RBC: Social Network Analysis

Description

In October 2013, the Royal Bank of Canada (RBC), Canada’s largest bank, hired a new head of Enterprise Fraud Strategy, a department tasked with protecting RBC’s global customers from fraud. The department head’s immediate priority was to prevent fraudulent transactions by RBC’s own customers—a phenomenon called first-party fraud—by implementing a bourgeoning technology called social network analysis (SNA). The technology used predictive analytics and big data to forecast the occurrence of first-party fraud. The head of Enterprise Fraud Strategy had three primary questions: First, how should SNA be used to bring down the ratio of fraud alerts to actual fraud at RBC? Second, how should the cost of maintaining SNA protocols be reduced? Finally, how should the issues around systemic performance of SNA be resolved?

Learning Objective

Students should take on the perspective of the case’s central figure to find out which rule (or combination of rules) provides the best fraud detection for RBC and how effective these rules would be if implemented for all RBC customers. After using the case, students will be able to do the following: ·Highlight the features of first-party fraud, particularly its “connected explosion.” ·Discuss the unique attributes of social network analysis. ·Demonstrate the importance of data analytics in detecting first-party fraud. ·Identify the important trade-offs in fraud detection. ·Use advanced analytical methods to test out various fraud detection rules and construct a “best” model for fraud detection.

Keywords

Big data Data analytics Fraud