Designing and Managing Manufacturer Distribution Systems

This cumulative dissertation consists of three research articles, two of which have already been published in scientific journals. The third one is currently under review at another renowned scientific journal. All three articles address the design and management of a manufacturer distribution system in various contexts. Unlike the traditional marketing literature that investigates distribution strategies or distribution intensity (i.e., wide spread availability of products) generally from the retailers’ point of view, this thesis focuses on the manufacturer’s perspective. Technology advancement and changing customer requirements have driven many manufacturers to utilize multiple distribution channels that use both direct and indirect sales channels. Manufacturers that adopt a direct channel may compete with their existing independent intermediaries (retailers, wholesalers), and thus a channel conflict may emerge between the parties. This type of competition is referred to as manufacturer encroachment. The first paper (Chapter B) responds by providing a systematic and exhaustive literature review of multi-channel distribution systems wherein the manufacturer competes with its independently-owned intermediaries and develops a conceptual framework for analyzing scenarios where a manufacturer intrudes into the retailers’ market, investigates determinant factors that induce the manufacturer to adopt a direct channel, explores possible mechanisms that the manufacturer may use to mitigate the channel conflict, and studies operational decision problems in a multi-channel distribution setting. Moreover, it explores potential research gaps and proposes future research directions. The research gaps identified in Chapter B pave the way for Chapter C which investigates the configuration of a last-mile distribution network for an encroaching manufacturer who sells multiple products to different customer segments. Today’s developed information technologies and the third-party logistics sector enable manufacturers to modify their distribution channels to meet rising customer expectations and gain from potential competitive advantages. Therefore, there is a need to explicitly consider customer channel preferences and delivery services in addition to strategic, tactical, and operational decisions to help manufacturers realize a cost-effective supply chain. Given the limited scope of earlier research, the second article analyzes three distribution network design choices of a manufacturer (single channel, multi-channel, and omni-channel) through the lens of location and routing decisions. In particular, the model formulated for the omni-channel distribution network fills a significant gap and contributes to the related literature. Here, we propose an integrated optimization model that includes a location-routing problem for designing of a combined two-echelon supply chain for an omni-channel distribution system with fragmented customer demand met via multiple shopping and delivery options. We also incorporate a customer service-level constraint. We further develop a decomposition solution method to solve large-scale instances efficiently. Based on our computational study, we conclude that an omni-channel distribution system is a feasible strategy that can reach more customer segments at low logistics costs. Our findings also show that an increase in the number of ‘buy online pick-up in-store’ (BOPIS) positively impacts the total logistics cost. Finally, Chapter D investigates an internal distribution problem in a warehouse handling returns of items sold online. A case company selling mainly apparel processes returns at two types of workstations: refurbishing and recycling. To improve the processing of returned items, the company plans to implement lane-guided transport (LGT) vehicles that pick up boxes of returned items at a depot and drop them off at workstations by following optical markers on the floor. In this context, we formulate a mixed-integer programming (MIP) model seeking an optimal solution to the following routing problem: Given a set of stations and multiple depots, which station should do what type of work, and what is the optimal number of LGT vehicles and their routes? Since the MIP model is an NP-hard problem, we develop a three-stage heuristic decomposition scheme that solves instances obtained from industry data to near-optimality in a reasonable solution time. Furthermore, to test to what extent our findings from the MIP model, particularly the number of vehicles, are optimal for real-world operations, we conduct a simulation study in addition. Our results show that the number of depots has a notable impact on the overall system performance, while the depot location has only a small influence on system efficiency.