Stefan Mohr Store Performance Optimization

Size: px
Start display at page:

Download "Stefan Mohr Store Performance Optimization"

Transcription

1 Stefan Mohr Store Performance Optimization

2 GABLER EDITION WISSENSCHAFT

3 Stefan Mohr Store Performance Optimization Demand and supply side implications Mit einem Geleitwort von Prof. Dr. Arnd Huchzermeier Springer Fachmedien Wiesbaden GmbH

4 Die Deutsche Bibliothek- CIP-Einheitsaufnahme Ein Titeldatensatz für diese Publikation ist bei Der Deutschen Bibliothek erhältlich Dissertation Wissenschaftliche Hochschule für Unternehmensführung Koblenz, Auflage Mai 2002 Alle Rechte vorbehalten Springer Fachmedien Wiesbaden 2002 Ursprünglich erschienen bei Deutscher Universitäts-Verlag GmbH, Wiesbaden 2002 Lektorat Ute Wrasmann I Nicole Schweitzer Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verla.gs unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Ubersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme, dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei zu betrachten wären und daher von jedermann benutzt werden dürften. Umschlaggestaltung: Regine Zimmer, Dipi.-Designerin, Frankfurt/Main Gedruckt auf säurefreiem und chlorfrei gebleichtem Papier ISBN ISBN (ebook) DOI /

5 Foreword Consumer value can be created or enhanced through low cost product afferings at high service Ievels. This is one of the primary goals of all ECR (Efficient Consumer Response) initiatives around the world. In general, consumer needs are better satisfied when retailers offer the right mix of products, with a right price at the right time. Uncertainty in market demand and random consumer choice behavior have major influence on product availability and supply chain costs. Hence, each day, store managers have to cope with the problern of reducing dead weight shelf space due to safety stocks. In this thesis, Stefan Mohr approaches this category management problern by providing i) a forecasting tool for store sales accounting for both demand uncertainty as weil as random consumer choice behavior and ii) a product portfolio optimization tool for efficient assortment planning under risk that can be applied, for example, to new product introductions, SKU (stock keeping units) rationalization or category optimization at the brand, the product group, the category or even at the store Ievel. The main results of the thesis can be summarized as follows: A high Ievel of variety increases randomness of consumer choice. The analysis of an extensive data set provided empirical evidence for the fact that the number of SKUs significantly increases randomness in consumer choice behavior. In particular, categories in which SKUs mainly differ by size show greater randomness of consumer choice. Similar products in low price, high variety categories exhibit greater randomness of consumer choice than dissimilar ones. Using the methodology of hypothesis testing, it can be shown that the high number of me-toos entering the market each year not only might fail to match consumer

6 VI expectations but also Iead to higher inventory costs, stock-outs respectively. Supply chain strategies like delayed product differentiation that increase similarity of products Iead to higher total supply chain costs. In an environment where consumers are increasingly confused about the number of variants and their similarity, operational concepts like mass customization or standardization have to be applied carefully. It is of vital importance to integrate consumer behavior in supply chain optimization concepts and to evaluate their impact on total supply chain efficiency. Applying the risk-return approach Ieads to an optimization of the trade-off between category productivity and supply chain costs. Marketing managers might argue that increasing the Ievel of variety within a store might Iead to higher category productivity (sales per m). Even though, Stefan Mohr was able to show that this might be true, he clearly demonstrates that there is a trade-off between category productivity and total supply chain costs. The riskreturn approach proved to be a powerful tool in order to optimize this tradeoff and hence, derive an efficient product portfolio. The results are promising: 8.3% increase in category productivity, yielding an equal coefficient of variation, randomness of consumer choice respectively. This thesis is noteworthy for a number of reasons. First, the data set utilized is current and quite comprehensive. It covers important food categories (salted snacks, sweet snacks, pasta and coffee) and thus compares quite weil to other empirical studies conducted in the retail industry. Second, the reviewed Iiterature covers recent streams both in the operations and the marketing literature. Third, the two-stage, twoproduct supply chain model is derived from other recent publications and extended to account for consumer choice behavior. Fourth, the portfolio model used is by itself not new, but adapted to the retailing data set. What is novel though is the explicit derivation of the returns and the variance as a function of the number of SKUs being offered. Fivth, the conclusions are convincing: early product customization may be superior when demand side factors are explicitly being accounted for. Sixth, the

7 vii portfolio model provides an effective way for store managers to optimize a large set of products in a store. Seventh, to date, few such models exist; they are mostly limited to a particular category only. Overall, the complex and difficult material is skillfully presented. Thus, the thesis will - at least in my opinion - be recognized as a very valuable reference by both academics and senior executives in operations or supply chain management as well as in marketing alike. Prof. Dr. Arnd Ruchzermeier

8 Preface During the last decade, retailing has gone through a complex transformation process triggered by increasing consumer requirements in terms of product variety and price awareness. The approach presented in this study shows how these market trends influence supply chain systerns in the food retailing industry. The theoretical rnodel was validated with data from a major German food retailer and the results are promising. The book is the printed edition of my dissertation. It was completed during my participation in the doctoral program of Roland Berger Strategy Consultants. As doctoral student I was working at the Department of Production Management at the WHU Otto Beisheim Graduate School of Management in Vallendar, Germany. Many people have contributed to this work. I would like to thank all of them, even though only a small nurnber can be named here. My advisor, Professor Ruchzermeier supported not only the whole process but also my stay at the Kellogg Graduate School of Management. I am very grateful to Professor Sunil Chopra, Department of Decision Seiences at the Kellogg Graduate School of Management for his valuable suggestions and for co-grading my work. I owe a lot of thanks to my friend Professor Kar! Schmedders, Department of Decision Seiences at the Kellogg Graduate School of Management for his constant motivating comments and his great support especially regarding the empirical part of the dissertation. Analyzing the data set is an important part of this work. Therefore I want to thank

9 X Mr. Hausteiner for providing the data set. Most of all I want to thank my parents. Their understanding support and love are the basis for all of my achievements not only this dissertation. Stefan Mohr

10 Contents Foreword Preface 1 Introduction 2 Store Performance Optimization in Food Retailing 2.1 Strategie Analysis ofthegerman Food Retailing Industry Growth Perspectives Profit Perspecti ves 2.2 Business Strategies Efficient Consumer Response Category Management 2.3 Efficient Assortment Roadmap of the Dissertation. 3 Drivers of U ncertainty - Empirical Results 3.1 The Data The Products Category Characteristics Category Structure Drivers of Uncertainty at the product group Level Model and Methodology V ix

11 xii Results Model Validation 3.4 Drivers of Uncertainty at the SKU Level Model and Methodology Results Model Validation 3.5 lmplications of the Empirical Results Product Variety Management 4.1 Product Variety- Driver of Demand Uncertainty Variety Management - The Operations Perspective Delayed Product Differentiation The Means of Effecting Postponement The Operating Mode The Supply Chain Cooperation Mechanism Product Differentiation Considering Consumer Choice Behavior - The Benefit of Early Differentiation 5.1 The Two-Stage Supply Chain The Two-Stage Supply Chain Model Considering Consumer Choice Behavior The Two-Stage, Two Product, Infinite Horizon Model Considering Consumer Choice Behavior The Early Differentiation System The Late Differentiation System Delayed Product Differentiation - Supply Side lmplications Delayed Product Differentiation - Demand Side lmplications The Benefit of Early Differentiation

12 xiii The Optimal Point of Product Differentiation 5.4 lmplications of the Model Efficient Assortment under Risk: A Portfolio Approach 6.1 Product Portfolio Models Principles of the Mean-Variance Approach Product Objectives The Efficient Set. 6.3 Efficient Assortment based on Mean-Variaace Optimization Drivers of Sales Drivers of Risk Efficient Assortment through Quadratic Programming Numerical Example Alternative Model Salutions 6.4 Discussion of Results Store Performance Optimization Strategy 7.1 Store Performance Optimization Approach 7.2 Store Performance Optimization Opportunities..... A Product Category Data B CV - Results, The Reduced Model C CV - Results, The Full Model D Model Results - Sales per m 2 E Example Nonlinear Programming F Bibliography

13 List of Figures 1.1 Strategie framework of the thesis German food market- Sales volume and growth rate 2.2 Profitability of food retailers and manufacturers in Germany 2.3 Food's share of total consumer expenditure. 2.4 Price decline January to December Top ten German food retailers 2.6 Spacc per store 1990 to Total benefit from ECR as % of costs 2.8 Focus areas of ECR The Category Management process 2.10 Status of implementation of Category Management tactics 2.11 Objectives of the Efficient Assortment approach The strategy of delaying the point of product differentiation 2.13 Roadmap of the dissertation Descriptive Statistics- Location of the stores. 3.2 Product structure Scatterplot CV /SKU for salted snacks- the Full Model 3.4 Scatterplot CV /MAG% for salted snacks- the Full Model 3.5 Scatterplot CV /SKU for swcet snacks- the Full Model. 3.6 Scatterplot CV /MAG% for sweet snacks- the Full Model 3.7 Scatterplot CV /SKU for pasta- the Full Model. 3.8 Scatterplot CV /MAG% for pasta- the Full Model

14 xvi 3.9 Scatterplot CV /SKU for coffee- the Full Model Scatterplot CV /MAG% for coffee- the Full Model Scatterplot CV /SKU for salted snacks- the Unbiased Model Scatterplot CV /MAG% for salted snacks- the Unbiased Model Scatterplot CV /SKU for sweet snacks- the Unbiased Model Scatterplot CV /MAG% for sweet snacks- the Unbiased Model Scatterplot CV /SKU for pasta- the Unbiased Model Scatterplot CV /MAG% for pasta- the Unbiased Model Scatterplot CV /SKU for coffee- the Unbiased Model Scatterplot CV /MAG% for coffee- the Unbiased Model Form postponement A multi echelon system Decentralized supply chain system The make-to-order system The two differentiating-feature product line The supply chain systems lmplications of the snacks supply chain on the supply chain system The value of Early Differentiation - a numerical example Impact of p on a The optimal point of product differentiation a~ and the optimal point of product differentiation p and the optimal point of product differentiation Efficient set and space of possible solutions Efficient set, indifference curve and point of tangency 6.3 Scatterplot Sales-SKU% for salted snacks. 6.4 Scatterplot Sales-SKU% for sweet snacks 6.5 Scatterplot Sales-SKU% for pasta 6.6 Scatterplot Sales-SKU% for coffee

15 XVll 6.7 Mean-Variance Pasta; Store Gap between Model and Practice Optimization objects Store Performance Optimization approach CV /SKU scatterplot Priorities for optimization approach Sales per m 2 /SKU scatterplot Sales per m 2 /CV efficient frontier 152

16 List of Tables 3.1 Business parameters of the stores 3.2 Category characteristics Descriptive statistics - product group Ievel 3.4 R 2 and adjusted R 2 in the Reduced Model 3.5 R 2 and adjusted R 2 in the Full Model Number of product categories -biased versus unbiased- 3.7 R 2 and adjusted R 2 for the Unbiased Model 3.8 Descriptive statistics - SKU Ievel Mean and standard deviation of the coefficient of variation for the products differing by size only Mean and standard deviation of the coefficient of variation for the products differing by fiavor H 0 evaluation H 0 evaluation for MAG, FREQ, PRICE Overview on research about delayed product differentiation Implications of the model of Lee and Tang [1997] about delaying differentiation R 2 and adjusted R 2 - Sales per m