Posted in Business, Communications

Fake review business flourishes at online stores, but new tools help shoppers identify fabricated opinions

A neighbor asked me to help her with a purchase from Amazon. We searched the kitchen appliance she wanted, and quickly discovered a couple of good candidates. She glanced at them, and chose the one I wouldn’t have chosen. The product had plenty of positive reviews, she said. I didn’t explain the business of fake reviews to her at that moment, but perhaps she will read this, and think twice the next time.

Customer reviews have been an efficient marketing strategy for online stores as long as the stores have existed (for instance, Amazon was founded in 1995). When people search for a hotel, a camera, a book, or something else, they like to read reviews written by other customers. The common belief is that these people actually have purchased the product or stayed at the hotel, and they want to share their knowledge without any strings attached to their reviews.

walmart product review sample screen shot

Unfortunately, the competition between product vendors and online marketplaces have driven many businesses to rely on dubious marketing tactics. Buying and selling fake reviews is a big business that reaches every corner of the internet where customer reviews are published. For instance, European Union has estimated that in some countries up to 45% of reviews maybe fabricated. In Italy, an entrepreneur was sentenced to prison for writing thousands of fake reviews for hotels for money. He had a perfect marketing strategy: after publishing a review for a hotel, he contacted the nearby competing hotel and told them how to get positive reviews.

In the UK, Which? magazine that is in the business of producing high-quality professional product reviews, tracked a group of 87000 of people on Facebook who were members of a review writing service. After The Guardian wrote about the group and its services, Facebook closed the page. It is, however, quite easy to move somewhere else to continue business. The operation model was simple. Marketers told about a product to the Facebook group they wanted to promote, asking who would be interested in it. Those who signed up would purchase the product to get the Verified Buyer label at a store, write a five star review, and the marketer would pay the product price along with a fee to the reviewer.

If you have received a product from an online store you didn’t order, your name and address may have been used for fake reviews. The scheme goes like this. A reviewer purchases a product using your name and address to make sure his actions look legitimate to the store. Then, he writes five star reviews using your name. BBC reported of USB sticks and other cheap products that were ordered to random people. Later they discovered their names had been used in product reviews.

Writing fake reviews is practically business as usual in the UK. A reviewer who gets paid for his write-ups has allowed his name and portrait photo to be published. He likes his review job, and he believes everyone is doing it.

No matter how hard TripAdvisor, Amazon, Yelp, and other online services fight against fake reviews, it is a game they can’t completely win. If the reviews – real and fabricated – are written by people (and not by robots applying artificial intelligence (AI)) who can tell the difference?

Well, there is a tool for that.

If you still like to read product reviews, give these tools a try

Verified Buyer label is a way for online stores to restrict suspicious behavior in product reviews section. It means that you have to buy a product before you can review it. There are other methods that online services are applying. For instance, online services can hire people to analyse the reviews, but usually companies want to outsource the analysis to an algorithm that doesn’t require salary.

fakespo product review analysis service

Two review analysis tools are available for anyone to try: Fakespot and Reviewmeta. Find a product from, for instance, Amazon, copy and paste its link to the tool, and wait for the results.

I tried the review analysis tools, and came to the same conclusion as Cnet: they are far from perfect, but can help in some cases.

Fakespot focuses on the product and individual reviews, trying to assess if reviews make sense and if they are credible. The analysis service provides an easy to read report.

Reviewmeta tries to conclude the credibility of reviews from each reviewer’s history and activity. The service can only analyse Amazon reviews. Reviewmeta scans reviewer’s other write-ups and actions to give a score for reviews.

Reviewmeta product review analysis tool

Let’s stop for a moment and think. Why do we allow businesses and people fool us like this? This is insane. This is too much for a shopper who just wants to make an informed purchase decision.

Research conducted by humans and automatic analysis tools estimate that about half of product and service reviews are fabricated. Dedicated tools have been developed to help people to assess the reliability of reviews of a specific product. Who can an average online shopper trust anymore?

Who did we trust before online reviews? We trusted professional reviewers, like journalists or engineers who tested products, ranked them and wrote about them. These professionals still exist. They just don’t write on Booking.com, eBay, or Walmart product review pages.

Thousands of new online publications, like dpreview, Digital Trends, or Cnet have emerged that focus on testing new products. Of course, they must be critically viewed, but photos and video clips they publish are a proof that they actually have used the product.

Travel booking services that market private holiday rentals are exceptionally problematic in three ways: products and product information have not been verified, amateurs are often managing the products (apartments), and fake reviews are common. Again, professionals can help: travel guidebooks have reliable information.

Leave a Reply

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


CAPTCHA Image