Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


thesis_font_embedded.pdf (516.91 KB)

ETD Abstract Container

Abstract Header

A Framework for Estimating Customer Worth Under Competing Risks

Routh, Pallav
http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1525688483331787

Abstract Details

2018, Master of Science (MS), Bowling Green State University, Applied Statistics (ASOR).
Customer relationship management facilitates efficient management of the relationship between a firm and its customers. The notion of customer value helps quantify these relationships so that they can be measured and managed by the firm. A customer's survival probability and contribution forms two components of customer value. However, there are uncertainties associated with customer behavior that impact their worth to a firm. Specifically, these uncertainties directly affect customer retention and, therefore, degrades the customer worth, if they are unaccounted for. In this thesis, I use a Competing Risk customer valuation framework that is more robust in terms of withstanding these uncertainties. First, I compute the survival probabilities of customers using a Random Survival Forest method that incorporates the competing risk methodology to account for uncertainties in customer churn. This nonparametric estimation technique overcomes the distributional limitations of extant traditional parametric or semi-parametric methods. As a result, I demonstrate how the underlying factors that determine a customer’s survival, contribute to the competing causes of churn in terms of their importance. This study also demonstrates how the risks due to a particular competing cause evolve over time. Next, I compute the customer’s margin using a multiple regression model that incorporates customer specific information and the effect of time. The margin and survival propensities are then used to compute customer lifetime values. My contribution in this area advances existing works in customer valuation under contractual business settings in the presence of multiple causes of churn. I analyze the computed lifetime values using various data mining techniques, such as customer segmentation, and compare them to traditional non-probabilistic methods. Finally, I address the firms decision making process during allocation of promotional offers by using the estimated lifetime values and risks from the Random Survival Forest method. The optimization framework adapts to the changing customer behavior and ensures the high value customers with high risk of churn are targeted for making a promotional offer at a particular time, while also ensuring the cost of carrying out the entire marketing campaign is minimized.
Arkajyoti Roy, Dr (Advisor)
Jeff Meyer, Dr (Committee Member)
Arthur Yeh, Dr (Committee Member)
75 p.
Marketing; Statistics
Customer Relationship Marketing; Customer Lifetime Value; Competing Risks; Random Surival Forest

Recommended Citations

Citations

  • Routh, P. (2018). A Framework for Estimating Customer Worth Under Competing Risks [Master's thesis, Bowling Green State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1525688483331787

    APA Style (7th edition)

  • Routh, Pallav. A Framework for Estimating Customer Worth Under Competing Risks. 2018. Bowling Green State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1525688483331787.

    MLA Style (8th edition)

  • Routh, Pallav. "A Framework for Estimating Customer Worth Under Competing Risks." Master's thesis, Bowling Green State University, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1525688483331787

    Chicago Manual of Style (17th edition)

bgsu1525688483331787
587
© 2018, all rights reserved.
This open access ETD is published by Bowling Green State University and OhioLINK.