Customer Lifetime Value: a framework for application in the insurance industry - Building a Business Process to Generate and Maintain an Automatic Estimation Agent

Research Project submited as partial fulfilment for the Master Degree in Statistics and Information Management

In recent years the topic of Customer Lifetime Value (CLV) or in its expanded version, Customer Equity (CE) has become popular as a strategic tool across several industries, in particular in retail and services. Although the core concepts of CLV modelling have been studied for several years and the mathematics that underpins the concept is well understood, the application to specific industries is not trivial. The complexities associated with the development of a CLV programme as a business process are not insignificant causing a myriad of obstacles to its implementation. This research project builds a framework to develop and implement the CLV concept as maintainable business process with the focus on the Insurance Industry, in particular for the nonlife line of business. Key concepts, as churn modelling, portfolio stationary premiums, fiscal policies and balance sheet information must be integrated into the CLV framework. In addition, an automatic estimation machine (AEM) is developed to standardize CLV calculations. The concept of AEM is important, given that CLV information “must be fit for purpose”, when used in other business processes. The field work is carried out in a Portuguese Bancassurance Company which is part of an important Portuguese financial Group. Firstly this is done by investigating how to translate and apply the known CLV concepts into the insurance industry context. Secondly, a sensitivity study is done to establish the optimum parameters strategy. This is done by incorporating and comparing several Datamining concepts applied to churn prediction and customer base segmentation. Scenarios for balance sheet information usage and others actuarial concepts are analyzed to calibrate the Cash Flow component of the CLV framework. Thirdly, an Automatic Estimation Agent is defined for application to the current or the expanding firm portfolio, the advantages of using the SOA approach for deployment is also verified. Additionally a comparative impact study is done between two valuation views: the Premium/Cost driven versus the CLV driven. Finally a framework for a BPM is presented, not only for building the AEM but also for its maintenance according to an explicit performance threshold.