Customer service has assumed strategic importance in most manufacturing environments. Corporate reputations often depend on how reliably promised lead times are met. A goal of filling orders within a service window of T time units is often encountered in practice. This goal ignores the differences among types of units, treating all customers as equally important. Keeping finished goods inventory of all items is only one, and often not the best, way to achieve this objective. High-demand items naturally have safety stock assigned to them, but in many organizations there are so many very-low-demand items that keeping any stock in these items is prohibitively expensive. Customer service can be maintained for these low-demand items by giving them higher production priority when a demand occurs. In this paper, stochastic analysis and simulation are employed to test the merit of this idea. Changes in management structure to allow this type of system are discussed.
