(1.3) Exponential Smoothing Sales Forecast
The final page of the media file, labeled "1.3 - Exponential Smoothing," features a complete example of a sales forecast using the exponential smoothing sales forecasting method. This method is completely different from the previous two sales forecasting methods.
First, the required equation to calculate a forecast with the exponential smoothing method is listed, followed by an explanation of the variables. To begin calculating the sales forecast with the exponential smoothing method, the first period's forecast must be known. This forecast should be already established or estimated, if it is a new company. Then, a company must figure out a good value for Alpha, α, which is the smoothing constant.
Alpha is multiplied to the difference between the previous forecast and previous actual sales. It works to allocate a portion of the gain or loss in sales to the previously forecasted value, representing the actual percentage of increase or decrease of sales due to natural market fluctuations. Alpha usually ranges between .3 - .4 for many businesses; however the value must remain between zero and one.
Table 6 shows the actual sales, the calculated forecast, and the difference between those two values; which is referred to as the "error" in the forecast. Table 7 explains the calculations for each period to calculate the individual exponential smoothing forecasts, in this example.
For more information on the three forecasting methods and the "best" forecasting method, please continue reading on Page 3.