spikes it may pause new reviews on a profile and show a notice. If you see that notice, read the remaining reviews with extra care.
The business itself can be the fake. Google says it removed or blocked more than 12 million fake Business Profiles in 2024. Before you rely on any reviews, confirm that the business has a working website, a phone number that someone answers and an address that matches what you see on the map.
Scores built from public ratings inherit the same limits. American Business Recognition publishes this blog and runs a recognition program based on Google ratings, and our published formula uses rating and review count as inputs. It has no term for authenticity, so it should support your own judgment, never replace it.
A Five-Minute Check Before You Call
You do not need special tools or paid software. This routine works for a mechanic, a roofer or a dentist, and it takes about as long as making coffee. The goal is a rough judgment, because that is all the public evidence allows.
- Sort by newest. Scroll through the last 20 or 30 reviews and look at the dates. Note any burst, especially after a quiet period, and whether recent reviews match the older ones.
- Open five reviewer profiles. Pick a mix of positive and negative reviews. You are looking for repetition across accounts, not one odd profile, and a few minutes here beats reading twenty more reviews.
- Read the low ratings first. The two-star and three-star reviews are often the most useful. Ask whether the worst realistic outcome is something you can live with.
- Compare a second source. The FTC advises checking several sources. Search the business name with the word "complaint," then look at another review site or ask a neighbor. A large gap deserves a question.
- Test with a small job. Where you can, start with an oil change or a minor repair before you commit to something larger. Our guide to choosing an auto repair shop applies this idea in detail.
If something looks wrong, report it. On Google Maps, open the review, choose the three-dot menu and select "Report Review". If you suspect paid reviews, the Maps app has a separate path under Suggest an edit, then Report business conduct. The FTC also asks consumers to tell it at ReportFraud.ftc.gov and to include "fake review" in the comments.
One caution applies to all of this. Weigh the overall picture, and do not accuse a business publicly based on one signal. You may simply be looking at an owner who finally started asking for reviews.
How AI and Enforcement Will Change Review Fraud
Two trends are pulling in opposite directions. Generative tools have made convincing text cheap, and the Kovács experiments suggest that readers cannot be counted on to catch it. That weakens language as a signal, and it will probably keep weakening.
Enforcement is moving the other way. The FTC now has a rule that allows civil penalties for knowing violations, and its December 2025 letters show that staff are applying it to real companies. Google's reported totals rose from 2024 to 2025, though that alone cannot tell us whether fraud grew or detection improved. Both may be true.
For consumers, the practical shift is toward signals that are costly to fake. A review's text costs almost nothing to produce. A long history of steady reviews, a track record across several independent sources and a verifiable reviewer history are far more expensive to manufacture.
We expect the most useful habits to stay simple. Readers who check timing and second sources will keep an edge, whatever tools the fakers adopt. Platforms will likely keep adding automated checks of their own, and some fraud will keep slipping past them for months at a time. No checklist will catch everything, and anyone who promises a foolproof method is overselling it.
Patterns Beat Instinct When Reading Reviews
The evidence points to one clear conclusion. You are unlikely to catch a fake review by reading it closely, and neither is almost anyone else. Controlled studies put human judges near chance, and AI-written text has made that harder still.
What you can do is look at the shape of a business's reviews. Check when they arrived, who wrote them, how they are spread across ratings and whether another source tells the same story. None of those signals is proof. A burst may be a new review routine, and a single-review account may be an honest customer who never reviewed anything else. Several signals pointing the same way, however, should change your decision.
Remember that distortion runs in both directions. A business can inflate its rating, and a competitor or scammer can drag one down. Federal rules and platform policies now address both, yet removals can take months, so a listing that looks clean may still contain reviews that will later disappear.
Spend five minutes on the check above before any purchase that matters. If the pattern looks wrong, keep looking, then report what you found to Google and the FTC. It takes little effort, and it helps the next customer who reads the same page.
Questions People Ask About Fake Reviews
Are fake reviews illegal in the United States?
Yes. The FTC's rule on consumer reviews and testimonials took effect on October 21, 2024, and it bans writing, buying or selling fake reviews. Courts can impose civil penalties for knowing violations. The FTC notes that the rule gives consumers no private right to sue under it.
How can you tell if a Google review is fake?
Usually you cannot be certain from one review. Look at patterns instead, including bursts of reviews, many single-review accounts and repeated wording across reviewers. The FTC itself says it can be nearly impossible to tell for sure, so treat your conclusion as a judgment rather than a finding.
Can a business pay customers to leave reviews?
Under the FTC rule, a business may offer an incentive only if it does not require a particular sentiment. Google is stricter and bars rewards for posting a review on Maps at all. A business that follows the federal rule can still break Google's policy.
Does Google remove a review because it is negative?
No. Google's help pages tell users not to report content simply because they dislike it. A review has to break a policy to be removed, and an assessment can take several days.
Is a perfect 5.0 rating a red flag?
Not on its own, especially with few reviews. Research from Northwestern's Spiegel Research Center found that purchase likelihood typically peaks at ratings from 4.0 to 4.7, then falls toward 5.0, which suggests buyers are already wary of perfection. For context, see what counts as a good Google rating.
Can AI-written reviews be detected?
Not reliably by reading them. In 2024 experiments, people could not consistently separate AI-written reviews from human ones, and a detection tool also failed. Reviewer history and timing stay useful, because they do not depend on the wording. AI text can also count as a fake review under the FTC rule when it is attributed to someone who does not exist.
Can you trust reviews on a company's own website?
Treat them as marketing. A business chooses which testimonials to publish, so they rarely show the full picture. The FTC rule bars a business from implying that its displayed reviews represent most or all of those submitted when it hides the negative ones. Outside review sites give you a useful comparison.
This article is general information and is not legal advice.
