Demand Aggregation and Forecast Disaggregation Feature
The Demand Aggregation and Forecast Disaggregation feature enhances the forecasting capabilities of the system by allowing the following:
• Aggregate Demand — For a part, this is the demand from multiple locations that is aggregated at a single location, also known as the Aggregation Location.
• Forecast — At the Aggregation Location, the Forecasting process generates the time-series forecast for the aggregated demand. The Best Fit functionality can also be used to find the best suitable forecast method for the aggregated demand.
• Disaggregate Forecast — The forecast from the Aggregation Location is distributed to all of the contributing child locations, using a system calculated average demand of x slices or the user provided weight.
Customers can use a blend of locally generated statistical forecast and disaggregated forecast to get an overall better forecast. Demand trend and seasonality could be better captured at an aggregate level than locally generated forecast. For slow moving or intermittent parts, aggregated demand based forecast can provide better forecast specially for statistical forecast methods.
The following changes were made to enable this new feature:
1. The Forecast Disaggregation Parameters page was added. You can view, add, or modify Forecast Disaggregation parameters on this page.
2. A new AutoPilot process, Demand Aggregation (Cannot be executed by Segment) was added.
3. A new Forecast Method, Forecast Disaggregation, was introduced. This Forecast Method will populate at all the child locations. At the source or parent location,(aggregated node or pooled), the true Forecast Method is used.
This is the recommended workflow for using the new feature:
1. Designate a location where all of the demand from multiple locations will be aggregated. This location can be a virtual location or a parent location. You can use an existing location or . create a new location.
2. Identify the parts that should use Demand Aggregation and Forecast Disaggregation.
3. Identify or create Demand and Forecast streams to be used in Demand Aggregation and Forecast Disaggregation
4. Create Forecast Disaggregation Parameters using Aggregation Locations, Demand Streams, Forecast Streams, and Segments that have SKUs of required locations, excluding the Aggregation Location, for each part.
5. Setup separate Stream Configurations for child and aggregation location. Child locations will have forecast method as Forecast Disaggregation for demand and forecast stream used in Demand Aggregation and Forecast Disaggregation. Aggregation locations will have a Statistical forecast method and can also use Best Fit functionality for demand and forecast stream used in Demand Aggregation and Forecast Disaggregation.
6. Run the following AutoPilot processes
◦ Synchronize Database
◦ Demand Aggregation (Cannot be executed by Segment)
◦ Forecasting (on the segment)