
Part 2. MM4XL Tools > 1. Strategic Tools > Forecast Manager > 1. How To Run Forecast Manager > Page 4: Special Events Forecast ManagerPage 4: Special EventsFinally, the tag Special Events in the opening window enters the area where it is possible to shape occasional events, such as promotional actions or any other event that may cause a temporary change in the sales level of the curve you analyse. In the upper right side of the page Special Events, when unchecked the option Measure special events deactivates all controls in the page.
You cannot analyse special events without selecting at least one forecast period when to apply the coefficient(s). The time period(s) accounting for special events can be chosen with the drop down list Forecast Period and the list Kind of event determines what sort of event you are forecasting. A promotional campaign of our direct competitor may produce a negative effect on our sales and vice versa.Labels in this list are read from the Special events range in page Input data. It is because of this list that the software does not allow to enter the page above without selecting an event range in page one. As shown below there are four available options in the listbox Smoothing method, Quadratic trend is set by default. The smoothing method instructs Forecast Manager on how to remove the effect of special events from input data.
The Linear trend method replaces in the fitted line values affected by special events with values obtained by means of a linear regression equation as the one below (see math notation in section Technicalities).
Where, the variable x refers to time and the beta coefficients are found with a linear regression. The Quadratic trend method applies the same concept as for the linear one, but, of course, the equation above is modified to include a third term for time squared. The Preceding value method replaces an affected value with the first unaffected value preceding it. The Average method replaces values affected by special events with a mathematical average of all terms preceding the value to be smoothed. The two buttons Remove>> delete the selected entry in the respective list. The listbox Coefficients five options that instruct the software on how to estimate the effect of future special events of the same nature as those listed in the Kind of event panel.
Average values is the default coefficient, and it estimates the value of special events averaging all special event effects belonging to the same event category. For instance, Forecast Manager estimated the following three percentage effects for an event category named Promo 15%: +12.3%, + 9.7% + 13.9%. In this case, 11.9% is the average value used for encapsulating in the forecast the value of the special event effects of kind Promo 15%. Last value uses as coefficient for one event category the %effect of the last special event found in the input data for the same event category. First value uses as coefficient for one event category the %effect of the first special event found in the input data for the same event category. Largest value uses as coefficient for one event category the largest %effect of all effects for the same event category. Finally, Smallest value uses as coefficient for one event category the smallest %effect of all effects for the same event category. Tip: Below is an example of input selection. Selecting range A1:A15 runs a single series forecast while selecting range A1:C15 runs a multiseries (batch) forecast.
