Data Level
The level of detail of the data refers to the planning detail required. If we are forecasting annual dollar sales by product line for a marketing plan, data at the same level and time horizon are fine. However, if we also need weekly unit forecasts by stock-keeping unit by location (SKUL), annual product line data will be of little help. Since we need sales forecasts for a number of different functional plans, data at the level of detail corresponding to each of these planning needs are necessary.
This level of detail is called the forecasting hierarchy and is defined as all the planning levels and time horizons and intervals at which forecasts are needed. Figure 5.6 illustrates one such forecasting hierarchy for a company with which we have worked. In this company, the logistics function needs forecasts by week, by SKUL (SKU); the production and purchasing functions need forecasts biweekly, by stock-keeping unit; the sales function needs dollar sales by product by quarter; the marketing function needs annual dollar sales for the next year by product line and for the next five years by division; and finance needs annual dollar sales for the next five years by strategic business unit (SBU) and for the overall corporation. The data detail required for developing a forecast for each of these functions must match each planning level, horizon, and interval. The figure is drawn as a triangle to rep- resent the number of forecasts that are required at each level of the hierarchy. Many more forecasts are required at the SKUL level than at the SKU level, more at the SKU than at the product level, and so on.