Managing backlog variation in order fulfillment: The case of Internet retailers

Matching supply, demand, and capacity is made more costly and less efficient when demand variation is amplified into work-in-process, or backlog, variation. Existing research has focused on managing variations in backlog quantity. This paper extends that research by examining how variation in fulfillment leadtimes can be managed effectively in the context of Internet retailing. Computer simulation, grounded in actual operational data, is used to test hypotheses concerning the links between order fulfillment rules and backlog variation. Two rules, based on due dates for the release of inventory and inventory thresholds, are found to be effective in controlling backlog variation amplification through the accumulation of orders or the postponement of order fulfillment. The due date release rules always generate the highest fill rate and timeliness in fulfillment, and, in most cases, the lowest total costs. This work highlights the need to consider both quantity variation and leadtime variation when managing order fulfillment systems.