Summary

[Page 702 ( continued )]

We have presented several methods of forecasting that are useful for different time frames . Time series and regression methods can be used to develop forecasts encompassing horizons of any length time, although they tend to be used most frequently for short- and medium-range forecasts. These quantitative forecasting techniques are generally easy to understand, simple to use, and not especially costly, unless the data requirements are substantial. They also have exhibited a good track record of performance for many companies that have used them. For these reasons, regression methods, and especially times series, are widely popular.

When managers and students are first introduced to forecasting methods, they are sometimes surprised and disappointed at the lack of exactness of the forecasts. However, they soon learn that forecasting is not easy and exactness is not possible. Forecasting is a lot like playing horseshoes: It's great to get a "ringer," but you can win by just getting close, although those who have the skill and experience to get ringers will beat those who just get close. Often forecasts are used as inputs to other decision modelsfor example, in inventory models, the subject of the next chapter. The primary factor in determining the amount of inventory a firm should order (i.e., prepare to have on hand) is the demand that will occur in the futureforecasted demand.