Amazon Reviews and Information Cascades
Amazon has been one of the top platforms for people all over the world to search, review, and buy products that they’re interested in. With the website being easily accessible and having quick shipping times, Amazon fostering millions of ratings for people to read through for all their products should be of no surprise. As a result of their overwhelming popularity, it is only natural for information cascades to occur within the site.
Information cascades relates to the idea where when people are connected by a network, there is a chance for them to influence the behavior and decisions of others. Therefore, this can be applied to endless situations whether it be movie ratings, election voting, and certainly in a website with as many users as Amazon does. An information cascade is where every person has their own personal information, but when they leave their information as a response to observing the actions of previous people, this is when a cascade forms. Therefore, in situations where people make decisions in a sequential manner, information cascades have a chance to occur, which is especially relevant to a platform like Amazon because it has a collection of past reviews that previous users have left based on their experience with a product.
A study at Stanford University actually set out to analyze Amazon reviews and whether subsequent reviews would create a herding effect. Surprisingly, with a random model the researchers created where previous reviews are not considered when people made decisions and a cascading model they designed, they found that the random model performed better and that people tended to give their honest evaluations. However, these results do not necessarily indicate that information cascading didn’t occur at all as the researchers did discover about 31% of the reviews show there was some sort of cascading behavior as well. Looking at this study as a whole, the results do make sense because if we consider other possible factors such as the fact that users who leave reviews on a product are the people that tend to have a passionate opinion about something, it is unlikely that they will be influenced by the past reviews.
The biggest thing we can learn about information cascades after analyzing their relevance to Amazon product reviews is that it is something that is ingrained to the real world in that it is a normal occurrence for people to conform to the opinions of others. Although the results of the Amazon study weren’t in favor that information cascades occur frequently, the evidence that shows that herding happened to a certain extent indicates that it is an important analysis of human behavior because decision-making is something we all do in our daily lives.
https://www.cs.cornell.edu/home/kleinber/networks-book/networks-book-ch16.pdf
http://snap.stanford.edu/class/cs224w-2012/projects/cs224w-033-final.v08.pdf