Process

These processing options let you specify whether the system runs the program in proof or final mode; creates weekly or monthly forecasts; and specifies the start date, length, and data that are used to create forecasts.

In addition, you use these processing options to specify how the system calculates the best fit forecast. The system applies the selected forecasting methods to past sales order history and compares the forecast simulation to the actual history. When you generate a forecast, the system compares actual sales order histories to forecasts for the months or weeks that you indicate in the Forecast Length processing option, and computes how accurately each of the selected forecasting methods predict sales. Then the system identifies the most accurate forecast as the best fit. The system uses two measurements for forecasts: Mean Absolute Deviation (MAD) and Percent of Accuracy (POA).

MAD is the mean of the absolute values of the deviations between actual and forecast data. MAD is a measure of the average magnitude of errors to expect, given a forecasting method and data history. Because absolute values are used in the calculation, positive errors do not cancel out negative errors. When you compare several forecasting methods, the forecast with the smallest MAD is the most reliable for that product for that holdout period.

POA is a measure of forecast bias. When forecasts are consistently too high, inventories accumulate and inventory costs rise. When forecasts are consistently too low, inventories are consumed and customer service declines. A forecast that is ten units too low, then eight units too high, and then two units too high, is an unbiased forecast. The positive error of ten is canceled by the negative errors of eight and two.

1. Mode

Specify whether the system runs the summary forecast in proof or final mode. When you run this program in proof mode, the system does not create any forecast records which enables you to run it again with different criteria until you produce appropriate forecast information.

When you run this program in final mode, the system creates forecast records. Values are:

Blank: Proof mode.

1: Final mode.

2. Weekly Forecasts

Specify monthly or weekly forecasts. For weekly forecasts, use fiscal date patterns with 52 periods. For monthly forecasts, use fiscal date patterns with 14 periods. Values are:

Blank: Monthly forecasts.

1: Weekly forecasts.

3. Start Date

Specify the date on which the system starts the forecast. Enter a date to use or select one from the Calendar. If you leave this processing option blank, the system uses the system date.

4. Forecast Length

Specify the number of periods to forecast. You must have previously established fiscal date patterns for the forecasted periods. If you leave this processing option blank, the system uses 3.

5. Actual Data

Specify the number of periods of actual data that the system uses to calculate the best fit forecast. If you leave this processing option blank, the system uses 3 periods.

The system applies the selected forecasting methods to past sales order history and compares the forecast simulation to the actual history. When you generate a forecast, the system compares actual sales order histories to forecasts for the months or weeks that you indicate in the Forecast Length processing option and computes how accurately each of the selected forecasting methods would have predicted sales. Then, the system identifies the most accurate forecast as the best fit.

6. Mean Absolute Deviation

Specify whether the system uses the Mean Absolute Deviation formula or the Percent of Accuracy formula to calculate the best fit forecast. Values are:

Blank: Percent of Accuracy formula.

1: Absolute Deviation formula.

7. Amounts or Quantities

Specify whether the system calculates the best fit forecast using quantities or amounts. If you specify to use amounts, you must also extract sales history using amounts. This processing option also affects forecast pricing. Values are:

Blank: Quantities.

1: Amounts.

8. Fiscal Date Pattern

Specify the fiscal date pattern type to use for the forecast calculations. If you run weekly forecasts, the fiscal date pattern that you specify here must be set up for 52 periods.

9. Negative Values

Specify whether the system displays negative values. Values are:

Blank: Substitutes a zero value for all negative values.

1: Displays all negative values.