A Practical Approach to Diagnosing and Tuning a Statistical Forecasting System

Most, if not all, commercial enterprises require some form of demand forecasting for financial and operations planning. For financial planning, a high level, aggregate forecast (e.g., in dollar value) of major product groups or geographies is sufficient. For operations planning, a more detailed forecast, such as forecast by product or even by product-location, is necessary. A manufacturing enterprise employing a make-to-stock strategy needs a demand forecast to plan what products and how much of each to build. A make-to-order manufacturer uses a demand forecast to plan the purchase of parts and materials and its production capacity. A retailer needs a demand forecast to determine how much of each product to stock at the different retail locations. Other service enterprises utilize a demand forecast to plan and locate their capacity (for both labor and equipment). We focus on the latter situation in this chapter, namely detailed, product level forecasts that drive the planning of a supply chain.