Planning and control for operations takes place at several levels. Therefore, it is unlikely that one kind of forecast can serve all needs. We require forecasts of different time spans to serve as the basis for operating plans developed for different planning horizons. These include (1) plans for current operations and for the immediate future, (2) intermediate range plans to provide for the capacities of personnel, materials, and equipment required for the next 1 to 12 months, and (3) long range plans for capacity, locations, changing product and service mix, and the development of new products and services.
The horizon of the forecast must be matched with the decision the forecast will month forecast would be valueless. On the other hand, it is unwise to select a forecasting model for daily or weekly decisions that has an acceptable error on a monthly or annual basis but poor accuracy for daily or weekly projections. Therefore, a major criterion for model selection is the match between decision time, forecast horizon, and forecasting accuracy.
When developing plans for current operations and for the immediate future (e.g. how many jobs should be released to a shop on a given day or how many tellers should be assigned during the lunch hour on a Friday), the degree of detail required in forecasting is high. The forecast data should be available in a form that can be translated into demands for material, specific labor skills, and time usage of specific equipment. Therefore, forecasts of gross dollar demand, demand by customer or client classification, or demand by broad product or service classifications are of limited value for short term daily operational decisions.
For such short term decisions, we need forecasting methods that are relatively inexpensive to install and maintain and that can be adapted to situations involving a large number of items to be forecast. This means that the data input and storage requirements should be modest and that computerized methods are a likely mechanism for updating forecast data as needed.
For intermediate range plans, such as plans for monthly production levels or work force levels, useful forecasts will probably be aggregated by product types. Detailed forecasts for each individual item may not be necessary. Also, since the relative frequency of forecasts is lower and the number of different product types for which forecasts are made is smaller than is smaller than is the case for the short term decisions, forecasting methods that require modest cost and effort could be employed.
Long range plans for capacity, location, and new technologies for plant and equipment require forecasts for the next 1 to 10 years. Because of the longer time involved, these forecasts will necessarily have greater uncertainty and a lower degree of accuracy. Often, the mechanical application of a model is not sufficient to obtain the desired forecast, and subjective inputs from the managers and other knowledgeable people are required. The methods of forecasting should therefore be able to integrate objective data and subjective inputs.
Forecasting methods can be divided into three main categories
1. Extrapolative or time series methods
2. Causal or explanatory methods
3. Qualitative or judgmental methods.
In some situations, a combination of methods may be more appropriate than a single method.
Extrapolative methods use the past history of demand in making a forecast for the future. The objective of these methods is to identify the pattern in historic data and extrapolate this pattern for the future. This process might seem like driving while looking only through a rear view mirror. However, if the time horizon for which the forecast is made is short, extrapolative methods perform quite well.
Causal methods of forecasting assume that the demand for an item depends on one or more independent factors (e.g. price, advertising, competitorâ€™s price etc). These methods seek to establish a relationship between the variable to be forecasted and independent variables. Once this relationship is established, future values can be forecasted by simply plugging in the appropriate values for the independent variables.
Judgmental methods rely on expertsâ€™ (or managersâ€™) opinion in making a prediction for the future. These methods are useful for medium to long range forecasting tasks. The use of Judgment in forecasting, at first blush, sounds in-scientific and ad hoc. However, when past data are unavailable or not representative of the future, there are few alternatives other than using the informed opinion of knowledgeable people. There are, however, good ways and bad ways to solicit judgments for making a forecast. We will discuss some approaches that structure and formalize the process of soliciting judgment so that individual biases are minimized. Often, in operations situations, judgmental methods are employed in conjunction with extrapolative or causal methods.