ABC Analysis Web Steps

BW’s Little-Known ABC Analysis Tool Has Wide Logistics
Application: Expanded Setup and Analysis Steps
by Ned Falk, Senior Education Consultant, SAP America
Step 1. Create the model: A data mining model in BW provides the specific fields and
settings you are to utilize in performing the data mining application at hand. The idea is
that only a few people in your organization know the details behind many of the model
settings, so it is easier to determine the settings once, by the experts. Once the model is
determined, others a little less technical, can utilize the model many times for various
data sets.
To access the data mining workbench and ABC analysis access the transaction code.
RSDMWB. Alternatives are to follow the menu path under the main menu Data Mining
>Date Mining Workbench, or to access the data mining button on the BW
Administration Workbench. Be aware that data mining existed as of version 2.1C, but
alternative access paths are not available until version 3.1. Once in the Data Mining
Workbench, access the ABC Analysis Application under the Further Analysis
Subsection. (It is located in the Classification subsection in older releases of BW.)
Figure 1 Data Mining Workbench
By using the context menu (right-click), you can use the wizard to help you create the
model. The screens of the wizard appear below. In the first screen, you assign the name
of the model and select the query you will use to help you define the field names to be
used by the model. This is not necessarily the query you will use to execute them later,
but it should be related, just for common sense reasons. For example, if you are planning
on doing an ABC of customers, then Sold_to or Customer or some thing similar should
be on the query you select for the model definition phase.
Figure 2 Wizard Model Definition screen
In the next part of the model creation, you determine which fields from the query will
actually be used to define the model. In the screen shot below, I have selected Revenue
and Business Partner, and the field Customer Classification to hold the resulting
calculated ABC code for our model. To do this, select the fields on the right hand side
and using the arrow icon in the middle of the screen bring them into the model.
Again, it is not critical which fields are selected for the model, but if you want to make it
more intuitive later in the process, pick the fields that are logical. For example, if the
ABC analysis model was called Customer ABC by Sales Revenue, pick the field most
closely related to sales revenue and customer to build the model. You also might want to
use a query with 0ABC_CODE as a more intuitive field, but for demonstration purposes
later, let’s leave it like this. In this case, the query is simple and the decision is easy.
Figure 3 Choose InfoObjects for the model
In the next phase of model creation, you tell the application which field represents the
object you are classifying (BPARTNER), which field represents the values to be used to
do the classification (Classification Criterion=Revenue), and finally the name of a
characteristic InfoObject that can hold the resulting ABC code (0CUST_CLA). In this
case, 0SOLD_TO, 0INVCD_VAL and 0ABC_CLASS were chosen (Figure 4).
Note! In BW 3.1x, you can also specify a grouping characteristic. A grouping

characteristic such as Sales Organization would create ABC codes for your business
partners considering each sales organization as a separate group of customers. When
using absolute and the percentage methods below, the best BPARTNERS in every sales
organization would be classified as A customers. Without grouping your A customers
would be the best ones in the entire corporation.
Figure 4 Model fields
In the final phase of model creation, you define exactly what variety of ABC will be
performed. This is the phase when specific settings are made that control the outcome of
your analysis.
Figure 5 Model parameters
The first option in this phase is to determine which classification method to use. The
default, Cumulative Percentage of classification criteria, is the most common. The
numerical values on the screen mean different things based on which ABC classification
type is chosen. In some cases, they refer to absolute values and in others they refer
percentages. You must decide which type to use, and they are all effective depending on
your business purpose. However, in most cases I prefer the default, as it focuses your
energy on a fixed percent of your objects (customers), either the best or the worst ones.
Step 2. Define a source of data for the model: After the model has been defined, the
next step is use it with any source query that can provide related data. The easiest way to
do this is to continue to use the wizard as shown in Figure 6. It is important to note that
the query being used as the source does not have to be the same as the one used to build
the model itself. The only requirement is that you must be able to map the fields
(InfoObjects) from the source query to the model.
Figure 6 Assigning a source query to the model
Now that the source query has been identified, you must perform this mapping task as the
next step in the wizard (Figure 7). To accomplish this, link the Source Query field with
the related field in the model. In this case, the model fields are the same as the query
fields, so mapping is performed automatically.
Figure 7 Column mapping
After this step, the model is complete and the screen confirming this appears. You now
need to execute the ABC analysis and then do something with the results!
Note! It is now necessary to return to the Data Mining WorkBench and activate the
model and the training source. In some cases, this may require you to leave and then
return to the data mining workbench, to facilitate a refresh.
Step 3. Execute the model: To execute the model open the tree structure to the Source,
then use the context menu to Run ABC Classification, selecting the Online option.
After allowing the program to execute, you may view the results of the model using the
same basic menu path (Figure 8).
Figure 8 Execute/view results of the classification process
The classification results are shown in Figure 9. It shows the data in a similar manner to
the data we discussed in Figure 5. It includes data to support any of the four classification
methods, even though only one option was chosen. As an alternative to the data view,
you can access the graph icon to see a graphical presentation of the data. For the purposes
of this article, I will discuss the graphical output in the context of an integrated Web page
output in the next section.
Figure 9 ABC classification results (table format)
Step 4. Use the ABC classification: Having ABC classified your customers is great, but
unless you use this new information, what good is it? In prior releases of BW, limited
(but still effective) uses of the results were possible. In the latest (3.1C) version of BW, a
lot can be done with the output from your ABC model. Options include:
1. Uploading the ABC values to a related BW master data object (for example,
0Bpartner) *** This could be done in prior releases of BW ****
2. Uploading the ABC values to the master record in CRM.

3. Uploading the data to an ODS.
4. Create a DataSource
5. Integrate the output in a BW Web cockpit

ABC Analysis Web Steps

Uploaded by

Copyright:

Available Formats

ABC Analysis Web Steps

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ABC Analysis Web Steps

Uploaded by

Copyright:

Available Formats

BW’s Little-Known ABC Analysis Tool Has Wide Logistics

Application: Expanded Setup and Analysis Steps

by Ned Falk, Senior Education Consultant, SAP America

Figure 1 Data Mining Workbench

Figure 2 Wizard Model Definition screen

Figure 3 Choose InfoObjects for the model

Note! In BW 3.1x, you can also specify a grouping characteristic. A grouping

Figure 4 Model fields

Figure 5 Model parameters

Figure 6 Assigning a source query to the model

Figure 7 Column mapping

Figure 9 ABC classification results (table format)

2. Uploading the ABC values to the master record in CRM.

5. Integrate the output in a BW Web cockpit

You might also like