THE LAST MILE CHALLENGE: EVALUATING THE EFFECTS OF CUSTOMER DENSITY AND DELIVERY WINDOW PATTERNS by Kenneth K. Boyer The Ohio State University Andrea M. Prud’homme The Ohio State University and Wenming Chung University of Texas at El Paso INTRODUCTION The last 10 years have seen an explosion of consumer direct businesses such as grocery, office supply, package and pharmaceutical delivery. Examples of businesses with consumer direct delivery include grocers such as Freshdirect and Ocado (Delaney-Klinger, Boyer, and Frohlich 2003), office supply retailers such as Office Depot, Staples and Office Max (all of which are in the top 10 Internet retailing business ranked by sales (Love and Peters 2006) and package delivery companies such as FedEx, UPS, Airborne and Deutsche Post. To illustrate, FedEx ground delivers over $5 billion in sales per year in business-to-business packages, with FedEx Home Delivery providing ground-based delivery services to residences. While FedEx does not break out the volume of home delivery versus business delivery packages, it does note in recent financial results (March 27, 2007) that growth in the home delivery segment is stronger than overall growth of 9% in ground delivery. The operational challenges underlying consumer direct delivery are daunting. Numerous companies have failed due to operational and logistical problems encountered with delivering orders directly to customers. The fulfillment process for consumer direct orders can be broadly characterized as consisting of three stages (Campbell and Savelsbergh 2005; Delaney-Klinger, Boyer, and Frohlich 2003): (1) order acceptance, (2) order selection and fulfillment and (3) order delivery. Each of these stages is critical to providing excellent customer service at a cost the customer is willing to pay. The focus in this study is on the third stage—order delivery. The delivery of the final product to the customer’s door is logistically challenging due to a number of factors and potentially very expensive (costs for a single delivery of groceries run between $10 and $20 per order according to Boyer, Frohlich, and Hult 2004). There has been substantial research on how to route and schedule vehicles from an algorithmic point of view toward solving a particular problem. What is not clear is the shape of the efficiency curve and the interaction of two key factors affecting routing efficiency: customer density and delivery window length. In other words, what is the relative change in efficiency/cost as customer density increases or delivery windows are lengthened? This is a problem of critical importance to customer direct businesses. The spectacular collapse of Webvan, the online grocer that went bankrupt after reaching a market capitalization of over $5 billion while only producing sales of less than $400 million, was partly tied to its promise of delivery within a pre-specified window of 30 minutes. While this was great from a customer perspective, it proved to be a huge logistical challenge and cost (Boyer, Frohlich, and Hult JOURNAL OF BUSINESS LOGISTICS, Vol. 30, No. 1, 2009 185