Financial services companies are concerned with identifying when customers have moved because of customer service and marketing concerns. Knowing when customers have moved can create opportunities for these companies to better market their financial products and provide improved customer service. This paper focuses on identifying features of customers who have moved for the purpose of predicting customer address changes. The data consisted of transactional and event data from the customer databases of a large financial services company. The analysis involved supervised learning techniques that used information about whether a customer had changed their address in the system as a response variable. The information gained about the features that explain customer address changes was extrapolated to all customers under the assumption that customers that have moved undergo similar spending patterns regardless of whether they changed their address in the database or not. A logistic model found that a combination of various spending factors leads to an 85.6% increase in positive predictive value in predicting customer movement over the general population. The customers labeled by the model as having moved can be marketed to by the financial company to reach a higher percentage of moving customers, including those customers who did not change their address in the database when they moved.