Ecommerce distribution: defining a new cost model

Abstract

This project introduces a novel approach for defining a cost model for ecommerce shipment distribution. The key objective is to determine the functionality of the advanced model, in order to obtain the cost of distributing ecommerce shipments to customers and enable the assessment of the cost impact for different business decisions. The cost model is based on the process costing models and the allocation criteria is determined for each specific process. In this regard, process costing is utilized as it is an organic full cost model reflecting all the costs of the production process, which in turn reflects the structure and organization of the different processes involved in the distribution. The model advanced in this project is built on the data gathered from publicly available information on sortation, labor and vehicle cost. Four different distribution ecommerce business decisions have been analyzed utilizing the model proposed: direct injection skipping intermediate hubs to reach the delivery unit, offer of different delivery windows to customers, impact of increasing pickup point’s adoption rate and enabling mailbox delivery. For direct injection, results show that there is no cost saving from skipping intermediate hubs for all delivery units, but there is for the units which concentrate the highest demand. Direct injection cost impact requires an individual analysis lane by lane as cost variation depends on fill rate, distance, vehicle cost, and sortations skipped for each specific case. For scheduled deliveries, the cost increases observed in the last mile versus a full day window are as follows: 30% for 4 hours windows, 60% for 3 hours, 80% for 2 hours and reaching up to 110% for 1 hour window. On the other side, increasing the adoption rate of pickup points or mailbox delivery increases the driver productivity and reduces last mile cost. Finally, by enabling mailbox delivery instead of door step delivery a 6% reduction in the total cost has been observed. By applying the model advanced in this project, we could further analyze the cost impact of different business decisions and compare the results with other metrics such as: sales uplift, customer experience and time from click to delivery.