Forecast Error Calculation

Common use cases provide flexibility and visibility into forecast error or forecast accuracy measures. Inventory optimization and demand planning use cases are different. While inventory optimization use cases calculate forecast error as a primary input at the base planning level, demand planning use cases define error measures to monitor forecasting performance and track the forecast value-add by comparing different forecast key figures.

SAP IBP for inventory provides forecast error measurement functions via the Forecast Error Profiles - Inventory app, which was made available with SAP IBP 2405 and which replaces the deprecated Manage Forecast Error Calculation - Inventory Optimization app (not usable with 2411 and later).

The Forecast Error Profiles - Inventory app provides options to set up a forecast error calculation profile and to use it both from the Excel add-in and the Application Job app.

To use this app, your business role must contain the business catalog Profile for Forecast Error Calculations for Inventory (SAP_IBP_BC_INV_KPI_PROFILE_PC).

Usually, the forecast error-related process is based on the following steps:

As part of the first steps to optimize the levels of inventory within the supply chain, we need to calculate the forecast error. The following figure presents a dashboard to understand the results of the algorithm:

The slide presents four graphs in a Dashboard:

• Historical sales and historical lag 0 forecast (Eaches):

  This is a line plot where two key figures (IO Sales and IO Forecast Lag) are grouped by weeks.

• Historical Lag 0 forecast error coefficient of variation:

  This is a heat map where customer groups and segments are displayed by location and product.

• Historical sales and Historical Lag 1 Forecast (Eaches):

  This is a line plot where two key figures (IO Sales and IO Forecast Lag) are grouped by weeks.

• Historical Lag 1 forecast error coefficient of variation:

  This is a heat map where customer groups and segments are displayed by location and product.