Can we trust Amazon reviews and Amazon ratings? The vast majority of shoppers do, but research shows the Amazon rating system is far from perfect. Paid reviews and reviews compensated by discounts and gifts of free products abound. These water-muddying factors have created a situation where fully 66% of all Amazon ratings are 5-star. While Amazon uses AI to fight fake reviews and to some extent succeeds, their rating system is far from perfect.
Consumers can use tools like Fakespot and Amazon Verified Reviews to gain a higher level of trust. Even so, a little digging beyond the easy Amazon answer can go a long way, as this article shows.
Can We Trust Amazon Reviews?
Basically, Amazon reviews can’t really be trusted. That statement comes with several spoonfuls of salt. Shoppers use Amazon reviews to find out if a product is worth buying. According to a study by SearchEngineLand, 88% of all shoppers trust these reviews. Most consumers are savvy enough to know that a star rating of a product with thousands of ratings is more trustworthy than one with just a few. However, Amazon ratings and reviews are overwhelmingly positive. They’re vulnerable to paid review farms and other compensated reviews. Many of them aren’t from verified purchasers.
Amazon is working to combat the problem, but in the meantime, shoppers shouldn’t put all their faith in the Amazon review and rating system.
Amazon Reviews Are 66% 5-Star
A 2016 study by BestReviews.com found that Amazon reviews and ratings skew heavily positive. In other words, looking only at Amazon ratings could easily convince a shopper that everyone likes everything. The study looked at 360,000 user ratings for 488 products in multiple categories. The reviews were overwhelmingly glowing:
Also see: Amazon vs eBay: Where to Sell Your Stuff
66% of all the Amazon reviews surveyed carried 5-star ratings. Meanwhile, reviews of movies, hotels, airlines and other products are far lower. Even Yelp, the review site widely derided for accepting paid reviews, turns in a collective review score far lower than Amazon’s:
Consumers using Amazon reviews and Amazon ratings to decide whether a product is worth it might therefore be getting a warped answer.
Fake Amazon Reviews: A Real Problem
The reason for the erosion of trust in Amazon reviews comes down to fake reviews. Not every review on Amazon is fake, but it’s in the interest of sellers and manufacturers to make sure their products get high review scores. More and more consumers make decisions on what to buy based on a product’s star rating. Take two identical products, one with a 4.5 star rating on Amazon and another with a 4 star rating. The one with the slightly higher rating will see much better sales. There’s real evidence that fake reviews exist.
Reviewing a Product Before it Exists
Fake Amazon reviews are a real problem. We found more than a couple solid five-star reviews of smartphone cases for phones that hadn’t yet hit the market.
Purchased Amazon Reviews
There are two kinds of fake Amazon reviews. The first is a review and/or rating that’s bought outright, for cash. There are review “farms” online that let people write fake reviews for money. The money comes from sellers and manufacturers who want to bump up their Amazon ratings. Amazon frowns on this practice. The retailer has gone as far as pursuing these fake review farms in court. To date they’ve filed three lawsuits against more than a thousand fake reviewers.
Also see: How to Sell Things on Amazon
Compensated Amazon Reviews that Aren’t Purchased
One type of Amazon review that the online retail giant isn’t battling is the non-purchased but compensated review. These reviews are still paid for, but not with cash. Reviewers can sign up with several review networks to rate products in return for freebies and discounts. Studies have shown that these compensated reviews tend to have much higher ratings than regular, non-compensated reviews. Below is a list of services pay users with discounts and free items in exchange for reviews:
Amazon Review Networks
|Elite Deal Club|
|Amazon Reviewer Network|
|AMZ Review Trader|
What Is Amazon Doing to Fight Fake Reviews?
To combat the diminishing trust in Amazon reviews and ratings, the retailer has filed three lawsuits against over 1,000 fake reviewers. It has also enlisted the aid of AI, or artificial intelligence, to fight fake reviews. Amazon’s site uses machine learning to identify the more trustworthy reviews and adjust its product ratings accordingly. As the screenshot below shows, Amazon’s star rating is no longer just an aggregate of all its customer reviews. While this is great and exciting, we can’t forget that BestReviews study that found 66% of all Amazon reviews are 5-Star. In other words, the online seller has a way to go yet.
Reasons to Trust Amazon Reviews
Here’s a caveat to the don’t-trust-Amazon-reviews outcry. Even if Amazon’s reviews do skew overwhelmingly positive, its machine learning algorithm may be taking that into account. In other words, it’s possible that the total of all ratings on Amazon is too positive, yet the aggregate of each product review isn’t.
If the Amazon AI really does its job, it should allow lots of too-high individual product reviews, yet filter those out when it adds up the score for each product. Further, BestReviews.com has a race in this horse. They make their money when customers use their site to double-check Amazon’s math. We’re not saying they’re biased, just that they have a very large financial reason to share exactly the results they are sharing.
Even so, there’s enough of a red flag here to trust Amazon’s reviews only with caution. The tools below give shoppers extra arrows in their quivers on the quest to learn the value of a product before purchasing it.
Using Fakespot to Screen Amazon Reviews
Anyone trying to see through the murk in the Amazon review story has an ally in Fakespot. Fakespot analyzes and reports on the accuracy of Amazon ratings. The service studies the language used by reviewers and reviewer profiles. It uses machine learning to detect reviews with a high probability of fraudulence. To use the service, just visit Fakespot.com and paste in the URL of an Amazon product. The image below shows a Fakespot review of a very high-rated Amazon product. This product gets a 4.5 star rating on Amazon as of 12/12/16.
The downside of Fakespot is that it isn’t infallible either. For instance, reviews of the Amazon Fire get only a score of B on Fakespot. According to Fakespot, over 80% of those reviews can be trusted. Does that mean Amazon faked nearly 20%? Almost certainly not. If Amazon tried that, it would lose a vast amount of trust from shoppers if discovered, and it would almost certainly be discovered. That would mean a lot less revenue for the retail giant.
Also see: 18 Big Ways to Save Money on Amazon
How to Use Amazon Verified Purchase Reviews
Another way to boost the trust level for an Amazon review or an Amazon rating is to focus on their verified purchase reviews. To look only at verified reviews, follow the steps in the image below.
First, in the Amazon product page, click on the number of reviews. Note that with this product we’ve got 284 reviews total giving a combined score of four stars. After you click on the number of reviews, you’ll need to click it again in the next page. This will bring up some options for sorting the reviews. Click “verified purchase only.” This means the reviewer actually bought the product on Amazon. Last, look at the number of verified reviews and compare it to the total review count. In this case there were 284 total reviews but ony 144 of those were from verified Amazon purchasers.
Looking at the verified purchase reviews gives a shopper a better idea of what the product is really like. One big downside is that it takes a lot longer.
The Difference Between Verified and Non-Verified Amazon Reviews
In the example above, the good news is that the score from the verified Amazon purchases matches the score Amazon is showing. We picked a product at random on Amazon: the Tile Mate Key Finder. The key finder has 284 reviews and a score of four stars. Using our method above, we learned that 144 of those reviews are from verified Amazon purchases. 144 of them aren’t. In other words, only 51% of the reviewers here actually bought the product on Amazon.
We did the math on the reviews and came up with the same rating for verified purchases: four stars. At first this looks like Amazon’s AI actually used only the verified reviews to come up with their overall score. However, looking at the number of 1-star reviews shows only 18 1-star reviews from verified purchasers but 20 more from non-verified purchasers. This is more than we expected.
It turns out that what’s going on here is that people who bought this product elsewhere and hated it came to Amazon to vent their frustrations. While this at first looks like a win for Amazon’s ratings system, it doesn’t work that way with every product. As the analysis in the next section shows, sometimes the system backfires.
Amazon Verified Reviews VS Non-Verified (Tile Mate Key Finder)
|Verified Purchase Reviews||144|
|% of Reviews Verified||51%|
|Overall Rating||4 Stars|
|Rating from Verified Purchase Reviews||4 Stars|
When Amazon Reviews Can’t be Trusted
To further test the Amazon rating system, we looked at the cell phone case shown below. This case’s score on Amazon gets an F from Fakespot.com. The case has 237 reviews and a 4.5 star rating. However, using our method from earlier in this article shows that only 17 of these ratings are from verified Amazon purchasers.
Doing the math on those 17 verified purchase ratings yields a lower four-star review, not the 4.5 stars displayed by Amazon. While that’s only a half star difference, it shows the fallibility of the system. In other words, if this product’s Amazon review is off by a half star, it’s definitely possible that other reviews are off by this amount or even more.
Amazon Rating Example (iPhone 7 Case)
|Verified Score||4 stars|
Why We Can’t Trust Amazon Reviews and Ratings
The bottom line with Amazon reviews and ratings is that we can’t really trust them. They’re getting better thanks to machine learning but they’re not perfect. Even verified purchases can be gamed, since all a fake review service has to do is reimburse the reviewer and voila. Also, Amazon doesn’t offer the option to toggle between a total score for all reviews or a total score for only verified reviews. That means customers don’t fully know what score they’re looking at.
How to Fine Tune Amazon Reviews
The only real way to trust Amazon ratings and reviews is to do a bit of research beyond the reviews themselves. The following are a list of tools consumers can use to look behind the reviews and see the real value of a product:
- Don’t trust Amazon reviews and ratings 100%.
- Check the top four or five magazine-style review sites.
- Use Fakespot.com but don’t fully trust that either.
- Use ConsumerReports.com and BestReviews.com.
Don’t trust Amazon reviews. Look beyond Amazon’s reviews and do a bit of original research, especially for high-dollar items.
Check magazine-style reviews. After looking at a product’s Amazon rating, google the product’s name and “review.” Skim the first four that pop up to see if this is really a good buy or a turkey. Be aware that these kinds of reviews can be coerced as well.
Use Fakespot.com but don’t fully rely on that either. Fakespot uses a good system to sift through Amazon’s ratings, but it isn’t infallible. It does however offer another layer of trust for any Amazon product’s review score.
Use BestReview and Consumer Reports. These two sites actually lab-test products, running them through rigorous testing. Lab testing has its drawbacks too, but it helps.
Finally, while the above methods can help, they’re not the end-all, be-all of shopping. Most people simply don’t have time to do this kind of research for small, inexpensive items. A great fallback is Amazon’s excellent return policy. Consumers can generally boomerang any purchase that turns out to be junk.