Inside the War on Fake Consumer Reviews

13 minute read

When Kay Dean turned her keen detective’s eye on the reviews for a Toronto dental clinic, she quickly discovered something alarming. The reviewers behind the clinic’s glowing testimonials had also reviewed an array of obscure small businesses across North America and Europe. A dry cleaner in Florida, a locksmith in Texas, an electrician in the UK, and more: a handful of reviewers recommended the same 14 companies. Even more tellingly, many reviews appeared word-for-word on both Facebook and Google under different names.

In a video on her consumer watchdog YouTube channel Fake Review Watch, Dean, a former federal criminal investigator, explains that she became leery of the dental practice when she spotted that 195 of its 202 Facebook reviews were posted over just three days in May. Previously, the clinic had gone years with only a handful of reviews.

“Given what I’ve just shown you here,” Dean says in the video. “Would you trust any reviews for this Toronto dental practice? I wouldn’t.”

The problem of fake positive reviews extends far beyond what Dean manages to flag in her YouTube channel.

As the COVID-19 pandemic continues to supercharge global e-commerce, this kind of fraud has emerged as one of the most significant factors contributing to an erosion of consumer confidence in the online marketplace. Fake reviews influenced around $152 billion in global spending on lackluster products and services last year, according to a report from the World Economic Forum. The world of online reviews is so overrun with fakes, it’s become like the Wild West, Dean says. “There’s no repercussions,” she says. “There’s no penalties. Cheating is rewarded in the current environment.”

Dean created Fake Review Watch after having her own harrowing ordeal five years ago. She chose a psychiatric practice based on reviews she read on Google and Yelp, but then had a very negative experience there that left her upset and extremely frustrated about being manipulated by the reviews. “We’re not just talking about five-star ratings,” she says. “We’re talking fake stories about people feigning various mental illnesses and claiming to be helped by the practice.”

Since then, she’s been devoted to exposing the various ways that businesses try to dupe consumers by manipulating reviews on their own or employing the services of shady organizations that sell fake reviews en masse (sometimes referred to as paid-review farms).

In her videos, Dean, wearing her signature glasses and dangle earrings, calmly talks viewers through her process for determining a business is allegedly faking reviews. “My aim is not necessarily to dime on these businesses, but you have to show examples to make it real so people can see what’s going on,” she says. “I’m just trying to get the word out as a consumer advocate.”

Kay Dean, seen here in a recent video on her “Fake Review Watch” YouTube channel, is a former federal investigator.TIME

Dean’s ongoing fight to publicize the growing threat of fake reviews coincides with renewed governmental efforts to tackle the issue. The U.K. government announced new plans in April to quell fake reviews, with the U.S. Federal Trade Commission (FTC) following suit in May by proposing to update related guidelines to bring them up to speed with the modern, digitized economy. Meanwhile, major review platforms like Amazon, Google, Meta, TripAdvisor, Trustpilot, and Yelp say they’re making their best attempts to keep deceitful reviews off their sites. Meta, for example, will try harder to curb fake reviews in the U.S., according to a new policy announced on June 20. These are steps in the right direction. But experts say the battle against fake reviews can’t be won without turning up the heat to full blast.

How we got here

The pandemic forced a less-experienced cohort of online shoppers to do more of their purchasing on the internet. These shoppers came to lean on reviews more heavily, says Adrian Palmer, head of Henley Business School’s department of marketing and reputation at the University of Reading. “Many consumers had to learn what indicators they should use to make it easier to buy online, rather than going to a shop,” he says. “Instead of picking up a product, touching it, feeling it, smelling it, they were confined to a two-dimensional screen and had to use all the tools they could to make a decision. Reviews were one of those.”

It makes sense, then, that when fake review schemes come into play, studies show they can and do pay off for businesses who employ them. A January survey by SEO platform BrightLocal found that 77% of consumers regularly consider reviews when making a purchase and 75% say that reading a positive review is the top factor that makes them feel good about a business. Fake reviews corrupt the market by restricting access to “free, fair, and proper information,” says Palmer. “If you’ve got a market where information that you think is coming from a buyer is actually coming from a seller, the market doesn’t work.”

Paid-review farms have sprung up around the world, providing their services to businesses paying to falsely boost their ratings on review platforms. “A lot of times businesses unwittingly will hire these firms because they try to position themselves as online marketing and reputation firms,” says Brian Hoyt, TripAdvisor’s head of global communications and industry affairs.

Highly-organized, largely off-shore fake review networks frequently use sites like Facebook, Craigslist, and Reddit to advertise and sell their deceptive practices to businesses. Dean says there’s a “tremendous amount” of review fraud being facilitated on social media.

“Groups that solicit or encourage fake reviews violate our policies and are investigated and enforced upon appropriately,” a Meta spokesperson says.

Other businesses take the matter into their own hands by offering incentivized reviews, coercing employees to write fake reviews, or otherwise influencing reviews to create a false impression. Dean says that review fraud extends across every profession you can imagine. “I’ve personally seen doctors, lawyers, dentists, contractors, wedding DJs, piano teachers, lactation consultants,” she says. “And I’m a single investigator using no automated tools, just my eyeballs and spreadsheets.”

Situations involving fake reviews get particularly dicey when consumer safety is on the line, says Beibei Li, an associate professor of IT and management at Carnegie Mellon University. “If you go to a doctor with a lot of fake reviews and you don’t recognize that, it could affect your health,” she says. “That’s definitely a higher risk scenario compared to going to a restaurant for dinner where, worst case, you just get a bad meal.”

It’s an issue the FTC considers to be one of the principal problems with online marketing. “There are thousands and thousands of fake reviews being generated almost on a daily basis,” says Richard Cleland, assistant director of the FTC’s division of advertising practices. “This really infects the entire commercial space of the internet.”

What review platforms are doing

In February 2021, U.K. consumer advocacy group Which? published an investigation into how fake reviews for products available on Amazon Marketplace were being sold online in bulk. It found that companies set up for the sole purpose of flooding Amazon sellers’ product listings with phony praise were fueling a huge global industry of coordinated online reviews. German company AMZTigers, for instance, was reportedly advertising bulk packages of up to 1,000 reviews for prices close to $10,000. Amazon attempted to crack down on fraudulent sellers later that year and suspended multiple high-profile sellers for using banned techniques to get reviews.

An Amazon spokesperson says that the company has “clear policies” that prohibit review abuse and that it suspends, bans, and takes action against those who violate those policies. “Amazon receives more than 30 million reviews each week, and uses a combination of machine learning technology and skilled investigators to analyze each review before publication,” the spokesperson says. “Each review provides us with dozens of pieces of information to detect suspicious activity. We can cross-reference this information to identify fraudulent patterns and potential abuse.”

In 2019, TripAdvisor became the first global review platform to begin putting out a transparency report outlining the steps it takes to fight fraud. The most recent report, released in 2021, stated that, of the 26,000,000 reviews submitted to the site in 2020, 943,205 reviews (about 3.6%) were determined to be fraudulent. Of those fake review submissions, Tripadvisor said it prevented 67.1% from ever making it onto the platform.

Hoyt says the travel booking site moderates user-generated content by supplementing the skills of its team of investigators with fraud detection technology that maps hundreds of discrete pieces of online information to track factors like review origin and potential connection to other reviews. Reviews must be recent, relevant to travel experiences, non-commercial, first-hand, respectful, and unbiased to meet TripAdvisor’s criteria. “It’s a race towards perfection, because no one in this space will ever truly be perfect. But we’re constantly trying to stay one step ahead of the fraudsters,” he says.

TripAdvisor’s transparency report states that it’s put a stop to the activity of over 120 different paid-review farms around the world in recent years. “We’ve been really successful in chasing down these companies that pop up like Whack-a-Mole advertising, ‘Hey, pay us 20 bucks a month or whatever and we’ll give you x, y, or z number of reviews to boost your business,’” Hoyt says.

In its own transparency report, Trustpilot, an online review community that connects businesses and consumers, said it removed 2.7 million fake reviews from the almost 50 million reviews it received in total last year (about 5.8%). Anoop Joshi, Trustpilot’s vice president of legal and platform integrity, says the platform’s automated systems analyze factors like text content, IP addresses, VPN use, email addresses, and patterns of behavior to determine the authenticity of reviews. “If we see a review that’s being submitted for a business that’s based in the U.K. and then 10 minutes later, we see one [from the same reviewer] for a business based in the U.S., that’s a signal that there may be some untoward behavior there,” he says.”

On Yelp, automated software identifies and recommends what the platform views as the most reliable reviews. Yelp says about 4.3 million of the approximately 19.6 million reviews contributed to the site in 2021 were not recommended by its software. Not recommended reviews are still visible to consumers on the bottom of Yelp business pages.

“Yelp invests in both technology and human moderation to mitigate misinformation on our platform,” a Yelp spokesperson says. “Our approach is driven by our automated recommendation software, reporting by Yelp’s community of business owners and users, human moderation, and Consumer Alerts.”

The Consumer Alerts program places warnings on the pages of businesses that have been flagged for suspicious review behavior to inform consumers about the violation of Yelp’s policies.

In a March blog post, Google said it put protections on over 100,000 businesses in 2021 after detecting suspicious review activity and abuse attempts using a combination of machine learning and human operators. According to the company, its machine learning models have been trained to watch out for specific words and phrases, examine the types of content an account has contributed in the past, and detect suspicious review patterns.

“Our policies clearly state reviews must be based on real experiences, and when we find policy violations, we take swift action ranging from content removal to account suspension and even litigation,” Google says. “We’ll continue to work closely with government officials to share more on how our technology and review teams work to help users find relevant and useful information on Google.”

What’s next in the fake review fight

Despite these measures, fake reviews have proven to be a particularly thorny issue—especially as rising inflation contributes to an increasingly fraught economic environment.

Dean contends her research shows that major review platforms often aren’t adequately looking out for consumers and honest businesses. She says they need to do more to alert consumers that a business has acted deceptively when they remove fake reviews. “Showing the public evidence of widespread fraud on their platforms is not in their business interest,” she says.

Now, legislators and regulators are moving to take the wheel, hoping to force businesses to toe the line and review platforms to step up their enforcement efforts.

In the U.K., government reforms put on the table in April would make it “clearly illegal” for businesses to pay someone to write or host a fake review for a product or service, while sites hosting consumer reviews would have to take reasonable steps to check they are genuine. Britain’s competition watchdog, the Competition and Markets Authority (CMA), would also have the power to fine businesses up to 10% of their global turnover for misleading consumers.

In May, the FTC announced that it was proposing changes to its Endorsement Guides to tighten rules around reviews and better reflect what’s going on in the market today.  ”The guidelines haven’t been updated since 2009,” Cleland says. “A lot has happened in the last 10 years. With the rise of influencer marketing and the dominance of digital retail sales, we felt it was time to address current marketing trends.”

The announcement came just months after the FTC settled a complaint against Fashion Nova in January alleging that the fashion retailer blocked negative reviews from appearing on its website. The case marked the first time the FTC had challenged a company for alleged review suppression.

These new guidelines are intended to serve as a warning to businesses seeking to bend the rules. Meanwhile, calls to overhaul Section 230, which shields digital platforms from lawsuits over content posted by their users, are also ramping up. Dean says this type of legislative reform would force review platforms themselves to do much more to self-police. “These platforms are getting a free pass under the current legislation,” she says.

Upping the ante

The scope of the nefarious underworld of review fraud demands more than scattershot retaliatory action.

Due to the nature of the fake reviews industry, Cleland says it’s difficult to enforce measures put in place against it. “The principal challenge is that much of this review brokering activity is being generated outside of the United States,” he says of paid-review farms. “Many of the targets that we’ve identified are either located somewhere in Europe or Asia. And those are difficult for us to address.”

He adds that it can be tough to locate even smaller scale fake review brokers because “you don’t need anything but a computer and a coffee shop.”

To Li, one solution to this is putting more of the onus on the review platforms. “Right now, there’s a very low cost of being caught,” she says. “So there’s potentially two levels of discouragement for these efforts. One is putting pressure on the platforms from a legislative perspective. The other is that the platforms themselves should institute information transparency policies to put pressure on fake review writers. Like, ‘If you do this, we’re going to publicly shame you to consumers.'”

Dean agrees that review platforms need to have “more skin in the game” to stop fake reviews from taking over. “If I can singlehandedly uncover the shocking amount of fraud I have using just publicly available information,” she says. “They should certainly be doing more,” she says.

More Must-Reads from TIME

Write to Megan McCluskey at megan.mccluskey@time.com