Logisticians collect products into groups to differentiate service levels among them or simply to manage them differently. These groups and the individual items within them form various demand patterns over time. When demand is "regular," it will typically be represented by one of the general patterns shown in Figure 8-1. That is, demand patterns can usually be decomposed into trend, seasonal, and random components. As long as random variations are a small portion of the remaining variation in the time series, good forecasting success is usually obtained from popular forecasting procedures.

When demand for items is intermittent, because of low volume overall and a high degree of uncertainty as to when and at what level demand will occur, the time series is said to be lumpy, or irregular, as in Figure 8-2. This pattern is often found in products that are phasing in or out of the product line, demanded by relatively few customers, divided among many stocking locations so that demand at each location is low, or derived from the demand for other items. Such demand patterns are particularly difficult to forecast using the more popular techniques. However, because such items may represent as much as 50 percent of the products a firm handles, they represent a special demand-forecasting problem for the logistician.