Forecast Model, Strategies and
Profiles
Use
In Service Parts Planning (SPP) you can use forecast profiles, forecast models, and forecast strategies.
A forecast model defines how the system proceeds when determining the forecast. For example, there are forecast models, where the system smooths historical data or calculates average values, or where trend and seasonal behaviour is considered.
A forecast strategy results from a forecast model and a calculation rule. The calculation rule defines which parameters the system calculates and which parameters are produced by it. For example, the system can calculate the Demand parameter, and thus also calculate, with first order exponential smoothing, the Average Demand per Order Item parameter. Then using these two forecast key figures it can determine the forecast key figure Number of Order Items.
A forecast profile contains a forecast strategy or a forecast model as well as all the parameters that the system needs to determine the forecast. For example: the amount of historical data that the system uses for creating forecasts, the type of forecast error that the system uses during automatic model selection and a profile for determining the final forecast.
Features
Forecast Models
For each location product, you can decide which forecast model the system uses when creating the forecast. You can either make this decision yourself, or the system decides during automatic model selection
The following forecast models are available in SPP:
● First-order exponential smoothing
● Second-order exponential smoothing
● Seasonal trend model with fixed period groupings
You can only use the Declining Demand Forecast forecast strategy if you have chosen one of the following periodicities in Customizing:
■ Month
■ Posting period with exactly 12 posting periods per year
Forecast Strategies
A forecast strategy consists of a forecast model and the forecast calculation rule. The forecast calculation rule defines which of the following roles are played by the parameters demand,number of order items, and the average demand per order item.
● The leading parameter is either the demand or the item. You can tell which it is by the name of the forecast strategy. For example, in the forecast strategy FOES (Item is a Dependent Parameter), the demand is the leading parameter.
● The constant parameter is always the average demand per order item. The system calculates this using first-order exponential smoothing.
● The dependent parameter is either the demand or the item, depending on which is the leading parameter. The system determines the forecast of the dependent parameter from the forecast of the leading parameter and the forecast of the constant parameter.
The following forecast strategies are available in SPP:
● A1 – FOES (Item Is a Dependent Parameter)
● A2 – FOES (Demand Is a Dependent Parameter)
● B2 – SOES (Item Is a Dependent Parameter)
● C1 – Moving Average (Item Is a Dependent Parameter)
● D1 – Linear Regression (Item Is a Dependent Parameter)
● E1 – Seasonal Trend Model (Item Is a Dependent Parameter)
● F1 – Seasonal Trend Model with Fixed Period Groupings (Item Is a Dependent Parameter)
● G1 – Intermittent Forecast Model (Item Is a Dependent Parameter)
● H1 – Dynamic Moving Average
● I1 – Declining Demand Forecast (Item Is a Dependent Parameter)
● J1 – Linear Regression (Item Ís a Dependent Parameter)
If you want to work with your own forecast strategy, you can implement the Business Add-In (BAdI) BAdI: Define Customer-Defined Forecast Strategy (/SAPAPO/BADI_FCST_STRATEGY) and define your own forecast strategy here that creates a forecast for demand. The system calculates the Average Demand per Order Item parameter using first-order exponential smoothing, and then using these two forecast key figures it can determine the forecast key figure Number of Order Items. For more information, see the Implementation Guide (IMG) for Advanced Planning and Optimization under Supply Chain Planning → Service Parts Planning (SPP) → Business Add-Ins for Service Parts Planning (SPP) →BAdIs for Forecasting → BadI: Determine Customer-Defined Forecast Strategy.
Forecast Profiles
In the forecast profile you define a forecast strategy or a forecast model and the parameters that the system needs to determine the forecast for each strategy. You assign a forecast profile to each location product for which the system is to carry out a forecast.
If you have not assigned a forecast profile to a location product, the system passes on a forecast profile to this location product. For more information, see Forecast Profile Inheritance.
You can also copy forecast profiles. For example, you can define a forecast profile for a product, and copy this forecast profile to all location products belonging to this product. For more information, see Copying Forecast Profiles.
If you work with leading-indicator-based forecasts, you must assign forecast profiles, not only to location products, but also to the leading indicators and coefficients. The forecast profiles for leading indicators and coefficients do not contain any forecast strategies, but instead they contain forecast models, since the system creates the leading-indicator-based forecast only for the Demand parameter. For more information, see Running a Leading-Indicator-Based Forecast .
You can define a forecast profile from the SAP Easy Access screen under Advanced Planning and Optimization → Service Parts Planning → Planning → Forecasting → Forecast Profile.