6. Finally, regression is used most often for medium-term, followed regression by long-term, forecasting horizons. This is consistent with theoretical reasoning that postulates that in the medium- and long-terms, more emphasis should be placed on understanding the variables to be forecast and the factors that in°uence them.

Such understanding can be substantially aided with regression analysis.

vThe level at which di®erent forecasting methods are used is shown in Table 11-4. The jury of executive opinion method is used more than any other method for all forecasting levels except product forecasting, while customer expectations and sales force composite v1. Measuring the wrong thing: In forecasting we often need to estimate demand, but demand data are rarely, if ever, available.
Thus, instead of measuring demand we measure such things as orders, production, shipments, or billings. It, of course, is obvious that such proxies of apparent demand introduce systematic biases in measuring the eal" demand and therefore decrease forecasting accuracy.