Fake Reviews and Information Cascades
How do you shop online? Do you buy the top hit, the name brand, or the cheapest product satisfying your needs? Or maybe you, like myself and most other consumers, buy the product with the most favorable reviews.
Now, what if I told you some of the most favorable reviews are fabricated? You might be more hesitant to purchase those highly-rated products.
The Forbes article written by Emma Woollacott, Amazon’s Fake Review Problem Is Now Worse Than Ever, Study Suggests, sheds light upon the issues of fake online product reviews. Vendors have a variety of tactics to product fraudulent product reviews, one of which is using Facebook groups, and offering full refunds to consumers for positive reviews. Social media platforms provide the perfect pool of people to coax into such a deal. Amazon’s product search algorithm takes into account number of reviews, and average review rating, when querying its product database and displaying results to users. Therefore, the shocking quantity of fake reviews drastically disrupts the most used and trusted source of product quality, deceiving online consumers into purchasing sub-par merchandise.
There is a direct correlation here to the study of information cascades. Information cascades are situations where people make decisions, not based off of their own information, but based off the observations and choices of others. Fake reviews lead more and more people to purchase a product, sending it higher and higher up the search result list as its popularity increases. The consumer has no personal information about what it is that they are purchasing, and like sheep, blindly purchase the top products: what others are buying. This can be perfectly fine in the case of reviews being legitimate, but that is not always the case, as shown by this article and many others like it.
As of October 2017, Amazon receives nearly 44 percent of all U.S. e-commerce sales. To preserve its market share and positive social status, Amazon needs to continue investing resources into fraudulent review detection research so that these information cascades, based on faulty reviews, don’t have the opportunity to come to fruition.