Abstract
This article investigates two alternative methods for estimating standard deviations of demand in inventory control systems. One method requires only the average demand rate (in units), while the second is based on the average number of units requested per customer order and the average customer order rate. For a sample of 489 products with 48 months of demand data, the second method pro vides more accurate estimates of standard deviations than does the first, with the largest improvement in the low-and medium-volume categories. For one numerical example, the improvement offers an expected reduction in total inventory-associated costs of 3.94 percent.