Fake Amazon Reviews: The Business Behind Five Stars

Technology News

In the Telegram group ‘Amazon Reviews’, almost 6,500 people receive more than 200 ads for products every day that they can review to get them for free. This is just one of dozens of groups created by companies that offer free products in exchange for five-star reviews on Amazon.com.

According to a study prepared by the Organization of Consumers and Users (OCU) on the opinions that consumers give about the products or services that are sold online, 8.6% of the products analyzed on Amazon are affected by anomalous opinions.

In research carried out by the platform, it is shown that brands promote positive opinions in different ways. Some give discount coupons, others send an email to remind the user that they can give a favorable opinion if they liked it, and others opt for false opinions.

On Facebook or Telegram, it is as easy as searching for ‘amazon reviews’ to find dozens of groups in which users are invited to comment positively on items in exchange for gifts or income via Paypal. These are generally Chinese stores that employ intermediaries for this purpose.

“The group of this type with the largest number of users in Spain has 20,000. It is as easy as subscribing and when an article interests you, you let them know. The seller or intermediary (sometimes through a bot) requests the Amazon profile to Verify that you do not have many purchases or many opinions and if it passes the filter it tells you to proceed to rate that item with 5 stars. Other times it asks you to buy it and even other times it asks you to value the store instead of the item. It depends . of course, revenues are always carried out by Paypal “explained from OCU.

According to the consumer organization, they have detected more than 70 companies that are carrying out this type of procedure.

To detect this type of ‘paid’ reviews, OCU has analyzed around 6,360,000 reviews on some 47,000 products marketed on Amazon, Italy, Spain and France, as well as hotels on TripAdvisor and Booking.com. “Specifically, unnatural behaviors are analyzed, such as a person concentrating many opinions on the same day, that their opinions are always very positive or that opinions for the same product are concentrated on specific days. When several of these behaviors (we analyze 11 different) are concentrated in the same product significantly influencing its final evaluation, we consider that it is affected by interested opinions “.

The result is that between June and August it was detected that 8.38% of the products have a significantly altered valuation due to unnatural opinions on Amazon, 6.20% on Tripadvisor and 2.10% on Booking.com .

In addition to this practice, “there are companies based in India, the Philippines, Bangladesh and elsewhere that have large teams of fake review generators,” says José Antonio Tovar, Founder of Dispitch.io.

Tovar ensures that on the internet and networks you can find companies that sell and traffic opinions. They write them in data farms but in addition to being false they have errors and serious errors . Obviously this damages the reputation of any business. “To detect them, users notice that they are too long and not very objective. Not to say that whoever signs it has the name of India, Pakistan, etc., not Spanish. Generally, any exaggeratedly good opinion is suspicious.”

Also, 1 in 5 consumers write negative reviews to get special discounts . “It is common to post a bad opinion on Google and wait for the call from the brand to remove it and extort money in return.”

Very famous was the case of Dabiz Muñoz complaining about the number of requests from influencers asking him to invite them to lunch or they would write a negative review.

Dispitch.io states that 90% of businesses suffer from the dreaded Fake Reviews, similar to fake news but that directly attack the reputation of brands. With the arrival of the Christmas campaign, these false opinions are multiplied by 10.

Both companies are aware of this problem and have measures in place to stop it.

At Amazon, they say they work hard to deal with this phenomenon. The platform uses an algorithm to order products taking into account the number and quality of your opinions (among other factors). In addition, it prohibits sellers from offering incentives in exchange for opinions and review comments before posting them. It does not allow you to comment if you have not made purchases of more than 50 euros.

On the other hand, it also allows you to comment on products even if they have not been purchased (without putting the Verified Purchase label) and allows stores to contact users through their email.

According to Amazon, the company invests significant resources to protect the integrity of comments. “When it comes to customer feedback, we have clear engagement guidelines for both users and vendor partners, and we suspend, prohibit and take legal action against those who violate our policies.”

“We use a combination of investigative teams and automated technology to prevent and detect false product reviews at scale, and to take action against the bad actors behind these abuses. We estimate that more than 90% of unauthentic product reviews are generated by computers, and we use machine learning to analyze all incoming opinionsand existing 24 hours a day, 7 days a week and block or remove inauthentic reviews. Our team investigates suspicious reviews, works on social media to stop inauthentic reviews from the source, takes legal action to prevent criminals from planning to abuse customer reviews, and feeds new information into our automated systems to keep it going. improving and being more effective in detecting abuse “, explains the company.

For its part, in 2018 on Tripadvisor they received 66 million opinions, of which 2.1% were identified as fraudulent.

The company uses automated systems to filter reviews. These systems are complemented by a team of hundreds of experts who investigate all those opinions that have set off alarms or that have been reported by users.

Leave a Reply

Your email address will not be published. Required fields are marked *