Model Events
Event modeling is a type of forecasting that accounts for irregular events that are not included in historical data, such as promotions and price changes. Events may be planned or unplanned:
- Planned events, such as promotions and plant closings, should be included in forecasts.
- Unplanned events, such as natural disasters, should be isolated from forecasts to avoid distorting projections.
CCH Tagetik Supply Chain Planning provides a unique method for modeling events. You can use each record in CCH Tagetik Supply Chain Planning as an event in another record. For example, you can include a Promotion record that tracks the seasonal pattern for a promotion as an event in the product's sales forecast. When a forecast uses another record as an event, that event can be applied to the forecast in several ways, including regression.
See About Record Dependencies.
To add an event to a forecast
- Open the record you want to edit.
See Open and View Records.
2. In the Adjustments and Overrides section, in the Overrides row of the data table, click the Add
Adjustment icon (
).
3. In the Adjustment screen, from the Adjustment Type list, select one of the following:
Adjustment Types
| Adjustment Type | Description |
|---|---|
| Local | Apply adjustments entered locally using the same syntax supported in the override line. This option creates an Override row in the Adjustments and Overrides table. Values typed in this row are treated as increments. To replace a forecast value with another value, always use the Overrides row. Ft = St + At |
| Add | Add the reference series to the Starting Value. Ft = St + Rt |
| Subtract | Subtract the reference series from the Starting Value. Ft = St - Rt |
| Multiply | Multiply the reference series and the Starting Value, and add the product to the Starting Value. Ft = St + St * Rt |
| Adjust | Scale the Starting Value to match the growth in the reference series. Ft = St * Rt/Rnow When one record (source record) is applied to another record (adjusted record) as an Adjust adjustment, the effect of the source record is removed from the history of the adjusted record (this adjusted history is displayed as a green line), the adjusted record forecasts based on this adjusted history, and the source record is then applied as an adjustment to the adjusted record. This removal from history is necessary to prevent double counting the effect of the source record. Mathematically, the Adjust adjustment calculates the average of all history in the source record and adjusts each historical data point in the adjusted record based on how the actual in the source record compares to the average of the history in the source record. This calculation for the adjusted history in the adjusted record is: (source record historical average / historical data point in source record) * corresponding historical data point in the adjusted record. To adjust the forecast, the calculation is the inverse so if the forecasted data point in the source record is higher than its historical average, a positive adjustment is applied to the adjusted record. Otherwise, a negative adjustment is applied. |
| Regression | Perform a regression analysis comparing the period-to-period change (first difference) in the reference series to the period-to-period change in this item, and adjust accordingly. Ft = St * Kreg * (Rt-lag - Rnow) |
| Supersession | See Add a Supersession Adjustment. |
| Comparison | Insert a line showing the forecast from another record that can be used as a comparison or as a period-by-period override. |
| Equate | Spread data over the course of a year (linked to a source record). |
| Supplementary Data | Insert a line to include any additional information. |
| Expiration | Import expiration data for inventory that expires such as groceries, pharmaceuticals, and batteries. This helps you see when inventory is going to expires so you can plan inventory and eliminate waste. See Add Expiration to Records. |
| Group Comparison | Add data from another group of records to use as a comparison or as a period-by-period override. |
| Firm Orders | Firm order adjustments can be added to point to a data source that includes orders on specific days in the future. These can be used for either comparisons, consumption, or merging into the forecast. See Add Firm Orders to Records. |
| Order Rollup | Used during an import of the supplier file to create a hub for your orders. See Order Rollups. |
| Addressable Market | Forecast using a percent penetration into an addressable market rather than forecasting based on absolute values. Type the name of an item that contains the addressable market forecast. |
| 4. If necessary, type the Name of the record to use as the adjustment and click OK. |
The record is added to the forecast table as an adjustment. The values of the record are included in a table row along with the record name.
Note: Override values are case-sensitive.
Next to the record name is a symbol indicating how the record is applied:
Adjustment Symbols
| Symbol | Description |
|---|---|
| + | Add |
| - | Subtract |
| x | Multiply |
| A | Adjust |
| R | Regression |
| T1 | Preliminary Top-down |
| T2 | Final Top-down |
| 5. To make changes to the record, next to the record name, click the Edit icon. |
If you do not see the Edit icon next to the record name, then the adjustment
record is a group adjustment applied to all Time Series Forecast records. Contact the owner of the adjustment record if you have questions about the
group adjustment.
6. To remove an adjustment, at the end of the adjustment's row, click the Remove Adjustment icon (
).
The entire row is removed. 7. Save your changes.