Statistical methods in customer relationship management

This book is a review of statistical methods applied to managing a firm’s relationships with customers. Kumar and Petersen bring their deep expertise in quantitative marketing and customer relationship to the fore by giving a concise and hands-on overview of different modelling techniques. Customer relationship management (CRM) is employed by many firms to get a better understanding of their customer behaviour. It consists of four phases: (a) acquiring customers, (b) retaining the customer base, (c) managing the attrition rate of customers, and (d) reacquiring the valued customers who left the firm. The authors begin by describing the mechanism for collecting customer data, and calculating appropriate metrics to monitor customer engagement. They then succinctly review appropriate modelling techniques in each phase while explaining the mathematical details behind the techniques in the appendices, although there are a few typographical errors in the appendices. For example, Sf (ti | Xi) in Equation (D.1) should be replaced with S(ti | Xi). A list of references is provided at the end of each chapter for further exploration. The authors explain the relevant attributes of each data set, and demonstrate the effectiveness of the models using empirical data. They also provide the SAS code for reproducing their results, and mention about implementing the models in other software programs like MATLAB, GAUSS, and R. They demonstrate the application of these models on several case studies. Finally, they throw light on how social media and mobile devices are shaping the future of CRM to target individual customers better. The authors have taken a unique approach in condensing the vast literature in CRM into a single volume. Thus, the book serves as an excellent survey of modelling techniques in CRM. It is accessible to applied statisticians by giving them an overview of the assumptions behind different models as well as their application. Since the book focuses on implementation and interpretation of models, it will be of interest to practitioners in CRM who want actionable insights. It can, however, take time and effort on the part of those who are unfamiliar with basic probability and statistical techniques. It can be used in graduate courses on quantitative models in marketing, and customer life cycle management. The book would have benefited from comparison between different competing models in addition to the sole comparison against the “random guess” model. The authors could have supplemented in-sample accuracy with cross-validation results to provide a more reliable estimate of the generalisation error. Also they could have fleshed out the details of modern data mining techniques that have fewer modelling assumptions, and are becoming popular with very large data sets collected on customer behaviour. The addition of R code to implement different models could have enhanced the appeal of the book, given the popularity of R as open-source statistical software.