University of St.Gallen
research platform alexandria
search projects
browse projects
by person
by year
by project type

Demand-based on-shelf availability - Integrating intraday store sales patterns in retail inventory management

abstract Providing items on retail shelves, i.e. on-shelf availability, lies at the core of retail supply chain management. Stocking too many items on retail shelves incurs excess handling cost, cost of capital, markdowns and write-offs. Stocking too little means risking stockouts (Sloot et al., 2005; Zentes et al., 2007). Stockouts have a negative impact on retailers and brand manufacturers, both directly on sales and profit and indirectly on customer satisfaction and store or brand loyalty respectively (Campo et al., 2000, 2003; Fitzsimons, 2000). For retailers, stockouts cause lost sales, dissatisfy consumers, diminish store loyalty and jeopardize marketing expenses (e.g. Marketing Online, 2004; EMFI, 2008).

Today, most retailers apply inventory management that target the on-shelf availability of any product at any time throughout store opening hours (Aastrup and Kotzab, 2009; Corsten and Gruen, 2003; McKinnon et al., 2007; Tan and Karabati, 2004). Yet, surveys on stockout rates over the past decade report that between 4 and 10 percent of the total stock-keeping units (SKU) at supermarkets are typically out of stock (EFMI, 2000; Gruen et al., 2002; ECR Europe, 2003; Roland Berger & Partner, 2003; IGD, 2004, 2005, 2006, 2007; Gruen & Corsten, 2008; Hofer, 2009). Despite efforts by retailers and manufacturers, these figures have remained constant over the past decade. During this time, scholarly and managerial efforts towards improving OSA have been directed primarily at overall improvements in stockout levels (Corsten & Gruen, 2003; ECR Europe, 2003; McKinnon et al., 2007; Aastrup & Kotzab, 2009). While this body of research has contributed significantly to our understanding of stockouts, the current advances treat demand as stationary (Tan & Karabati, 2004). This rigid assumption has led to the implementation of OSA policies that target undifferentiated on-shelf availability throughout store opening hours, as today's retailers protect themselves against unknown intraday demand variation. However, such policies seem inefficient considering the cost associated with maintaining a fixed level of on-shelf availability for all items, all day, all year. Also, OSA is only relevant when actual demand occurs. Shopper behavior research indicates the existence of differences in shopper purchasing habits over the day (e.g. Geiger, 2007; Reimers & Clulow, 2009), and the intraday effects of stockouts found on shopper purchasing behavior (Lee, 2004). This raises the question whether static daily OSA levels represent adequate targets for retail inventory management (van Woensel et al., 2007; van Donselaar et al., 2010).

To address this question, this study seeks to explore setting intraday OSA levels designed to ensure availability modelled on intraday store sales patterns. Thereby, this study's contribution is threefold. First, using econometric analysis, it aims to foster our understanding of intraday store sales patterns to identify at what point in time an item’s OSA is relevant to shoppers and when it is less relevant. Second, it tries to identify logistically relevant attributes, such as case-pack size and sales velocity that affect these patterns. Third, the study takes a modelling approach to improve inventory control rules by accounting for intraday purchasing patterns, which previous research has not regarded. Theoretically, it applies service-dominant (S-D) logic as the foundation for designing on-shelf availability on customer purchasing patterns. Methodically, the study uses on econometrics and quantitative empirical modelling.
keywords On-shelf availability, inventory management, demand modeling
partner Thomas W. Gruen, Ph.D., Professor of Marketing, College of Business & Administration, University of Colorado at Colorado Springs ; dr. Tom Van Woensel, Associate Professor of Operations Management and Logistics, School of Industrial Engineering, Eindhoven University of Technology
type dissertation project
status completed
start of project 2011
end of project 2012
additional informations
topics Service-Dominant logic, retail supply chain management, and operations
methods Literature review, derivative propositions, time-series analysis, modeling
contact Joachim C.F. Ehrenthal