What is Available in DP?

We have pointed in a number of posts that the forecasting methodologies in SAP SCM DP are not particularly a differentiator in the marketplace, and that the majority of complexity is in the data modeling tool called the Business Warehouse Workbench (which appears to be in transition to be replaced by Business Objects at some point).

____________________________________________________

The Problem With DP Forecast Methodologies

In fact the forecasting methodologies in DP are both generic and substantially behind the competition. This fact is unknown to executives who come to believe that DP is a leading edge forecasting engine primarily because its expensive and developed by SAP. One example of DP’s old design is the fact that it can not optimize forecast parameter values, a factor which is now standard among DP’s competitors. To learn more about these issues goto. http://www.scmfocus.com/forecastsystems/

These reasons and more are why I typically recommend a blended forecasting solution for clients, as it is very rare that DP can meet all of the client requirements and pull off a successful implementation without some help from a best of breed vendor.
____________________________________________________

The Forecast Profile

The forecast profiles the combination of settings that can be saved and applied to the demand history. The profiles configuration is segmented into two the forecasting categories.

• Univariate Forecasting
• Multiple Linear Regression Forecasting

The forecasting profile configuration also allows has a third tab for composite forecasting which is a combination of the two forecasting categories.

Master Profile

The profile begins with the Master Profile tab. This tab holds the forecasting method independent fields which selects things like the forecast horizon.

Univariate Tab

This tab is used to create profiles for univariate forecasts. The Alpha, Beta and other factors can be controlled here. By saving different profiles with different values different univariate profiles are created, and can the be applied flexibly to products. Additionally, there are a number of standard univariate profiles that ship with SAP.

These can connected to different forecast strategies. Many of these should be familiar to experienced forecasters.

The MLR Tab

This profile allows any regression model to be created and then applied. Independent variables are added below.

The Composite Tab

This is simply a tab that combines two profiles (one univariate and one MLR) and applies a weighing factor to each profile to blend the forecast.

Conclusion

Forecast profiles in DP are quite flexible however, this is all very standard. As far as statistical forecasting, best fit and MLR (which is very infrequently used) DP has it covered, but there is nothing here that is not in just about every other forecasting package. Configuration is relatively straightforward. However, the trick is figuring out which is the right forecasting method for which products. At one time it was thought that best fit could always be used to perform the right selection. The extremely simply forecasting methods that are generally used out there indicates that this is not true. Several clients I have worked with enabled, and then disabled best fit forecasting in SAP DP. Secondly, DP is not really designed for the first analysis of data, or the initial selection. For this I recommend external tools which you can build yourself in Excel, or which you can purchase.