The effect of optimizing warehouse locations on omnichannel designs

Abstract

Omnichannel distribution integrates stores and warehouses to fulfill online and store demand. This research studies the performance of four channel designs: (i) ship-from-store, (ii) ship-from-warehouse, (iii) ship-from-store-and-warehouse, and (iv) ship-from-warehouse with backhauling from the store inventory to the warehouse. Our focus is on omnichannel retailers who carry a single product category. A profit maximizing mixed integer linear programming model (MILP) is presented to determine the optimal locations and capacities of omnichannel warehouses, and the product flows from suppliers to warehouses, warehouses to the stores, and warehouses and stores to online markets. The online demand earned by the firm in each market is determined endogenously in the model as a function of the shipping time (distance) and the level of competition in the online market. Different warehouse capacities capture economies-of-scale in operating costs. To cover a wide range of different products, computational results are presented for five product categories, each considered separately, with widely differing cost characteristics. Results show that the ship-from-warehouse design generates greater profit than ship-from store for four product categories. However, the ship-from-store-and-warehouse design is better than both of the more restrictive ship-from-store or ship-from-warehouse designs, with the advantage growing as the percentage of online demand increases. Results also document increasing benefits from optimizing the number, locations and capacities of omnichannel warehouses as the demand increases. The magnitude of the profit advantage of different omnichannel designs is shown to depend strongly on the product category and percentage of demand that is online. © 2021 Elsevier B.V.