Social Networking in E-Commerce Sites
While the social networks behind websites such as Facebook and Twitter are well known, there are lesser-known social networks that exist in nonsocial websites like e-commerce sites. For example, companies like eBay and Amazon use product reviews to connect buyers, and websites like Tobi use recommendation engines to show people what their friends have been buying. Additionally, websites such as Groupon or LivingSocial reward people for social interactions, and in turn, promote them. A study performed at Stanford University was used to demonstrate the role of social connections in online shopping. The motivation for their study was to illustrate that the same behavior in offline consumers is also represented in online social commerce. Meaning, that the way that consulting friends or family when making a purchase influences people can occur online as well.
The study focuses on data taken from one million users on Taobao, the largest e-commerce website in the world. Aside from its size, Taobao serves as a good representation of social influences in shopping sites as it contains an instant messaging feature, which makes it more likely that people will communicate. The information that they collected was then analyzed through social graphs with buyers and sellers each represented as nodes, connected by three types of edges: “trades”, or purchases, “messages”, or any type of communication, and “contacts”. The results from the data set prove how different types of edges create different relationships, and how these can manifest themselves in triadic closures, which can then predict consumer behavior. For example, the relationships demonstrate that message edges result in many more purchases than contact edges. The explanation they use for this is that a contact takes a small amount of time to add, however, a message requires both time and effort to send, which creates a closer relationship, as both parties need to demonstrate interest. Additionally, they consider who a buyer would purchase from between a seller with a higher trust on the website, a lower price, or the most mutual friends. They state that if the website did not consider a social feature, that the seller would buy the lowest price, however the data demonstrates that a buyer would pay a higher price to purchase from a trusted seller with whom they share an edge.
This analysis directly references many topics that have been studied in the Networks class. It references subjects such as network graphs and demonstrates examples of triadic closure in networks that are not primarily social. Additionally, it takes some of the topics that have been discussed in the class and explains how they can be used to predict future purchases, and in turn, serve as a tool for online marketing strategies.
http://cs.stanford.edu/people/jure/pubs/taobao-ec11.pdf