Amazon admits it has a fake review problem, but does its best to spread the blame around in a new post detailing the issue. After numerous reports for years that the online retail giant is overrun with knock-off products and faked or farmed reviews, the company aims to look as if it is finally putting its foot down, but no new efforts or rules are discussed — rather, it is others that need to step up their work to keep Amazon safe.
Amazon reviews have become notoriously unreliable as indicators of quality as the store has given itself over willingly to counterfeits, AliExpress resellers, and promotion of the company's internal brands (developed with the benefit of seller data). Multiple reports have found organized efforts to spam the store with meaningless 5-star reviews in exchange for free products or cash. I have myself received such offers, or sellers promising payment for raising a star rating.
After the requisite preliminary palaver about being “obsessed with delighting customers” and all that, Amazon explains that it, like all big tech giants, uses automated systems to vet reviews before they go up. The company has always been cagey about the actual numbers, but in this post it drops a whopper: “In 2020, we stopped more than 200 million suspected fake reviews before they were ever seen by a customer.”
200 million is a lot no matter how you look at it, but it's really a lot when you consider that Amazon told CNBC that same year that it “analyze[s] over 10 million review submissions weekly,” which adds up to somewhere north of 520 million submissions yearly. These two Amazon-provided numbers suggest that a third of all reviews submitted, at a minimum, are rejected as fake.
Hard numbers on Amazon's total reviews are hard to come by. Speaking to Buzzfeed, Amazon listing analysis site ReviewMeta's Tommy Noonan estimated that in 2020 Amazon hosted around 250 million reviews (of which, incidentally, he calculated about 9 percent were “unnatural”). But if over 500 million were submitted in 2020 and about 200 million of those were fake, that indicates a far larger total. I've asked Amazon for more precise information on this, and will update the post if I hear back, but the company is not communicative about these numbers in general.
Groups organizing on social media numbering in the tens of thousands have been repeatedly pointed out as major contributors to the fake review ecosystem. Amazon writes that in the first quarter of 2020, it reported 300 such groups to the platforms hosting them, and the same period in 2021 it reported over 1,000. Takedown times have increased, but it's hard to take this increase as anything other than a thriving business model — certainly not something in the process of being stamped out.
“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 declares. Indeed social media companies are being pressed from multiple directions to take more responsibility for what users do on their platforms, but they make the same noises Amazon does: “we're doing what we can,” (and, it is left unsaid, clearly it's not enough).
“We need coordinated assistance from consumer protection regulators around the world,” Amazon writes. But the company lobbied forcefully and successfully against the INFORM act, which would have helped identify sellers with bad intent and add transparency to online marketplaces (strangely, Amazon has taken some of the actions it objects to independently). And that line is suspiciously absent when it is Amazon that consumers need to be protected against.
“It is also critical that we hold bad actors—and the service providers that provide them with fake reviews—accountable for their activity,” the post continues. But while lawsuits and partnerships with law enforcement are part of that, once again the call to “work together” rings hollow when Amazon itself is where the activity occurs and the company is in complete control of that ecosystem. Though it has banned some major players from the store, innumerable others flout the rules with impunity.
Nowhere in the post does Amazon detail any new steps it will take to deter these bad actors or crack down on the pervasive gaming of the system for which it sets the rules. It will “continue to enhance” its detection tools, “streamline processes” for partnerships, and “work hard” at keeping scammers accountable. In other words, it will keep doing exactly what it has been doing this whole time — which is what put it in this position in the first place.