Demand Sensing

Demand Sensing is used to compare the current forecast to the actual sales for measuring the current period actuals against the forecasted plan. Deviations can exist in the actuals from the planned forecast, therefore, safety stock must be maintained. However, when the forecast deviates more than the permissible limit, corrective action must be taken before updating the forecast at the end of the current planning cycle.

Demand Sensing enables you to identify the forecasts where errors exceed the specific tolerance percentage or multiple standard deviations. These tolerances can be modified based on user requirements.

This table lists the levels to track the errors between actuals and forecasts:


Level	Item/Location dimension level	Default Hierarchy level - Item	Default Hierarchy level - Location
Base	Base level of dimensions	Item_id @ Location_id	Customer_id
Item	Base level of Item dimension Lconst of Location dimension	Item_id @ Location_id	Top Node (All Customers)
Item Grp	Group level of Item dimension Lconst of Location dimension	Item_id	Top Node (All Customers)
Item Grp - Location	Group level of Item dimension Base level of Location dimension	Item_id	Customer_id

Note: The levels are based on the base elements in the hierarchies and the Group level in the item hierarchy.

You can run demand sensing from the following levels using the default hierarchies:

Item, warehouse, and customer combination (Base)
Item and warehouse combination (ITEM)
Item (Item Grp)
Item and customer combination (Item Grp - Location)

Note:

You can analyse the forecast versus sales to date using the different workbenches for different selections.
The Percentage Tolerance exceptions are created for all levels. The Standard Deviation Tolerance exceptions are only created for the Base level.