Joint Research Conference
June 24-26, 2014
Studying and gaining insights based on corporate client payment transactions allows banks to deliver specific customer communications based on the individual client's transaction patterns, helping to increase loyalty and improve customer retention. In addition, an insightful view of client transaction behaviors will help banks to understand how actions taken by the bank affected their customer's behavior, and also allow for predictions of customer reaction to proposed changes. However, today’s banks lack a clear view of large corporate clients’ payment transaction behaviors. Banks are challenged with siloed operations and latency from legacy systems. Legacy back-end systems typically resolve transactions after the close of the business day. Distinct payment systems are used based on the payment type, for example, ACH or WIRE. Multiple departments in a bank interact with different organizations within a large customer’s enterprise making in-context decisions difficult for any one bank entity. There is a need to gain a holistic view of customer transaction behaviors based on their payment transaction patterns. In this research, we studied bank transaction data from corporate clients. We apply statistical analysis and machine learning techniques to examine client transaction behaviors and explore client historical cash flows, credit and debits over time from several different perspectives. Exploring cash flow patterns will help banks segment customers based on their different transaction patterns and thus identify customer and product/service relationships to better understand the most profitable customer product combinations. In addition, cash flow pattern recognition will facilitate client cash flow prediction for more efficient cash management.