The forecast calculation process results in updated forecast being posted in the “Calculated”row of the forecasting grid. It is launched in one of several ways and can be influenced by a number of used controlled mechanisms. Forecasts can be generated for all or a specific group of items in a single run, or for one item at a time interactively. In addition,there are several options available by which the normal forecast process can be overridden. Some of these are detailed in the rest are discussed in this section.
The most common procedure for forecasting is to use the “Forecast All SKUs” option in the File menu. Selecting this opens the “Forecast All SKUs” window.
It is possible to calculate forecasts for all items in the database in one run using the same forecasting parameters for each. This is done by leaving all of the selection boxes on the left side of the window blank and checking the “Group by” boxes for SKU and Location.
If different parameters are to be used when forecasting different groups of items, a separate forecasting run must be executed foreach. To select a group of items, entervalue(s) in one or more of the selection boxes on the left of the window,select one or more of the Active Codes in the center of the window, and/or a selection of ABC codes on the right, in the same way as items are filtered on the Welcome Screen (see CP Sec 3.5.1). As above, the “Group by” boxes for SKU and Location should both be checked.
The “Group by” boxes to the right of the selection fields are provided to permit forecasts to be calculated by accumulating the history of all of the items in the group, calculating the aggregate forecast for the group,and then prorating the forecasts back down to the individual items. For example, if only Category 1 is checked,history totals will be calculated for each unique entity in Category 1, and the forecasts will be calculated for each entity based on the totals. Then those forecasts will be distributed back down to the items within each entity in proportion to their contribution to the total. If any of the individual items do not already have a forecast, they will not be allocated a portion of the total forecast. Ensure that all items have been forecast before using this feature.
To use this feature, the box for SKU would normally be left unchecked, except in the case where the forecasts are to be calculated at the SKU level and distributed back down to the various Locations.
There are 3 options which can be selected to affect the forecast:
· FSIM Personality Select the personality desired for the current run. See Section 2 for more details on Avercast’s Personalities.
· ReplaceAll Adjusted Forecasts or Replace Non-overrides Only These options cannot both be selected together. When using All Adjusted Forecasts, themonthly Adjusted forecast values are reset to agree with the new Calculated forecasts. Selecting Replace Non-overrides Only will reset the Adjusted forecast values in those months where the Calculated and Adjusted forecasts were equal prior to there-forecasting, and will not change those which were not equal. This allows the preservation of previous adjusted forecasts through the re-forecasting process. If both options are unchecked, this means that only the calculated forecast will be updated without altering user adjusted forecast.
· Use Grand Rapids with Slow Moving Items This option will use a special forecasting methodology (called Grand Rapids) to calculate forecasts for items coded as slow moving(i.e., Active Code = S.
There are three forecast types associated with monthly forecasting.
· Monthly Point of Sale
· Customer Forecast
Standard is the most commonly used forecast type. Forecasts will be calculated based on the demand history, using the forecasting parameters as set in Forecast Policies (see Section 4.3). When this type is selected, there are four additional options available:
· Override Forecast Assignments – if selected, any forecast assignments which have been madill be ignored
· Ignore All Default Curves - see Section 5.5.
· Skip First Non-Zero Sale for Curves
· Use Curve
Monthly Point of Sale forecast type uses POS history as the basis for calculating forecasts. The storage of POS history is optional but must be implemented to take advantage of this forecast type. Refer to Standard Forecast type (above) for details on the additional options available when using this type.
Customer Forecastuses customer history as the basis for calculating forecasts. The storage of customer history is optional but must be implemented to take advantage of this forecast type. Refer to Standard Forecast type (above) for details on the additional options available when using this type. The option to replace weekly adjustment only applies when weekly forecasting has been implemented.
It is possible to forecast an individual item using either the Flat Forecast icon or the Standard Forecast icon. Before initiating the calculation, on the right side of the screen select the FSIM Personality that you wish to use for this forecast.
Using the Flat Forecast icon will distribute the current 12 month total of the Calculated forecasts equally over the months. Any residual amount due to rounding is allocated by adding 1 to the months beginning with the current month. If the existing Calculated forecast is allzeros, this function will have no effect. This action does not alter the values in the Adjusted forecast row, but after the recalculation of the Calculated forecast, they will likely be different and therefore will be displayed in red. To copy the new Calculated forecast into the Adjusted row, use the “Move Calculated to Adjusted’ option on the Edit menu. Note that the previous Calculated forecast is permanently overwritten and can only be restored by running Forecast All SKUs from the File menu for that item.
When executed using the Standard Forecast icon the results of the re-forecast analysis are displayed in ascending sequence by MSE (mean square error) in a pop-up window which includes the method that was selected in the previous forecast run. Clicking on one of the methods in the list displays the resulting forecast for that method in the Adjusted row; the Calculated row remains unchanged. In this way, it is possible to compare the results from the two calculated forecasts. To return the Adjusted row to its previous values, click on the “Previous Forecast” method or use the “Move Calculated to Adjusted’ option on the Edit menu.