Amazon Fake Reviews: 5 Reasons You’ll Never Get Rid of Them Completely

Don’t mistake me: fake reviews are a cancer in the consumer review system.

In theory, online consumer reviews are an ideal engine for a free market system. Consumers receive goods or services from companies. They can then take immediately to online review sites like BestCompany.com, Angie’s List, or Yelp to log their experience. Companies can then see this feedback, make needed changes, and then engage consumers with a new, improved product, service, and/or experience. Other consumers also read these reviews and optimize their purchasing decisions based on the feedback.

It all forms a nice, tidy feedback loop that is bound to improve the lives of consumers in perpetuity and weed out bad companies.

To some degree, in the real world, this system is working, with 67% of consumers now saying their purchases are influenced by online reviews and 54% admitting that online reviews are an important part of their decision-making processes.

It’s a thing of beauty really, something worth protecting from corrupting influences like fake reviews, which have a tendency to steer consumers to undeserving companies, hurt more deserving businesses, and, in turn, damage consumers’ faith in reviews.

Enter Amazon, one of the first heavies to really attack the issue of fake reviews head on. Starting in April 2015, the online retail giant began launching lawsuits against thousands of sellers and buyers of fake reviews on their site. Their mission: to eradicate fake reviews from their site and restore faith in their review system.

But is such a feat really possible? As noble and correct as their intentions might be, getting rid of all fake reviews, even just enough to make their review system trustworthy again, might be impossible — highly unlikely, at least. Here are five inescapable reasons why:

1. There are just too many reviews

To understand what kind of problem Amazon is dealing with here, you have to understand the volume of reviews involved in this undertaking. Amazon is likely the biggest review site in the world. Unofficial estimates say the site experiences 35 orders per second. Approximately 5% of all of those purchases is acknowledged with a review.

Whitney Gibson at the Internet Defamation Attorneys blog estimates that, as of 2015, Amazon had amassed around 90 million cumulative reviews.

Now imagine sifting through 90 million reviews — a number which has surely grown since 2015 and will continue to grow by approximately 50,000 per day — and attempting to verify the trustworthiness of each one. Like I said, maybe not impossible, but certainly unlikely, even for a company of Amazon’s means.

2. It’s hard to identify fake reviewers

How do you determine with certainty who is a fake reviewer and who is not? Certainly, some fake reviewers will be identified by their offers on sites like Fivrr, but what about the others? As much as Amazon wants to sift out fake reviewers, they also want to be careful to not cut out some genuine reviewers in the process. How do you tell the difference, especially across such a large volume of reviews?

Some, like Fakespot, claim to be able to code their way to identifying fake reviewers. For example, based on patterns seen across fake reviews that have actually been discovered, Fakespot looks for similar patterns in reviews and reviewers, according to their site:

“Every fake reviewer has patterns. And the more data we collect via analyses completed, the more our engine is able to adapt and learn.”

While such a program could, over vast numbers of reviews, identify reviewers who are highly likely to be fakes, it wouldn’t be capable of actually verifying which ones are fakes, according to Dr. Bing Liu, a professor of sentiment analysis and opinion mining and the University of Illinois at Chicago:

“The problem with this task is that there is often no hard proof that the detection is actually correct unless the author of the actual fake reviews (not made up fake reviews) from a review hosting site confirms it.”

3. It’s hard to define what exactly a fake review looks like

As mentioned above, there are certainly patterns across some fake reviews, but how to determine for sure which ones are indeed fake?

If they wanted to, Amazon could adopt a hardline stance and delete all reviews that were flagged as possibly (but not verified as) fake — in essence, throwing out the baby with the bath water. But such a move could potentially do as much damage to the validity of their review system as allowing fake reviews to stay.

Instead, Amazon has chosen a less aggressive approach. According to a company spokesperson, “Reviews that were received prior to this policy change are being retroactively removed only if they are excessive or do not comply with the then policy.”

In a recent Yahoo Finance article, writer Ethan Wolff-Mann pointed out the problem with this statement:

“The word ‘excessive’ gives Amazon some latitude, making it unlikely that many reviews will be pulled; many for-profit reviewers are careful — they give five stars to skew the ratings and ranking in Amazon search results yet write neutrally and critically in the text itself.”

What exactly makes a review excessively fake? Is it too much hyperbole about how great the product is? Is it the presence of too many solid five-star ratings? Is it that the reviewer absolutely loves every product they leave a review about? What if the reviewer is just really easy to please?

In this move, it seems, Amazon has pulled the teeth out of its anti-fake reviews crusade and created a massive loophole for fake reviewers to crawl through. They have, quite possibly, doomed their efforts to being regarded as symbolic only and not a true threat to be feared by fake reviewers. Certainly not something grand enough to fix their broken review system.

4. ‘Fake’ is par for the course on the Internet

Since the Internet’s beginnings, the anonymity offered therein has created a breeding ground for fakers of all kinds. As often as this anonymity invites people to be honest — in, say, online reviews — it also invites people to lie and deceive, simply because they can. Expecting the denizens of the Internet to act any differently could be excessively optimistic.

In the cynical but true words of online marketer Ryan Stewart at Webris:

“It’s not slander — it’s the internet. It’s full of trolls (I work here, I would know).”

5. Eradicating fake reviews might not be worth it

Given all the work required to actually rid the world of fake reviews, is it actually worth it? Surely, there is a moral argument to be made about the need to have trust in the consumer review system, but is it worth the resources that will need to be expended? Or do the numbers and nuances involved make this a problem that can never be chased down and defeated?

Perhaps, the best line of defense against fake reviews is consumers themselves. Today’s consumers are used to swimming in murky waters, so to speak. Between messaging via social media, TV, display ads, and content marketing, consumers are used to being bombarded with truths and falsehoods about every product or service they encounter. And consumers do their best to sift through the fake stuff and find the truth. So why not trust them to do the same with reviews, which function in so many ways like just another marketing channel? Review sites like Amazon might have no other realistic option.