Skip to main content

Samsung Admits to Paying for Deceitful Product Reviews

In numerous lectures and problem sets, we discussed the importance of reputations in the online marketplace. For example, we figured out the probabilities of an eBay seller deceiving a buyer given that they have a certain amount of positive and negative ratings. This is important because when shopping on the internet, consumers have a lot more doubt than they would if they went to purchase a t-shirt or gadget at a local store in person. Because of this it is really important that sellers and companies monitor their online reputations. Unfortunately, there is only so much that a company can do to actively affect their online reputation. On Thursday, April 18, 2013 it was reported that Samsung took their reputation monitoring efforts to a deeper level when it was revealed that “Samsung paid people to post favorable reviews about the products online while at the same time offering up negative feedback on rival HTC’s products”. While the attempted actions obviously aimed to improve the reputation of Samsung, it results in a huge hit to the company’s image. Knowing that Samsung has paid people to write positive reviews for their products – and negative reviews for their competitors’ goods – will make many customers more skeptical of reviews praising Samsung gadgets, even if they are legitimate reviews from satisfied customers. Samsung is not the only company who ever did this, but it still does not bode well for the international electronics maker.

A major effort to combat such behavior has been studied by researchers at Cornell University, who have been trying to develop strategies to detect deceitful reviews. They characterize deceitful reviews as those that a company writes to praise their product or criticize a competitor’s. Myle Ott, Yejin Choi, Claire Cardie, and Jeffrey Hancock preformed their study on online hotel reviews and have found some great results. They discovered that humans can only identify whether a review is deceitful about 50-55% of the time, while the algorithms they designed were effective almost 90% of the time. This has two implications. First, humans have proven to be somewhat incapable of identifying deceitful reviews. As a result, deceitful reviews can be extremely effective at persuading a consumer without his or her knowledge. However, since automated algorithms have proven quite effective at recognizing deceitful reviews, as the technology is developed more it will be interesting to see if online marketplace and gadget review blogs/forums will implement software to flag potentially deceitful reviews to help safeguard customers from this unethical, yet unfortunately frequent practice.

– cfs76

Sources used:


Leave a Reply

You must be logged in to post a comment.