Decentralized Networks: the Impact on Return Policies for an Assemble-to-Order System

The last years have displayed an impressive e-commerce boom. To protect the customer stricter policies have been imposed to e-tailers, most important of which is the return policy. Companies have to deal now with more complicated management of inventories due to uncertainty in the reverse flow of materials. Moreover, returns do not re-enter the supply chain immediately after leaving the customer. Although there is a broad literature on returns management and on how companies should handle returns, mathematical models have not yet addressed efficiently the customer’s needs. Assemble-to-order systems build products to customer’s requirements. However, returns in such systems cause losses due to the uniqueness of the products. The present paper implements a customer-friendly return policy, at a component level, for an assemble-to-order system. Finished products are assembled in response to customer demands; the inventory is kept at a component level; once the customer reports the intention of returning a product because of a particular component, the company will take care of the dissembling and shipment processes and costs and pays the customer a return credit; no time limits are imposed for returns because any returned product crosses refurbishing processes and quality checks. We describe a simple network (manufacturer, two suppliers, third-party logistics provider, customers) and derive both the optimized profit and the optimal base-stock inventory level for different return rates. Through a simulation study we show that a return policy as such brings improvements to the company’s profits, for certain levels of return rates.