Segmentation, Revenue Management and Pricing Analytics

267 Pages·2014·2.261 MB·English

by  Tudor Bodea, Mark Ferguson

Segmentation, Revenue
Management, and
Pricing Analytics
The practices of revenue management and pricing analytics have transformed the trans-
portation and hospitality industries, and are increasingly important in industries as
diverse as retail, telecommunications, banking, health care, and manufacturing. Segmen-
tation, Revenue Management, and Pricing Analytics guides students and professionals on
how to identify and exploit revenue management and pricing opportunities in different
business contexts.
Bodea and Ferguson introduce concepts and quantitative methods for improving
profit through capacity allocation and pricing. Whereas most marketing textbooks cover
more traditional, qualitative methods for determining customer segments and prices,
this book uses historical sales data with mathematical optimization to make those deci-
sions. With hands-on practice and a fundamental understanding of some of the most
common analytical models, readers will be able to make smarter business decisions and
higher profits.
Tudor Bodea is a Revenue Optimization Manager in the Global Revenue Management
and Systems Department at the InterContinental Hotels Group in Atlanta, USA. He
earned his Ph.D. in Civil Engineering at the Georgia Institute of Technology, USA, with
an emphasis on transportation systems, logistics, and statistics. He holds a B.S. in Trans-
portation Systems from the Technical University of Cluj-Napoca, Romania and an M.S.
in Civil Engineering from the Georgia Institute of Technology.
Mark Ferguson is a Distinguished Business Foundation Fellow and Professor of
Management Science at the University of South Carolina, USA. He received his Ph.D. in
Business Administration, with a concentration in operations management, from Duke
University, USA. He holds a B.S. in Mechanical Engineering from Virginia Tech, USA
and an M.S. in Industrial Engineering from the Georgia Institute of Technology.
This page intentionally left blank
Segmentation,
Revenue Management,
and Pricing Analytics
Tudor Bodea and Mark Ferguson
First published 2014
by Routledge
711 Third Avenue, New York, NY 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2014 Taylor & Francis
The right of Tudor Bodea and Mark Ferguson to be identified as authors of this
work has been asserted by them in accordance with sections 77 and 78 of the
Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced
or utilized in any form or by any electronic, mechanical, or other means,
now known or hereafter invented, including photocopying and recording,
or in any information storage or retrieval system, without permission in
writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and
explanation without intent to infringe.
Library of Congress Cataloging in Publication Data
Bodea, Tudor.
Segmentation, revenue management and pricing analytics /
Tudor Bodea & Mark Ferguson.
pages cm
Includes bibliographical references and index.
1. Revenue management. 2. Pricing. 3. Market segmentation.
I. Ferguson, Mark, 1969– II. Title.
HD60.7.B63 2013
658.15'54—dc23
2013039078
ISBN: 978–0–415–89832–4 (hbk)
ISBN: 978–0–415–89833–1 (pbk)
ISBN: 978–0–203–80215–1 (ebk)
Typeset in Minion
by Swales & Willis Ltd, Exeter, Devon
CONTENTS
List of figures vi
List of tables viii
Acknowledgments x
Chapter 1 The Ideas Behind Customer Segmentation 1
Chapter 2 Forecasting 8
Chapter 3 Promotion Forecasting 47
Chapter 4 Capacity-Based Revenue Management 77
Chapter 5 Unconstraining 98
Chapter 6 Pricing Analytics 136
Chapter 7 Dynamic and Markdown Pricing 167
Chapter 8 Pricing in Business-to-Business Environments 186
Chapter 9 Customer Behavior Aspects of Pricing 212
Appendix A Dichotomous Logistic Regression 220
Appendix B Advanced Analytics Using R 228
Index 251
v
LIST OF FIGURES
2.1 Measuring Forecast Accuracy 17
2.2 Illustration of the Role of Holdout Samples 19
2.3 Simple Moving Average 20
2.4 Weighted Moving Average 22
2.5 Simple Exponential Smoothing 28
2.6 Double Exponential Smoothing 31
2.7 Additive and Multiplicative Seasonality 34
2.8 Triple Exponential Smoothing 36
2.9 Triple Exponential Smoothing—Initialization of S and T 39
0 0
2.10 Sales of SPSS Manual (2nd Edition) 40
2.11 Sales of SPSS Manual (2nd Edition)—One-Step-Ahead Forecasts 42
2.12 Sales of SPSS Manual (2nd Edition)—1983 Forecasts 44
3.1 Illustration of the Deviations in the Linear Regression Analysis 50
3.2 Estimation of Regression Coefficients via the Least Squares Method 50
3.3 Daily Sales of Ice Cream and Daily Average Temperatures 52
3.4 Trendline Option in Excel 52
3.5 Ice Cream Sales with Trendline 53
3.6 Data Analysis Menu 54
3.7 Regression Pop-up Box 55
3.8 Output of the Linear Regression on the Ice Cream Sales 56
3.9 Illustration of the Maximum Likelihood Estimation Mechanism 57
3.10 Estimation of the Linear Regression Model via Maximum Likelihood 59
3.11 Ice Cream Sales with Promotional Flyer Variable Included 61
3.12 Regression Output with Promotional Flyer Variable Included 61
3.13 Promotion History for a Staple Fashion Item 69
3.14 Sales and Price Plots: Quick Quaker Oats at River Forest Store 71
4.1 Normal Distribution Curve 81
4.2 Protection Level and Booking Limit in a Hotel Context 86
4.3 Nested Fare Class Buckets 88
vi
Figures (cid:129) vii
4.4 Flowchart of Bid Price Revenue Management Logic 93
4.5 Linear Programming Formulation of Revenue Management Problem in
Excel 94
4.6 Solution Report to the Linear Programming Formulation 95
4.7 Solution Report After Total Capacity is Set to Ten Rooms 96
5.1 Royal Hotel Business Booking Curves 100
5.2 Product Availability Progression in an Online Environment 102
5.3 Out-of-Stock Rates for the Sampled Style/Color/Size Items 104
5.4 In-Store Item Availability Measures 105
5.5 Tocher’s Inverse and Inverse Gumbel Cumulative Distribution Functions 114
5.6 Illustration of Demand Unconstraining via Double Exponential Smoothing 120
5.7 Demand Unconstraining via Double Exponential Smoothing 122
5.8 PD E -like Step vs. EM E -Step 129
6.1 The Pricing Analytics Process 138
6.2 Linear Price-Response Function 140
6.3 Uniform WTP Distribution 141
6.4 Density of Demand at Price p from a Uniform WTP Distribution 142
6.5 Density of Demand at Price p from a Normal WTP Distribution 142
6.6 Linear Price-Response Function 143
6.7 Constant-Elasticity Price-Response Functions 144
6.8 Reverse S-Shaped Price-Response Function 144
6.9 Power Price-Response Function 145
6.10 Nonlinear Price-Response Function 147
6.11 Estimated Price Elasticities for Various Goods (Absolute Values) 149
6.12 Profit as a Function of Price 150
6.13 Developing Pricing Capabilities: Process Roadmap 154
6.14 Price-Response Functions and Elasticity Curves 159
7.1 Price-Dependent Demand Profiles 183
8.1 (A) Historical Demand Data for Customized Pricing, (B) Fitted Reverse
S-Shaped Probability Function to Win/Loss Data 189
8.2 (A) Marginal Deal Contribution vs. Unit Price, (B) Win Probability vs.
Unit Price, (C) Expected Profit vs. Unit Price 191
8.3 Online Auto Lender—CHAID Decision Tree 198
8.4 Online Auto Lender—CHAID Logit Bid-Response Functions 200
8.5 Online Auto Lender—Logistic Regression Tree 205
8.6 Auto Online Lender—Bid-Response Functions and Expected Profit
Functions for a Holdout Sample Auto Loan Application 209
9.1 Changes in Consumer Utility as Explained by Prospect Theory 216
A.1 (A) Fitted Line for the Linear Regression Model (B) Fitted Curve for the
Logistic Regression Model 222
A.2 (A) Fitted Line for the Linear Regression Model (B) Variance Plot 223
A.3 Residual Plot 223
B.1 Sales Forecasting at Company X 246
B.2 Price-Response Function, Profit Function and the Optimal Price 248
LIST OF TABLES
2.1 Summary of Forecasting Accuracy Measures 17
2.2 Simple Moving Average 21
2.3 Weighted Moving Average 23
2.4 In-Sample Forecasting Accuracy Measures—SMA(5) 24
2.5 Summary of Forecasting Accuracy Measures 25
2.6 In-Sample Forecasting Accuracy Measures—Simple Exponential Smoothing 29
2.7 In-Sample Forecasting Accuracy Measures—Double Exponential Smoothing 32
2.8 Triple Exponential Smoothing—Initialization of Seasonal Parameters 37
2.9 In-Sample Forecasting Accuracy Measures—Triple Exponential Smoothing 38
2.10 Sales of SPSS Manual (2nd Edition)—Smoothing Parameters and
Forecasting Accuracy Measures 42
2.11 Sales of SPSS Manual (2nd Edition)—1983 Forecasts 43
3.1 Illustration of the Maximum Likelihood Estimation Mechanism 58
3.2 Promotion Planning and Optimization 64
3.3 Additive and Multiplicative Promotion Models 66
3.4 Promotion History for a Staple Fashion Item 68
3.5 Summary Statistics and Model Fit 69
3.6 Log-Transformed and Original Multiplicative Models (Full Model) 74
3.7 Log-Transformed and Original Multiplicative Models (Reduced Model) 74
4.1 Demand Distribution 78
4.2 Expected Profit Calculation for an Order Quantity of 200 bagels 79
4.3 Expected Profits for Each Order Quantity Option 80
4.4 Standard Normal Distribution Table 82
4.5 Fare Class Prices and Distribution Parameter Values 88
4.6 Nested Protection Levels Calculated Using EMSR-b 92
5.1 Daily Sales of Pain de Boulogne at an Albert Heijn Store 107
5.2 Product-Limit Method Applied to the Pain de Boulogne Data 108
5.3 Hourly Sales Rates (Units/Hour) of Pain de Boulogne 109
5.4 Hourly Sales Rates to Cumulative Sales Ratios 111
viii
Tables (cid:129) ix
5.5 Cumulative Hourly Sales and Unconstrained Demand 112
5.6 Cumulative Demand Distribution Function 113
5.7 Parameter Estimates for Tocher’s Inverse Cumulative Demand Distribution
Function 114
5.8 Parameter Estimates for the Gumbel Cumulative Demand Distribution
Function 115
5.9 Operationalization of Averaging Method for Demand Unconstraining 118
5.10 Demand Unconstraining via Averaging Method (AM) 119
5.11 Demand Unconstraining via Double Exponential Smoothing (DES) 121
5.12 Operationalization of the EM Algorithm for Demand Unconstraining 127
5.13 Demand Unconstraining via the EM Algorithm 128
5.14 Operationalization of the PD Method for Demand Unconstraining 132
5.15 Demand Unconstraining via the PD Method 133
6.1 Price-Response Functions 146
6.2 Estimated Price Elasticities at the Industry and Brand Level 148
6.3 Estimated Price Elasticities for Various Goods 148
6.4 Types of Data Used to Make Pricing Decisions 153
6.5 Online Price Experiment Results 158
6.6 Linear Price-Response Function 160
6.7 In-Store Price Experiment Results 160
6.8 Constant-Elasticity Price-Response Model 161
6.9 Customer Survey Results 162
6.10 Logit Price-Response Function 163
7.1 Critical Ratio σ (.)for Assessing the Economic Viability of Price Markdowns 178
7.2 Product Group Demand Analysis 181
7.3 Preseason Optimal Markdown Policies 184
8.1 Output From Logistic Regression on Alpha’s Historical Win/Loss Data 193
8.2 Output From Logistic Regression After Removing Quantity 193
8.3 Bid Characteristics 194
8.4 Online Auto Lender—Data Dictionary 197
8.5 Online Auto Lender—CHAID Logit Bid-Response Functions 199
8.6 Online Auto Lender—Logistic Regression Results 203
8.7 Online Auto Lender—Behavioral Interpretation of the ΔRate Interaction
Effects 204
8.8 Online Auto Lender—Logistic Regression Tree Bid-Response Functions 206
8.9 Online Auto Lender—Behavioral Interpretation of the APR Effects 207
8.10 Online Auto Lender—Characteristics of a Holdout Sample Auto Loan 208
8.11 Online Auto Lender—Bid-Response Functions and Expected Profit
Functions for a Holdout Sample Auto Loan Application 209
A.1 Dichotomous Logistic Regression Model 225
B.1 Model Selection Using Forecast Accuracy Measures 245
B.2 Long-Term Forecasts and Prediction Intervals 247
B.3 Product-Level Optimal Prices and Expected Profits 248