CASE STUDY

Demand Planning (DP)

In this case study the full demand planning process will be explained and
organized: from the configuration of the demand planning, choosing of the
adequate forecasting methods and the execution of the demand planning
till its release for the supply network planning.

Product MOTIVATION PREREQUISITES

SAP SCM Demand planning allows Before you start working with the
G.B.I forecasting customer requirements case study, you should create your
Release 7.0 EHP 3 by means of different historical and own planning version.
statistical forecasting methods.
Level NOTES
The demand planning for the
Advanced This case study uses the model
locations DL10 and for products
company Global Bike Inc. (G.B.I.)
DXTR1000 and DXTR2000 will be
Focus that was developed exclusively for
performed within this case study.
SAP UA Curricula.
Demand planning in the supply
chain

Authors
Hristina Ivanova

Version
2.31

Last Update
March 2017

© SAP SE
 CASE STUDY

Process Overview

Learning Objective Understand and perform demand planning. Time 95 min

During demand planning the historical data (such as billing quantity or volume of sales) gets
adjusted for the future by means of the forecasts. The result of the forecast is the demand plan.
The plan can be released in order to generate demand for the products in the respective locations
and periods (planned independent requirements). The calculated demand can be satisfied by
means of internal/external production. Small alterations in demand planning lead to large changes
in production planning. The goal of the demand planning is to provide the most correct forecasts
for the sales volume.
The graphic below displays the complete process and at which point we are currently

© SAP SE Page 2
 CASE STUDY

Table of Contents
Process Overview ....................................................................................................................... 2
Step 1: Administration of Demand Planning............................................................................ 4
Step 2: Executing the Demand Planning ................................................................................ 15
Step 3: Other Univariate Forecast Models.............................................................................. 27
Step 4: Multiple Linear Regressions (MLR) .......................................................................... 38
Step 5: Composite Forecast ..................................................................................................... 45
Step 6: Release the Demand Planning .................................................................................... 50

© SAP SE Page 3
 CASE STUDY

Step 1: Administration of Demand Planning

Task Create one planning area, one planning book and one forecast profile. Time 15 min

Before we start making the demand planning we have to create one planning
area, one planning book, and one forecast profile.
At first we should create our planning area.
A planning area is a central data structure that is used as a storage for
planning data for demand and supply network planning. Characteristics and
key figures with their functions are determined here.
A key figure is a numerical figure that denotes either an amount or a value.
Characteristics are the objects to which operational data is aggregated,
disaggregated, and evaluated.
A planning area summarizes main parameters for the coverage of the
planning. In this way it determines where and how the planning results will
be saved. The demand planning plans only in time series that are saved in the
version of the planning areas. Before you can work with a planning area you
have to create time series objects. This process is also named initialization of
the planning area. The system creates a network that consists of
characteristics and key figures in the liveCache.

To create the time series objects, please follow the menu path:
Advanced Planning and Optimization ► Demand Planning ► Menu path
Environment ► Administration of Demand Planning and Supply
Network Planning

Execute a right click on the planning area 9ADP01 and choose 9ADP01
from the list.

© SAP SE Page 4
 CASE STUDY

GBI-##
As a planning version type in GBI-##.
One year ago
The Start Date should be one year ago and the End Date should be one year One year in the future
in the future.
Click on .

If the system offers to adjust the date, click on :

Click on again in the popup Create Time Sries Objects.

You get the notification message:

Click on to get back to the main menu.

Now we create the planning book.

The demand planning process is usually conducted in the background.
However, the involved into the process users need the tool that enables them
to perform interactive corrections of the automatic forecast and to exchange
the data with other users. The most important tool for creating a demand plan
is the planning book. A planning book is based on the characteristics and key
figures of the planning area. The user does not maintain the planning area
but the data saved there.

© SAP SE Page 5
 CASE STUDY

The key figures define the master data, such as products, locations, etc. and
the characteristics define the transaction data, such as forecasts, customer
orders, etc.

To define a planning book, please follow the menu path:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Environment ► Current Settings ► Define Planning Book

GBI_DP##
Type in GBI_DP## for planning book and click on .

Planning Book
In the next screen enter for Planning book text Planning Book GBI_DP## GBI_DP##
9ADP01
and for Planning Area 9ADP01.
Univariate forecast
Moreover enable the checkboxes for Univariate forecast, Multiple linear Multiple linear
regression, and Composite forecast. The forecasting procedures will be regression
Composite forecast
explained later.
Click on in the lower right corner to get to the next step.

© SAP SE Page 6
 CASE STUDY

Total Forecast
Click on the Total Forecast and then click on to add the selected key
figures. The key figure appears on the left side and belongs now to your
planning book.
Click on in the lower right corner to get to the next step.

© SAP SE Page 7
 CASE STUDY

Mark the corresponding APO Location and APO Product and then click on
to add the selected key figures.

Click on in the lower right corner to get to the next step.

© SAP SE Page 8
 CASE STUDY

GBI_DP##
The data view should be named GBI_DP## and described as Data View Data View GBI_DP##
9A52WEEKS
GBI_DP##. Assign to both TB Profile IDs 9A52WEEKS (52 WEEKS) by
means of input assistance. Then press Enter.

Visible
Now you can mark the fields Visible and Input and assign both to the first Input
The first entry you can
entry you can select via the selection feature. In the Table 1 type in select
Demand Planning GBI_DP##. Demand Planning
GBI_DP##
Click on in the lower right corner to get to the next step.

© SAP SE Page 9
 CASE STUDY

Total Forecast
Mark the key figure Total Forecast then click on to add the selected key
figures.
Then click on .

© SAP SE Page 10
 CASE STUDY

Answer if you are asked whether the planning book and the data
view should be completed. Then click on to go back to the main menu.

Now we will create a new forecast profile.

At the beginning of the implementation of the demand planning your
company has to create the profiles that include the parameters that are
essential for the controlling of the forecast in the demand planning. How
many profiles should be created depends on the aggregation level and the
number of the models used for one particular demand forecast.

To create a new forecast profile, please follow the menu path:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Environment ► Maintain Forecast Profile

Please, enter the following values for the Master Profile:

Planning Area: 9ADP01 9ADP01

Master Prfl.: GBI_DP## GBI_DP##

Master Profile GBI_DP##

Description: Master Profile GBI_DP##
9ATOTFC
Forecast Key Figure: 9ATOTFC (Forecast complete)
W
Period Indicator: W (week)
Forecast Horizon From: next Monday next Monday

Periods: 52 52

History Horizon To: Sunday before the forecast start Sunday before the
forecast start
Periods: 52 52

All in all, the past and the forecast we use cover one year.

Enable the checkbox for Univariate Forecast and enter as a profile Univariate Forecast
GBI_DP##_UNI
GBI_DP##_UNI.
Choose the tab Univariate Profile.

© SAP SE Page 11
 CASE STUDY

Enter the following data:

Text: Univariate Forecast GBI_DP## Univariate Forecast
GBI_DP##
Key Figure: 9ATOTFC (forecast complete) 9ATOTFC

GBI-##
Version: GBI-##
31 (Season according to
Forecast strategy: 31 (season according to Winters) Winters)

Alpha, Beta, Gamma: 0,5

0,5
Periods: 13
13
Outlier Correction : Ex-Post Method (scroll down)
Ex-Post Method
This characteristic is used when the outlier in the past data that is important
for the forecast should be corrected.
The smoothing factors Alpha, Beta and Gamma mean:
Alpha base value
Beta tendency value
Gamma season component for trend-seasonal smoothing

© SAP SE Page 12
 CASE STUDY

With the help of the forecast strategy you set the forecasting procedures that
31 (Seasonal Winters´)
need to be applied. Here choose 31 Seasonal Model Based on Winters'
Method. A forecast procedure should be applied which considers seasonal
fluctuations in the market development.

Click on and then click on to adopt. In the tab change

again to the Master Profile and click on and then on to get back to the
main menu.

© SAP SE Page 13
 CASE STUDY

© SAP SE Page 14
 CASE STUDY

Step 2: Executing the Demand Planning

Task Execute a demand planning Time 15 min

The persons who take part in the demand planning process need a special
tool in order to be able to analyze, create, and edit the marketing program.
With the help of the planning books and data views the users are able to
interact with the SAP SCM. The planning table is a perfect tool for these
purposes. It consists of two parts.
On the left there are 4 sections:
• Selection window – for choosing the objects (locations, products,
resources, etc.)
• Selection profile – for saving the chosen objects into your profile
• Planning book – for choosing the planning book and data views
• Macros – we can create a macro for every planning that can, for
example, generate alerts.

© SAP SE Page 15
 CASE STUDY

On the right we have a planning book. The planning book is used as an

interactive planning instrument in the APO. A planning book can be attached
only to one planning area. In one planning area the data from the liveCache
(database) is saved in the structure or table defined by the user.

A planning area can have one or several planning books. Every book stands
for one particular role in the demand planning process. Every role can
require other information or a particular display format of the information. It
can also happen that particular roles are not allowed to access some features.

© SAP SE Page 16
 CASE STUDY

Every planning book can contain an unlimited amount of data views. Data
views determine the layout of the planning table and are the most important
tool for creating, analyzing, and modifying of the planning data. Every data
view stands for one task that has to be done by one corresponding person
(role).
SAP has prepared many data views for standard plannings.
Now we can start the demand planning. To demonstrate various possibilities
of the forecast strategies, follow the severely simplified examples.

Follow the menu path:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Interactive Demand Planning

If you are asked, whether a planning version should be installed, press no.
Search for the data view of your planning book in the section Planning
book/Data view, open the navigation tree and execute a double click.

© SAP SE Page 17
 CASE STUDY

Now press and type in the following data for the filter:
Show: APO product APO product

APO Planning Version: GBI-## GBI-##

APO Location: DL10 DL10

DXTR1000 / DXTR2000
APO Product: DXTR1000 and DXTR2000 (via multiple
selection and saving the new window via )

© SAP SE Page 18
 CASE STUDY

Click on to save the selection. In this way you can access both products
later without creating a filter one more time.

© SAP SE Page 19
 CASE STUDY

As a selection description type in GBI_DP##. Click on to save and close GBI_DP##

the popup.In the previous window click on to adopt.

Now you see both products at the top left in an area that is called Shuffler.

The next step is going to be assigning the selection to our profile. Click on
to maintain selection profile.

The window Maintain Selection Profile for Users opens. Search for the just
GBI_DP##
configured profile GBI_DP##. It will be on the right. Pull it via Drag&Drop
to the left in your user folder and click on to adopt.

© SAP SE Page 20
 CASE STUDY

Now we just have to assign our forecast profile. In the menu press
Settings ► Forecast Profile. Menu bar

GBI_DP##
Choose your Master forecast profile GBI_DP## and click on to adopt.

© SAP SE Page 21
 CASE STUDY

A double click on DXTR1000 in the Shuffler. It will load the product and we
can now type in our past data.

© SAP SE Page 22
 CASE STUDY

The demand volume is given in the following table. Scroll to the left in the
planning area and start filling out the table with the figures given below. You
should end in your own week.

1 2 3 4 5 6 7 8 9 10 11 12 13
2 6 11 15 19 24 34 24 19 15 11 6 2
14 15 16 17 18 19 20 21 22 23 24 25 26
2 6 11 15 19 24 34 24 19 15 11 6 2
27 28 29 30 31 32 33 34 35 36 37 38 39
2 6 11 15 19 24 34 24 19 60 11 6 2
40 41 42 43 44 45 46 47 48 49 50 51 52
2 6 11 15 19 24 34 24 19 15 11 6 2
*the top line is for guidance only.

The demand volume is cyclic and repeats as you can see every 13 weeks
(except the 36th week). It is so intended. It represents a seriously simplified
example for a seasonal fluctuating demand volume.

The 36th week exemplifies an outlier. This is a past value that lies outside
the awaited range of values. Outliers can falsify the forecast results. By
means of the outlier control feature this value will be corrected. We have
appointed for this purpose the Ex-post-Method before. Using this method the
system utilizes the Ex-post-Forecast to determine the tolerance range. If a
past value lies outside the tolerance range, the system understands it as an
outlier and corrects it.

The width of the tolerance range for the automatic runaway value control is
defined through the sigma factor. It sets the amount of the tolerated standard
deviations. The smaller the sigma factor, the lesser the tolerance and the
bigger the amount of the measured and corrected runaway value. The
standard sigma factor is 1,25. If you set your own sigma factor, SAP
suggests the setting between 0,6 and 2.

In the same way like for DXTR1000 produce the demand forecasts for
DXTR2000.
Double click on DXTR2000. If you are asked, whether you would like to
save, click always on .

© SAP SE Page 23
 CASE STUDY

1 2 3 4 5 6 7 8 9 10 11 12 13
1 1 1 1 3 3 3 3 5 5 5 5 5
14 15 16 17 18 19 20 21 22 23 24 25 26
7 7 7 7 9 9 9 9 11 11 11 11 11
27 28 29 30 31 32 33 34 35 36 37 38 39
13 13 13 13 15 15 15 15 17 17 17 17 17
40 41 42 43 44 45 46 47 48 49 50 51 52
19 19 19 19 21 21 21 21 23 23 23 23 23
*the top line is for guidance only.

As you can see the demand rises with time. It is a simple trend example.

Double click again on the selection DXTR1000. If you are asked, whether you
would like to save, click on . Then click on for Univariate
Forecast.

If you asked again about a forecast profile, choose your own (GBI_DP##) GBI_DP##

and verify your choice by clicking on .

© SAP SE Page 24
 CASE STUDY

The results are presented in a table:

You can also observe the results in the form of a diagram if you click on
.

Your result should look somewhat like this:

© SAP SE Page 25
 CASE STUDY

It is obvious that a forecast has been conducted. 52 weeks in the future are
sufficient. The seasonality was taken into consideration. As expected, the
values of the forecast should correspond exactly to our past data. Click on
to save your results and then on to go back to the main menu.

© SAP SE Page 26
 CASE STUDY

Step 3: Other Univariate Forecast Models

Task Create another univariate forecast profile. Time 15 min

Company's product portfolio consists of a variety of different products that

are illustrated at different stages of their life cycle and with various
characteristics. A universal profile for creating precise statistical forecasts
does not exist. A profile for fully developed, slow-moving, or new products
does not exist either. Approaches that try to enclose the majority of different
demand types are extremely complex and from user’s point of view have
characteristics of a „black box“.
Demand planning by SAP SCM still has a variety of useful and proven
forecast methods. A user can choose the best method for his requirements.
The forecast models that examine the past data for continual, trending, and
seasonal processes and display the corresponding forecast errors are named
univariate. Before we have conducted a seasonal forecast.
The univariate model is often called a time series model. In time series
models the forecasts derive from the evaluation of the past demand patterns
and trends. The key argument for the selection of a time series model is
therefore the process model of the previous demand data. In this case it is
assumed that the future demand follows the past one. The demand patterns of
the past are determined and reproduced in the forecast.

To create a new forecast profile, use the menu path:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Environment ► Maintain Forecast Profile

Choose your planning area 9ADP01 and your Master Profile GBI_DP##. In 9ADP01 / GBI_DP##
50 (Forecast with
the tab change to Univariate Profile and change the forecast strategy to 50 automatic selection)
(Forecast with Automatic Model Selection) and the smoothing factors
0,3
Alpha and Beta to 0,3.

© SAP SE Page 27
 CASE STUDY

Click on and then click on to adopt. In the tab change

again to the Master Profile and click on and then on to get back to the
main menu.

Open the demand planning:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Planning ► Interactive Demand Planning

Select your selection profile GBI_DP##finished products the product

DXTR1000 finished goods and click on for Univariate Forecast.

The result should look like in the forecast strategy 31.

© SAP SE Page 28
 CASE STUDY

The system has different statistical tests and test combinations that can
determine a model. To choose the best forecast profile, the system analyzes
the past data. If no explicit time series processes are detected, the system will
choose the continual model automatically. In our case the system detects the
seasonal changes and selects the seasonal strategy.
You don´t need to save this forecast.

Change the forecast strategy:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Environment ► Maintain Forecast Profile

In the tab change to Univariate Profile and change the forecast strategy to 21 21 (Holt´s Method)
(Holt's method).

Click on and then click on to adopt. In the tab change

again to the Master Profile and click on and then on to get back to the
main menu.

© SAP SE Page 29
 CASE STUDY

Go back to the demand planning:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Interactive Demand Planning

Select DXTR2000. Click on for Univariate Forecast.

If you asked again about a forecast profile, choose your own (GBI_DP##) GBI_DP##

and verify your choice by clicking on .

The forecast is conducted according to the Holt’s method and is suitable if the
past values are depicted through a rising or falling trend. Click on
for graphical view.

© SAP SE Page 30
 CASE STUDY

As you can see this time the past data was rising. With the given pattern the
trend procedure predicts rising of the future demand data.
We have given Alpha a small value because we wanted to involve the past
data into the forecast in the most balanced way.
The smoothing factor balances the past data in a way that the most up-to-date
values are considered stronger than the older ones.
You don´t need to save this forecast.

How fast the forecast reacts to changes depends on the smoothing factor. If
Alpha is 0, the new average value equals the previous one. In that case the
basic value that was once calculated does not change. It means that the
forecast does not react to the current consumption data. If Alpha is 1, the new
average value will equal the last consumption value.
Reasonable values for Alpha are between 0.1 and 0.5.

Now change the forecast strategy to 94 (Linear Regression) and execute the 94 (Linear Regression)
forecast with DXTR1000 in the demand planning.

© SAP SE Page 31
 CASE STUDY

Because of the constant seasonal changes the forecast is also constant.

The linear regression is a statistical method that can be applied to the

forecasting of trends. On the contrary to the most forecasting methods for
trends, the forecasting parameters are not determined by the values that come
out after the first approval and then are improved between the periods. The
linear regression considers all the data much better. It sets a straight line
through the data in a way that even the smallest errors (sum of square) are
determined.
The linear regression does not need any parameters such as Alpha or Beta.
The only parameter you can set is the trend dampening profile. In this case
we do not need any trend dampening.
Execute the forecast also with DXTR2000. One can notice the linear
procedure here as well.

© SAP SE Page 32
 CASE STUDY

Choose DXTR1000 again and delete the whole available data. Scroll back to
the week again and type in the following data:

Use the following data:

1 2 3 4 5 6 7 8 9 10 11 12 13
3 7 12 16 22 27 37 27 24 20 16 11 7
14 15 16 17 18 19 20 21 22 23 24 25 26
9 13 18 22 28 33 43 33 30 26 22 17 13
27 28 29 30 31 32 33 34 35 36 37 38 39
15 19 24 28 34 39 49 39 36 32 28 23 19
40 41 42 43 44 45 46 47 48 49 50 51 52
21 25 30 34 40 45 55 45 42 38 34 29 25
Click on to save the datay you entered.

Now select DXTR2000 and proceed like before:

1 2 3 4 5 6 7 8 9 10 11 12 13
19 8 11 0 2 0 25 15 0 4 0 39 1
14 15 16 17 18 19 20 21 22 23 24 25 26
56 0 0 16 0 37 12 3 21 25 10 19 11
27 28 29 30 31 32 33 34 35 36 37 38 39
32 39 51 0 0 10 53 57 0 14 74 18 9
40 41 42 43 44 45 46 47 48 49 50 51 52
8 42 3 0 36 0 0 51 5 20 24 0 26
Click on to save and then on to go back to the main menu.

© SAP SE Page 33
 CASE STUDY

Change the forecast strategy to 41 (Holt and Winters' Exponential 41 (Holt and
Winters´Exponential
Smoothing) and change the smoothing factors (Alpha, Beta, Gamma) again Smoothing)
to 0,1 each. 0,1

Select in demand planning DXTR1000 and execute Univariate Forecast by

clicking on .

© SAP SE Page 34
 CASE STUDY

The forecast works according to the multiplicative procedure by Winter/Holt.

It is suitable if the past values fluctuate seasonally over the rising or falling
trend. With the rising trend the fluctuation becomes stronger.
You don´t need to save this forecast.

At last choose the forecast strategy 80 (Croston’s Model). 80 (Croston´s Model)

© SAP SE Page 35
 CASE STUDY

Select in demand planning DXTR2000 and execute Univariate Forecast by

clicking on .

Croston’s method if a forecast strategy for products with sporadical demand.

It consists of two steps. First of all, the separate estimated values based on the

© SAP SE Page 36
 CASE STUDY

exponential smoothing will be deduced from the average demand rate.

Afterwards the calculation of the average duration between demands will be
done. It will be presented in the form of a continual model for the forecast of
the future demand.
The exponential smoothing in connection with the portfolio leading systems
is often used for the forecasting of demand.
However, if you have sporadic demand, the application of this method leads
almost always to inadequate stocks. Croston’s method is suitable in the case
when demand appears randomly, or when in many periods if not even in the
most of them there is not demand at all. In the cases when demand exists the
past data is distributed in any order: quasi independently from the demand
interval. Such demand procedures are called irregular or sporadic demand.
Click on to save and then on to go back to the main menu.

© SAP SE Page 37
 CASE STUDY

Step 4: Multiple Linear Regressions (MLR)

Task Execute a MLR. Time 15 min

A Multiple Linear Regression is based on the casual factors, such as prices,

advertising tactics, budget, and campaigns. The system uses the MLR in
order to calculate the influence of the casual factors on the demand in the
past. In this way you can analyze the success of single activities. With the
help of the computed correlation between the casual factors and the past
demand one can model future activities.
MLR is a form of a Multiple Linear Regression where one can examine the
connections between one single dependent variable and several independent
variables. You use the independent variables you know to forecast the single
independent values.
This forecasting method is not suitable if the independent variables behave in
the similar way and the MLR cannot find out the influence of the single
variables (correlation); or if the values of the past periods in the forecast
model influence the values of the current periods (autocorrelation).
Time series with distinct seasonal or cyclical progress are often strongly
auto-correlated.

To execute a MLR, follow the menu path:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Environment ► Maintain Forecast Profile

In the tab choose MLR Profile. Type in GBI_DP##_MLR for the Profile and GBI_DP##_MLR
MLR PROFILE
describe it as MLR Profile GBI_DP##. GBI_DP##
9ATOTFC
In the section past for the dependent key value Y choose as Key Figure
9ATOTFC (Total Forecast) and as Version enter GBI-##. GBI##

Now we need time series for our independent variables. Click on

for selecting Time Series.

© SAP SE Page 38
 CASE STUDY

Price
Promotion
In the popup enable checkboxes for Price and Promotion and click on to
continue.

Look at both time series in the popup Time Series Maintenance and click on
to adopt.

© SAP SE Page 39
 CASE STUDY

Next Monday
In both cases look for the next Monday of 11 months ago and choose the - 11 months
+ 4 weeks
date 4 weeks later as Start.

Click on and then click on to adopt. In the tab change

again to the Master Profile.

© SAP SE Page 40
 CASE STUDY

Multiple Linear
In the tab Master Profile enable the checkbox for Multiple Linear Regression
GBI_DP##_MLR
Regression and choose the just now created profile GBI_DP##_MLR, if
this hasn´t been done automatically.

Click on to save the master profile and then on to go back to the main
menu.

Now go back to demand planning:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Interactive Demand Planning

© SAP SE Page 41
 CASE STUDY

Select the product DXTR1000 and overwrite its data using the following
values:
1 2 3 4 5 6 7 8 9 10 11 12 13
16 23 22 23 23 25 25 25 24 23 22 23 23
14 15 16 17 18 19 20 21 22 23 24 25 26
33 35 35 35 34 34 34 34 34 35 35 35 35
27 28 29 30 31 32 33 34 35 36 37 38 39
44 45 45 45 46 45 45 45 45 47 46 47 47
40 41 42 43 44 45 46 47 48 49 50 51 52
37 34 34 34 34 35 35 35 35 34 34 34 34

Click on for Multiple Linear Regression (It may happen that here you will
be asked to select your forecasting profile one more time. You may also be
asked whether you want to save your changes, click on .

The result should look like that:

© SAP SE Page 42
 CASE STUDY

If you click on to switch Parameter On/Off in the table view, you can see
the fit measures. (If your current year has 53 weeks, your results might be a
little different)

© SAP SE Page 43
 CASE STUDY

Click on to save and then on to go back to the main menu.

© SAP SE Page 44
 CASE STUDY

Step 5: Composite Forecast

Task Execute a composite forecast. Time 20 min

This feature unites the forecasts that have different forecasting procedures
(for example, time series, multiple linear regression, and/or estimation
procedure) and were made for particular brands, product lines, or single
products. Every forecast exploits the same past data, however they have
different procedures. The goal is to combine single forecasts with different
strengths to obtain one forecast. You can either find the average value of the
forecasts by means of equal emphasis of the single forecasts, you can assess
every forecast differently, or you can change the weighting of the single
forecasts through time. Combination of the forecasts is supposed to provide a
better forecast for a company. It became evident that the composite forecasts
that are based on different mathematical and/or estimation procedures are
often superior to the single forecasts and their underlying procedures.

Follow the menu path:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Environment ► Maintain Forecast Profile
GBI_DP##_COM
Choose the tab Composite Forecast, type in GBI_DP##_COM as a profile, Composite Profile
GBI_DP##
and Composite Profile GBI_DP## as a description. The mode is standard. Standard mode

Click on .

Select GBI_DP##_UNI from the list. You can select several profiles here if GBI_DP##_UNI

they were created before. Click on to continue.

© SAP SE Page 45
 CASE STUDY

Click on .

Confirm the following popup by clicking on .

Select GBI_DP##_MLR and click on to continue. GBI_DP##_MLR

Now we can create another weighting. In our example we have chosen 40 for Univariate: 40
MLR: 60
the Univariate and 60 for the MLR.

© SAP SE Page 46
 CASE STUDY

Click on and then click on to adopt. In the tab change

again to the Master Profile.

Enable checkbox for Composite Forecast and choose the just now created Composite Forecast
profile GBI_DP##_COM, if this hasn´t been done automatically. GBI_DP##_COM

Click on to save the master profile and then on to go back to the main
menu.

Go to the interactive planning:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Interactive Demand Planning

Select the product DXTR1000 and start the Composite Forecast by clicking
on . It might happen that you will be asked again to select your forecast
profile.

© SAP SE Page 47
 CASE STUDY

Now the selected forecast profiles will be shown and the results will be
displayed in the table.
In our example the results will look like this:

© SAP SE Page 48
 CASE STUDY

Click on to save and then on to go back to the main menu.

© SAP SE Page 49
 CASE STUDY

Step 6: Release the Demand Planning

Task Release the Demand Planning for the Supply Network Planning. Time 15 min

Before releasing the demand volume we shall adjust the forecast for the last
time.

Follow the menu path:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Environment ► Maintain Forecast Profile
31 (Seasonal
Go to the univariate profile and change the forecast strategy back to 31 and Winters´)
0,50
change Alpha, Beta, and Gamma to 0,50 each. Set the periods to 52. 52

Click on and then click on to adopt. In the tab change

again to the Master Profile.
Click on to save the master profile and then on to go back to the main
menu.

Now go back to the Demand Planning one more time:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Interactive Demand Planning

Execute the Univariate Forecast by clicking on for DXTR1000 and for

DXTR2000.

The results should look like that:

© SAP SE Page 50
 CASE STUDY

For DXTR1000:

For DXTR2000:

© SAP SE Page 51
 CASE STUDY

Click on to save and then on to go back to the menu.

We have the forecasts about our future demand. But now we should make for
them a production plan. When we release the demand planning, the system will
generate the necessary planned independent requirements.

For that we need a Forecast Split and Distribution Profile, so the demand will
be well distributed within the periods.

To release the demand planning we need a Distribution Profile.

To create the profile, follow the menu path:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Environment ► Current Settings ► Maintain Distribution Function

GBI-##
For the distribution function type in GBI-## and click on .

Our demand volumes are planned for every week. We want that they will be
distributed to the first (Monday) out of 5 working days (5 periods) (holidays
will be skipped automatically).

For description enter Distribution Function GBI-## and for length enter 5. Distribution Function
GBI-##
Click on . 5

© SAP SE Page 52
 CASE STUDY

Here you can distribute your demand to 5 periods you set up before. For the
1
first period type in 1.
Save the data by clicking on .

Click on to go back to the main menu.

Now we are going to create a Period Split Profile.

Follow the following menu path:

Menu path
Advanced Planning and Optimization ► Demand Planning ►
Environment ► Current Settings ► Maintain Period Split Profile

Click on to create a period split profile.

© SAP SE Page 53
 CASE STUDY

GBI-##
For profile enter GBI-## and for description enter Period Split Profile GBI-
Period Split Profile
##. Click on to add split level. GBI-##

52
In the created row enter 52 for the number of periods. As mentioned before we
Week
want to distribute weeks to days. That is why we select week as starting
periodicity and day as target periodicity. For the Distribution Function enter Day
GBI-##
the already made Distribution Profile GBI-##.
Click on to save and then on to simulate period spilt profile.

© SAP SE Page 54
 CASE STUDY

Current date
In the following screen select current date as starting time and the date in a Today in a year
GBI-LOC
year from today as end time. For the calendar you should take GBI-LOC and
for the quantity per interval enter 100. Now execute the simulation. 100

Click on to execute.

The results should look like this:

© SAP SE Page 55
 CASE STUDY

The amount 100 should always be assigned to Mondays (except on holidays).

Click on to go back to the main menu.

Now we can release the demand planning.

To release the demand planning, follow the menu path:

Advanced Planning and Optimization ► Demand Planning ► Menu path
Planning ► Release ► Release Demand Planning to Supply Network
Planning

© SAP SE Page 56
 CASE STUDY

Type in the following information:

Data Source:
Planning Area: 9ADP01 9ADP01

GBI-##
Planning Version: GBI-##
9ATOTFC
Key Figure: 9ATOTFC (Total forecast)
Target:
GBI-##
Planning Version: GBI-##
FA
Category: FA (Planned Independent Requirements)
Horizon:
From date: next Monday next Monday

A year from the start

To date: a year from the start date date
Periodicity:
Period Split Profile
Period Split Profile: Period Split Profile GBI-## GBI-##

Object Selectn:
DXTR1000
Product: DXTR1000 and DXTR2000 (through DXTR2000
Multiple selection) DL10
Location: DL10
Results Log
Enable the checkbox for Results Log.
Click on to execute

© SAP SE Page 57
 CASE STUDY

If you get this notification and the to date changes, click again on .

You will get this log:

© SAP SE Page 58
 CASE STUDY

Now you can check the results. Click on to go back to the menu and go to
the product view:
Menu path
Advanced Planning and Optimization ► Production Planning ►
Interactive Production Planning ► Product View

In the popup “Welcome to the Product View” choose SAP_SCO and confirm it
by clicking on .

GBI-##
For the Planning Version enter GBI-##, for the product DXTR1000, and for DXTR1000
DL10
location DL10. Click on to continue.

© SAP SE Page 59
 CASE STUDY

The planned independent requirements should be terminated every Monday,

0:00:00 and the amounts should match with our forecasted demands.
Single amounts can be distributed to other days (for example because of
holidays).

Change the product to DXTR2000 and click on to update. DXTR2000

Now you can control the demand planning for DXTR2000 as well.

© SAP SE Page 60
 CASE STUDY

If you spotted an error, you can delete the just made transaction data and
execute the planning one more time. Use the transaction:
Menu path
Advanced Planning and Optimization ► Supply Network Planning ►
Environment ► Delete Transaction Data
Here in Objects Selection choose By Categories und type in FA. For the
Planning Version type in GBI-##, then the Product Numbers DXTR1000 and
DXTR2000 (via multiple selection), and for the Location type in DL10. Press
Execute. Check the selected data in the log and choose Delete selected objects.

Click on to go back to the menu.

© SAP SE Page 61