Analyzing Customer Churn in the Software as a Service (SaaS) Industry

Predicting customer churn is a classic data mining problem. Telecommunications providers have a long history of analyzing customer usage patterns to predict churn. Many other industries, such as banking, routinely analyze customer behavior to predict customer satisfaction and renewal rates. The Software as a Service (SaaS) model enables software vendors to collect customer usage data that is not available to traditional software vendors. The SaaS market, and cloud computing in general, is growing rapidly yet to our knowledge little work has been done to apply existing methods to analyze customer churn in the SaaS industry. This paper describes a study conducted on behalf of a leading performance and talent management company. Our work uses churn analysis in telecommunications as a baseline to study the SaaS industry. Although there are many similarities between the two industries our study identifies several important differences. The paper presents four experiments and outlines opportunities for future research.