Demand Plan Options for Demand or Replenishment Plans

Define or change demand options for a demand plan or replenishment plan on the Demand tab on the Plan Options page.

For a replenishment plan, the Demand tab is enabled only when you select the Generate forecast checkbox on the Plan Options page.

The following table explains how you use the Demand tab:

Demand Plan Option	Explanation
Aggregate non-key customer data to an All Other level member	Select this checkbox to aggregate data for non-key customers to placeholder sites named <zone name>:All Other or a placeholder site named Default:All Other. The data for key customers is stored at the Customer Site level. For information about the functionality for aggregating data for non-key customers, see Overview of Aggregating Data for Non-Key Customers. This checkbox is available for only demand, demand and supply, and sales and operations plans. The checkbox is disabled for a demand plan if you've selected a planning level profile on the Scope tab on the Plan Options page. Caution: Select this checkbox only while creating your plan and before running it for the first time. If you change your selection later, you might run into data integrity issues.
Forecasting Time Level	Select the time level at which the forecast is generated. The selections are restricted to the planning time level that you selected on the Scope tab or higher levels. For example, for a planning time level of a week, the available forecasting time levels are week and period. For a demand plan that uses a hybrid time hierarchy, the available forecasting time levels are parent levels of the Hybrid time level for the hierarchy that you selected as the planning calendar on the Scope tab. The bucket you select for the forecasting time level is also used for the historical buckets, history offset, forecast buckets, and locked forecast buckets displayed on the Demand tab. Note: The forecasting time level doesn't apply to forecasting profiles that use feature-based machine learning.
Forecasting Profile, Input Measure, and Output Measure	In the Forecast Profiles section, specify the forecasting profiles for the plan. Only those forecasting profiles that are enabled for the Demand Management or Replenishment Planning work area are available for selection. When you run the plan, the forecasting profiles run in the specified order. If the forecasting profiles use the same output measure, it stores the output of the last run forecasting profile. You can't edit the input measure or output measure for a forecasting profile. The data in the input measure serves as the basis of historical demand for the forecast. The output measure stores the generated forecast. For information about predefined forecasting profiles, see Create a Forecasting Profile.
Analysis Set	Select an analysis set. If you don't select an analysis set, the forecasting profile applies to the full scope of the plan. If you select an analysis set, the forecasting profile is applied to only the subset of the plan scope that's defined by the analysis set. For example, you've an analysis set containing a specific set of products, and you select that analysis set for a forecasting profile. In this example, the forecast is generated for only the products included in both the analysis set and the plan, instead of for all the products in the plan. You can use segments as member filters while configuring the analysis sets. Then, you can apply forecasting profiles to specific segments by selecting the associated analysis sets.
Historical Buckets	Define the amount of historical data to use during the forecasting process. The number of historical buckets is multiplied by the forecasting time level for determining the time horizon (number of days between the history start and end dates) for considering the collected historical demand. The historical buckets substantially affect the demand forecast generated when you run the plan. Use at least 12 months of history, but using 18 to 36 months of history is the best practice. When setting the historical buckets, keep the following in mind: Very long history affects the plan runtime and makes the forecast less relevant to current demand patterns. Less than 1 year of history makes determining the impact of seasonal occurrences and holidays difficult or impossible. The historical bucket setting is the maximum amount of history used during forecast generation. The actual amount depends on the data available for a given item and organization or item-location combination. For each item and organization or item-location combination, the forecasting process identifies all available historical data, removes any leading zero demand before the first positive demand point, and generates a forecast.
History Offset	Define the buckets by which the history start date is moved backward. The forecast start date is then set to one bucket ahead of the new history end date, and the history start and end dates and forecast end date are automatically updated. The bucket size is determined by your selection in the Forecasting Time Level field. You use the history offset when you want to perform out-of-sample forecast tuning or back testing. By moving the history start and end dates backward, you can generate a forecast for the period between the history end date and the forecast start date, for which period you already have data. You can compare your forecast with the actual data to refine your statistical forecast.
History Start and End Dates	For a demand plan, the history end date is determined by the plan run date and the planning calendar. For a replenishment plan, the history end date is determined by the plan start date (displayed on the General subtab on the Supply tab on the Plan Options page) and the planning calendar. The history start date is determined from the history end date and the number of days between the dates. The number of days between the history start and end dates is determined by multiplying the number of historical buckets by the forecasting time level.
Forecast Buckets	Indicates the number of time periods for the demand forecast and is driven by the planning horizon set for the plan and the forecasting time level.
Locked Forecast Buckets	Indicate the number of buckets from the end of historical data that don't receive a new forecast when you run the plan. The bucket size is determined by your selection in the Forecasting Time Level field. Ensure that the number you enter doesn't exceed the displayed number of forecast buckets.
Forecast End Date	Indicates the end date for the generated forecast. The end date is set to an end-of-period date based on the history end date, forecasting buckets, and forecasting time level.