A decision system for routing returned product to the optimal recovery channel

Abstract

Dell, a leading computer manufacturer, must deal with systems returned from its customers. Historically, it has refurbished most of its returned systems for resale on its Dell Outlet website. While this has provided high net recoveries (revenue less incurred costs) compared to its peers, Dell believes there is ample opportunity in cannibalizing some returned systems for the piece parts (i.e. "teardown"). These harvested piece parts can be used to service field systems, repair refurbished systems, or directly sold to customers as spare parts. Dell is concerned about ensuring an optimal disposition of system to teardown vs. direct resale. Written as part of research internship at Dell, this paper proposes, simulates, and evaluates a decision support system to address the question of disposition. The decision engines use historical data and statistics to estimate net recoveries in resale and forecasted demand to estimate net recoveries through teardown. Linear regressions were found to have poor power in predicting overall net recoveries; however, simple heuristics were found to identify likely low recovery systems. Overall, the implementation of the decision support system will drive improved net recoveries, with savings estimated to be greater than $1 million annually.