SAP SOP Forecasting

Study Material Contributed by Ulhas Kavle - Senior SAP Consultant


The use of SOP in SAP is mostly to use the historical data stored in the information structures and use it for forecasting the values of a period in future. Though one may rightly argue that they would not use the forecasting feature and instead carry out the forecast outside SAP using their own forecasting methods and tools and pass the same to SAP. Forecasting models and tools are available in SOP but these have to be studied along with the sales patterns displayed and the use of the best model would give you the best results.

Forecasting as the word suggests (in material planning) is forecasting of the future requirements using the historical data. The historical data can be the material’s historical consumption data as recorded by standard SAP on a real time basis every time there is material movement (goods issues predominantly) or it can be the historical sales data as recorded, on a real time basis, in absolute values, in the sales information systems. One of the primary uses of Sales and Operational planning is its capability to forecast the historical data and use the forecasted requirements as production figures in the form of planned independent requirements.

In consistent planning, one can forecast any key figure based on the key figures values in the information structure, provided the key figure has been marked for forecasting in the “set parameters for information structure and key figures” configuration, whereas in standard SOP the forecasting is carried on sales quantities (only on sales key figure) of product groups or materials using the historical consumption data (goods issue) as recorded in SAP consumption history for the material; this demonstrates that standard SOP forecasts at material level as the case is with level by level planning type. It should be noted that standard SOP plans based on level by level planning and not based on consistent planning type.

Fig – SOP_60

Companies with no existing Historical data:

Companies, who have just started on SAP or have just started using SAP SOP (Sales & Operational Planning), would not have any historical consumption data or historical key figure data with them; for such cases, you may need to enter the historical data manually or upload the same through batch inputs.

Forecasting Models and Strategies

The aim of this document is not to explain forecasting in detail; though with the flow of the topic, the basic models and strategies need to be stressed on.

When you are using the smoothing factors, the system puts stress on the most recent historical values rather than the less recent ones. The recent values in the time series are given more weight than the less recent ones, i.e., the weights for the recent ones exceed the weights for the less recent ones (the weights decrease at an exponential rate as they move to the less recent values). The greater the value of the smoothing constant the more weight is given to the recent values in the historical time series. The use of the smoothing factors means that you would have to go through a trial and error method to arrive at this value.

When the historical values are constant and line across the mean value, you can use the “first order exponential smoothing”, but if the values waiver away from the mean and show a consistent trend (upward or downward), then use of first order exponential smoothing will have the forecast lag behind in an amount equal to the trend. Therefore when there is a “consistent trend” the second order exponential smoothing recognizes the consistent trend in the historical data (it evaluates the smoothed estimate of the trend in the time series data).

Therefore, it can said that if you are using a first order exponential smoothing forecast for a trend, then you are bound to get a wrong result, instead it is recommended to use the second order smoothing method. The second order smoothing model can be chosen with or without model parameter optimization.

Along with the forecast model, there are forecast tools such as the “outlier”. The outlier as the name suggests recognizes any value which is does not fit in the constant mean graph or consistent trend or in seasonal graph or in seasonal trend graph and such out-casted values are corrected by the system. The outlying values, cause considerable damages to the resulting forecast therefore to correct these outlying values, the system forecasts in the past period and then compares with the outlier value and if the difference (difference of historical value and the forecasted historical value) exceeds the specific value, the system replaces the original historical value with the ex-forecasted value for the periods observed on the time series. Once the correction is done, the forecasting process resumes with changed historical values. The sigma factor controls such outlier values and the greater the sigma value the more the system is resistant towards such outlying values.

Another forecast tool is setting any negative key figure value to zero. Negative values may cause the resulting forecasts to be damaged. Another tool is “ignoring initial zeros” in the historical time series.

Fig – SOP_61

The forecasting types/models are categorized as “constant model” where the historical data is pretty constant across and stays very much at a mean value.

· Forecast with Constant Models (10)

· Forecast with First Order Exponential Smoothing (11) - Alpha Smoothing value is required

· Forecast with Automatic Alpha Smoothing factor Adoption (First Order Exponential Smoothing) (12)

· Forecast with Moving Average Method (13) - Number of Historical Values is Mandatory

· Forecast with Weighted Moving Average Method (14) - Weighting group is Mandatory

The “trend model” is where; there is an evident trend in the consumption of the material (evident fall or rise).

· Forecast with First Order Exponential Smoothing (21) - Alpha and Beta Smoothing value is required

· Forecast with Second Order Exponential Smoothing with or without model parameter optimization (22) ) - Alpha and Beta Smoothing value is required

· Forecast with Automatic Alpha Smoothing factor Adoption (Second Order Exponential Smoothing with or without model parameter optimization) (23)

The “seasonal model” is where; the consumption pattern changes seasonally, where fall and rise is evident and seasonal.

· Forecast with Seasonal trend based on winters model (31) - Alpha and Gamma Factors per Season is Mandatory

The “seasonal trend” model is where; the consumption is seasonal and there is an apparent trend observed in the figures. Therefore this model is a combination of seasonal and trend.

· Forecast with Seasonal Trend Models using First Order exponential Trend Smoothing (40) - Alpha, Beta and Gamma Factors per Season is Mandatory

The “Automatic” model is where; the system applies its own calculates the optimum smoothing factor values and applies the same on the historical data and run the historical data through forecasting formulas. SAP decides the required model automatically depending on the pattern of historical data.

· Forecast with Automatic Model Selection (50)

· Forecast with Automatic Model Selection Process 2 (52) - More Precise but time consuming

The “historical data way” is where; the historical data is identically pasted in the future period as the forecast values. This evidently is not a forecasting type or model but is available for use in Sales & Operational Planning.

· Historical Data Adopted (60)

Creating a Forecast Profile:

Forecasting can be carried out in the planning table by directly configuring the forecast profile in the planning table or by using a using an existing pre-configured forecast profile. The forecast profile can be configured separately using the transaction code MC96 (The screen shot of this configuration transaction is given below).

The forecasting profile contains the following parameters:

a) The period or interval for the historical data, that needs to be used during forecasting.

b) The period or interval of the required future periods for which the forecasting should be triggered for.

c) The forecasting strategies/models that should be used to carry out the forecast.

d) The factory calendars to specify the holidays or the fiscal year variant.

e) The number of periods in a season, in case there is seasonal trend.

f) Maximum number of historical periods.

g) Indicator to base the forecast of historical consumption as in standard SOP’s level by level planning forecast based on material consumption history.

h) Correcting historical values

i) Alpha factor to smoothen the basic values, Beta factor to smoothen the trend values and Gamma factor to smoothen the seasonal index.

Fig – SOP_62

Carrying out Forecasting in the Planning table

Forecasting can be carried out in the planning table for a key figure on which the cursor is kept. After keeping the cursor on the key figure, you should use the path – Edit > Forecast, to forecast the key figure values in a future period. A Forecast Model selection pop-up appears. Here you can enter the following:

Fig – SOP_63

a) Choose a forecast period or the actual forecast interval

Choose a forecast period or the actual forecast interval in terms of month/year and also choose the historical data period or the historical data interval in terms of month/year. You cannot enter both period and interval.

b) Chose one of the forecasting models

Chose one of the forecasting models – a constant model, trend model, seasonal model or just have the system paste the historical figures in the future periods (an exact copy).

Fig – SOP_64

c) Choose a forecast profile

The forecast profile should be selected in the pop-up, which will pull in the forecast parameters. SAP uses the standard profile called “SAP”. It is possible to change the “forecast profile code” itself or even change some of the configured parameters of the profile. To select the forecast profile, the planner should press the button – “Forecast profile” on the “Forecast: Model Selection” popup as shown on the screen below. To change the parameters of the selected profile, the planner has to press the “pencil button” on the “forecast profile selector” popup (the second popup on the screen).

Fig – SOP_65

Fig – SOP_66

d) Adding Historical values for the key figures

You can also add the historical values for the key figures, if there are none. This can be a case when SOP or in fact even SAP is being just implemented (first time use); this can also be a case when you don’t want to use the actual historical patterns of the previous period set, which may have been misleading (for example because of environmental, political or natural factors) and instead prefer on using the crafted historical values. For this you should press the “Historical” button on the “Forecast: Model Selection” pop-up, as shown below and enter the historical values. Choose the indicator that specifies correction of the historical values.

Fig – SOP_67

e) Execute the actual forecasting process

Finally with all the forecasting parameters set properly on the “Forecast: Model Selection” popup, you can execute the actual forecasting process by pressing the “Execute Forecasting” or F7 Button on the popup. Once you press the execute forecast button, the system may ask for certain details depending upon the forecast model used, if not it will directly provide you with the output on a “forecast result” screen.

On the “forecast result” screen (popup) if you “press enter”, the results will flow in to the planning table for the key figures’ forecast period. The forecasted values can be corrected on the ‘forecast result’ screen and then passed on, in to the planning table for the key figure. Alternatively the planner can also change the forecasted key figure values in the planning table.

Fig – SOP_68

Fig – SOP_69

All the site contents are Copyright © and the content authors. All rights reserved.

All product names are trademarks of their respective companies. The site is in no way affiliated with SAP AG. Every effort is made to ensure the content integrity.

Information used on this site is at your own risk.

The content on this site may not be reproduced or redistributed without the express written permission of or the content authors.

Sitemap Sitemap