Amazon today said it can’t stop fake product reviews without help from social media companies, and it blamed those companies for not doing more to prevent solicitation of fake reviews.
In a blog post, Amazon said its own “continued improvements in detection of fake reviews and connections between bad-actor buying and selling accounts” has led to “an increasing trend of bad actors attempting to solicit fake reviews outside Amazon, particularly via social media services.”
Some use social media services on their own; in other cases, they hire a third-party service provider to perpetrate this activity on their behalf. However, bad actors regularly try to take this transaction outside Amazon to obscure our ability to detect their activity and the relationship between the multiple accounts committing or benefiting from this abuse.
Amazon wants social media to detect fraud before Amazon reports it
Amazon went on to say that it reports fraud to social media companies and that “some” of the companies have improved their responses to those reports. However, Amazon said it wants social media firms to find and eliminate fraud before Amazon reports it to them.
“While we appreciate that some social media companies have become much faster at responding, to address this problem at scale, it is imperative for social media companies to invest adequately in proactive controls to detect and enforce fake reviews ahead of our reporting the issue to them,” Amazon said.
Amazon did not say which social media companies aren’t meeting its expectations. Facebook is the most likely culprit in part because of its sheer size—just about any problem that happens on social media is most likely to happen on Facebook because it’s the biggest platform.
There’s also a history of Facebook’s platform being used to sell fake reviews. As The Guardian noted today, “the UK’s Competition and Markets Authority secured an agreement from Facebook [in January 2020] to ‘better identify, investigate and remove groups and other pages where fake and misleading reviews were being traded, and prevent them from reappearing.’ However, a follow-up investigation in 2021 forced the CMA to intervene a second time.”
After that intervention, Facebook “removed a further 16,000 groups that were dealing in fake and misleading reviews” and “made further changes to its systems for identifying, removing and preventing such content on its social media platforms to ensure it is fulfilling its previous commitments,” the UK government said in April.
An October 2020 article by The Verge explained how review traders use social media “to evade detection on Amazon. Once a reviewer buys the item, they send the seller a receipt, along with a photo of their review. The seller then sends a refund through PayPal. This ensures that the rating has a ‘verified purchase’ tag on Amazon, cementing its supposed authenticity. Some sellers pay an additional fee, between $2 and $15, on top of the refund.”
“We need social media companies whose services are being used to facilitate fake reviews to proactively invest in fraud and fake review controls, partner with us to stop these bad actors, and help consumers shop with confidence,” Amazon wrote today.
Amazon said there was a big improvement between 2020 and 2021, however. When Amazon reported over 300 fake-review groups to social media companies in the first three months of 2020, the companies “took a median time of 45 days to shut down those groups from using their service to perpetrate abuse,” Amazon wrote. By contrast, Amazon reported over 1,000 such groups in the first three months of 2021, and “social media services [took] a median time of five days to take them down,” Amazon wrote.
Bulk sales of fake reviews
Amazon has struggled to eliminate various types of fraud. A UK consumer group recently posed as an Amazon seller to investigate bulk sales of fake reviews, saying in February that it found “702,000 product reviewers across just five businesses” and “one site selling contact and social media details for Amazon reviewers.”
In September 2020, six people were indicted in the US on allegations of paying over $100,000 in bribes to Amazon employees and contractors as part of a scheme to give third-party sellers unfair advantages on the Amazon marketplace. Last month, Amazon said it “seized and destroyed” over 2 million counterfeit products that sellers sent to Amazon warehouses in 2020 and “blocked more than 10 billion suspected bad listings before they were published in our store.”
Amazon’s blog post today said the company has “filed lawsuits against those who have purchased reviews and the service providers who provided them, and we will continue to pursue these bad actors through the courts.” While the fake-review problem is by no means solved, Amazon also said that in 2020, it “stopped more than 200 million suspected fake reviews before they were ever seen by a customer” and shut down accounts that contribute fake reviews.
Amazon said it uses a mix of “sophisticated technology powered by machine learning” and “expert human investigators” to find fake reviews. Since the fake-review problem originates partly on social media sites outside of Amazon’s control, the company said it also uses machine-learning technology and other techniques “to detect groups of connected entities—customer accounts, selling accounts, products, brands, and more.”