Advanced Business Analytics (eBook)

Saumitra Bhaduri received his Master’s degree in Econometric from Calcutta University, Kolkata, India, and his PhD in Financial Economics from Indira Gandhi Institute of Development Research (IGIDR), Mumbai, India. He currently works as a professor at Madras School of Economics, Chennai, India, where he regularly offers courses on Financial Economics and Econometrics, and on Advanced Quantitative Techniques. In terms of his former career he also worked at GE Capital, the financial services division of the General Electric Company, and has held various quantitative analysis roles in the company’s finance services. He also founded and headed the GE – MSE decision Sciences Laboratory, where he was responsible for developing state of the art research output for GE. He has also published several research articles in various international journals. His research interests include: Financial Economics and Econometrics, Quantitative Techniques and Advanced Analytics. David Fogarty received his BS in International Relations from Connecticut State University, USA, his PhD in Applied Statistics from Leeds Metropolitan University, UK, and his MBA with a concentration in International Business from Fairfield University, USA. He also has a post-graduate qualification from Columbia University in NYC. In terms of his professional career, he currently works at a Fortune 100 health insurance company as the Chief Analytics Officer or Head of Global Customer Value Management and Growth Analytics. In terms of his former career, Dr. Fogarty also worked for 20 years at GE Capital, the financial services division of the General Electric Company, and has held various quantitative analysis roles across several functions, including risk management and marketing, both internationally and in the US. He currently holds over 10 US patents or patents pending on business analytics algorithms. In addition to his work as a practitioner Dr. Fogarty has over 10 years of teaching experience and has held various adjunct academic appointments at both the graduate and undergraduate level in statistics, international management and quantitative analysis at the University of Liverpool (UK), Trident University (USA), Manhattanville College (USA), University of New Haven (USA), SUNY Purchase College (USA), Manhattan College (USA), LIM College (USA), the University of Phoenix (USA), Chancellor University (USA), Alliant University International (USA) and the Jack Welch management Institute at Strayer University (USA). Dr. Fogarty is also an "Honorary Professor" at the Madras School of Economics in Chennai, India and has given guest lectures in Asia at East China Normal University (Shanghai, China), Ivey Business School (Hong Kong, China), and the City University of Hong Kong. He has also taught business analytics courses at the esteemed GE Crotonville Management Development Institute in Crotonville, New York. Since obtaining his PhD, he has continued to collaborate with several universities and leading academics to pursue academic research and has several published research papers in peer-reviewed academic journals. His research interests include: how to conduct analysis with missing data, the cultural meaning of data, integrating genetic algorithms into the statistical science framework, and many other topics related to quantitative analysis in business.

Contents 5
Authors and Contributors 9
List of Figures 11
List of Tables 12
1 Introduction and Overview 14
References 29
2 Severity of Dormancy Model (SDM): Reckoning the Customers Before They Quiescent 31
Abstract 31
2.1 Introduction 31
2.2 Severity of Dormancy Model 33
2.2.1 Methodology 33
2.2.2 Severity of Dormancy Model 33
2.2.3 Prediction 35
2.2.4 Estimation 35
2.3 Data 35
2.4 Variables Used 37
2.5 Results 37
2.6 Beyond Conventional Dormancy Model 39
2.7 Conclusions 42
References 42
3 Double Hurdle Model: Not if, but When Will Customer Attrite? 44
Abstract 44
3.1 Introduction 44
3.2 Double Hurdle Model 45
3.2.1 Methodology 45
3.2.2 Tobit 45
3.2.3 Double Hurdle Model 46
3.2.4 Prediction 47
3.2.5 Estimation 48
3.3 Data 48
3.3.1 Variables Used 49
3.4 Results 49
3.5 Beyond Logistic Regression 53
3.6 Conclusion 56
References 56
4 Optimizing the Media Mix—Evaluating the Impact of Advertisement Expenditures of Different Media 58
Abstract 58
4.1 Introduction 59
4.2 Efficiency Measurement 59
4.2.1 Input-Oriented Measures 60
4.2.2 Output-Oriented Measures 61
4.2.3 Date Envelopment Analysis 61
4.2.4 Estimation 63
4.3 Data 63
4.3.1 Deseasonalization 63
4.3.2 Adjusting Spillover Effects 64
4.3.3 Model 65
4.4 Results 65
4.5 Conclusions 66
References 67
5 Strategic Retail Marketing Using DGP-Based Models 68
Abstract 68
5.1 Introduction 69
5.2 Methodology 71
5.2.1 Model Likelihood Function 71
5.2.2 Derivation of P(active|x, n, m) 72
5.2.3 Expected Number of Future Transaction 73
5.2.4 Average Money Value of Future Transaction 73
5.2.5 Prediction 74
5.2.6 Estimation 74
5.3 Data 74
5.3.1 Variables Used 75
5.4 Results and Retail Strategy Booster 75
5.4.1 Model Results and Validation 75
5.5 Conclusions 80
References 81
6 Mitigating Sample Selection Bias Through Customer Relationship Management 82
Abstract 82
6.1 Introduction 82
6.2 Methodology 84
6.2.1 Simultaneous Approach to Correct the Selection Bias 85
6.2.2 Estimation 86
6.3 Data 87
6.3.1 Variables Used 87
6.4 Results 87
6.5 Understanding and Identifying the Likely Responders from Non-selected Base 92
6.6 Conclusions 93
References 94
7 Enabling Incremental Gains Through Customized Price Optimization 95
Abstract 95
7.1 Introduction 95
7.2 Methodology 97
7.2.1 Customized Price Optimization Solution 97
7.2.2 The Generic Construct 97
7.2.3 Price Differentiation 99
7.3 Price Optimization Framework 99
7.3.1 Adverse Selection 100
7.3.2 The Response Model 100
7.3.3 Early Settlement 102
7.3.4 CRM Through Cross-Sell and Up-Sell 104
7.3.5 Segmentation 104
7.4 Segmentation Through GA 105
7.4.1 Optimization—Local Versus Global Optimum 106
7.4.2 Regulatory Constraints, Market Dynamics, and Competitive Conquest 106
7.5 The Optimization Model 106
7.6 Simulation 107
7.7 Summary 109
References 109
8 Customer Relationship Management (CRM) to Avoid Cannibalization: Analysis Through Spend Intensity Model 110
Abstract 110
8.1 Introduction 110
8.2 In-Store Purchase Intensity Model 112
8.2.1 Methodology 112
8.2.2 In-Store Intensity Model 112
8.2.3 Prediction 114
8.2.4 Estimation 114
8.3 Data 115
8.3.1 Variables Used 115
8.4 Results 115
8.5 Beyond Conventional Intensity Model 118
8.6 Conclusion 119
References 120
9 Estimating Price Elasticity with Sparse Data: A Bayesian Approach 121
Abstract 121
9.1 Introduction 121
9.2 Methodology 122
9.2.1 Methodology for Missing Value Techniques 122
9.2.2 Methodology for Sparse Data Techniques 126
9.3 Empirical Model 128
9.4 Data 129
9.5 Results 130
9.6 Distribution of Price Elasticities 133
9.7 Conclusion 135
References 136
10 New Methods in Ant Colony Optimization Using Multiple Foraging Approach to Increase Stability 138
Abstract 138
10.1 Introduction 138
10.2 k-Means and Ant Colony Optimization as Clustering Techniques 140
10.3 Methodology 141
10.4 Algorithm Details 141
10.5 Conclusion 144
References 145
11 Customer Lifecycle Value—Past, Present, and Future 146
Abstract 146
11.1 Introduction 146
11.2 Fundamentals of CLV 148
11.3 CLV Approaches in Literature 149
11.3.1 Probability Based Models 149
11.4 Econometric Models 153
11.4.1 Customer Acquisition 154
11.4.2 Customer Retention/Activity 155
11.4.2.1 “Lost for Good”—Hazard-Based Models 156
11.4.2.2 “Always a Share”—Markov Models 157
11.4.3 Customer Margin and Expansion 158
11.5 The Future of CLV 159
11.5.1 Moving Beyond Static Hazard Models 159
11.5.2 Reconciling Future Uncertainties Using Fuzzy Logic 160
11.5.3 Recognizing the Need to Model Rare Events 160
11.5.4 Scope of Bayesian Framework to Overcome Future Uncertainties 161
11.6 Conclusion 161
References 161