CHAPTER 6
Social Network Modeling*
Social network analysis can reveal possible correlations between business events, such as churn or purchasing, and other events within the social structure. Much of the following two topics are based on the work of Pinheiro and Helfert, published in the SAS® Global Forum in 2010. This work presented the first advances in the social network analysis method to increase bundle diffusion and decrease churn in the telecommunications industry.1 This correlation proves that there is a stronger impact when an event is triggered by an influential node in the social structure and a weaker impact when it is triggered by a noninfluential node. Influence is defined as the number of customers who follow an initial business event in a chain process or the percentage of related customers affected by an influential node.
Monitoring and analyzing social structures over time, particularly networks that are interconnected, allows companies to evaluate the impact of business events within the network in terms of revenue. According to the likelihood of churn events and the amount of influence assigned to customers, telecommunications companies can take straightforward actions to decrease churn and increase bundle diffusion in a chain of events.
Social network analysis, when applied to telecommunications, can help companies recognize the behavior of their customers and then predict the strength of links between customers and the possible impact of events among them. ...
