Editing Data Sources and Attribute Lists
Attribute lists are required for modeling segments during segmentation. The creation of attribute lists is the responsibility of the database manager, who aims to provide these lists to the marketing manager in such a way as to allow him or her to use the lists without requiring assistance from the database manager. Which attributes (and therefore which data sources) are relevant depends on the respective marketing activity, that is, the purpose for which the target groups are to be created.
The settings you require vary according to the data source in question. Depending on the origin type of the data source, the following prerequisites must be fulfilled:
Origin Type of Data Source | Prerequisite |
|---|---|
Attribute set | You have created attribute sets in CRM Marketing. For more information about how to do this, see Working with Attribute Sets. If you want to use this origin type for segmentation with high data volume (BWA), you must first have extracted the marketing attributes to an InfoCube, and entered the technical details of this InfoCube in Customizing. For more information about this, see Creating an InfoCube for Marketing Attributes. |
InfoSet | You have created an InfoSet query. |
SAP NetWeaver Business Warehouse (BW) Cube (only in classic segmentation) | You have created a BW report in SAP NetWeaver BW. You can do this using the Business Explorer Analyzer. |
External List Management (only in classic segmentation) | You have created an address list in external list management. |
File import (only for segmentation with high data volume (TREX)) | You have defined up to 10 different data sources with the origin type |
BW InfoProvider (only for segmentation with high data volume (BWA)) | An attribute cannot be used for selection in an InfoProvider, if the following points apply simultaneously:
The system automatically filters out these attributes when creating a data source. |
If you want to activate the fast find for certain attributes (step 6), you have installed SAP NetWeaver Search and Classification (TREX).
You create an attribute list as follows:
In Customizing for Customer Relationship Management
, choose .
Create data sources
When creating the data sources, you define where the attributes for your attribute lists come from (attribute set, SAP NetWeaver Business Warehouse (BW) Cube, InfoSet, external list management, CSV file import). If you have data sources of the origin type InfoSet
and are using a data reading program, you need to define whether the filters of an attribute are linked with AND or OR within an attribute and across all attributes. Filters that are mutually dependent require an InfoSet that always links the individual attributes with AND. In the standard delivery, no filter dependencies are defined.
For classic segmentation, the following origin types are possible for data sources:
Attribute set
InfoSet
SAP NetWeaver Business Warehouse (BW) Cube
External List Management
For segmentation with high data volume (TREX), the following origin types are possible for data sources:
Attribute set
InfoSet
File import (CSV file)
Note that you can create up to 10 different data sources of this origin type.
With data sources for high data volume (TREX), you also have the option of summating key figures. For more information, see Summating Key Figures in Data Sources (TREX).
For segmentation with high data volume (BWA), the following origin types are possible for data sources:
Attribute set
BW InfoProvider
Note
If you create a data source with the origin type BW InfoProvider
and the InfoProvider that was specified in the InfoProvider
field for this data source contains a non-aggregating key figure (such as a calculated key figure for a Query as InfoProvider
), the system automatically selects the Group By (grouping criterion )
indicator for all characteristics in this InfoProvider that have the object type CHA, TIM, or UNI. This indicator cannot be changed. Selecting this indicator prevents non-aggregating key figures being aggregated and producing incorrect results.
Create attribute lists
Choose a name and a usage for the attribute list.
Attribute lists that you want to use for segmentation with high data volume must have the category High Volume Segmentation
.
For more information about usages and usage categories, see Customizing for Customer Relationship Management
, under .
Assign data sources to attribute list
After saving your attribute list, you can assign the required data sources. To select multiple data sources, press and hold down the Shift or Ctrl key on your keyboard when selecting with the mouse. You can add as many data sources to an attribute list as you require. Open each data source and select the attributes that you want to include in the list. To select all attributes for a given data source, place your cursor on the data source and choose Select All Attributes
from the context menu.
Note
If the assigned data source has the origin type BW InfoProvider
and the InfoProvider that was specified in the InfoProvider
field for this data source contains a non-aggregating key figure (such as a calculated key figure for a Query as InfoProvider
), the field Summation
is set to As Defined in Data Source
in the attribute properties for all attributes in the attribuite list that correspond to this type of key figure and cannot be modified. This means that the Group By (grouping criterion)
indicator is selected for all characteristics in this InfoProvider that have the object type CHA, TIM, or UNI. See the note for point 1.
This affects the way in which this key figure is handled during high-volume segmentation.
Example: You have defined an attribute list that only has one assigned data source and this data source has the origin type BW InfoProvider
and the InfoProvider specified in the InfoProvider
field is as follows:
Business Partner ID | Year | Product Category | Number of Orders | Number of Order Items | Score |
|---|---|---|---|---|---|
33 | 2012 | Electronics | 12 | 20 | 5 |
33 | 2012 | Furniture | 3 | 5 | 2 |
48 | 2011 | Electronics | 4 | 8 | 6 |
56 | 2011 | Stationery | 23 | 42 | 9 |
56 | 2013 | Furniture | 2 | 3 | 3 |
56 | 2013 | Clothing | 6 | 10 | 5 |
If the Score
is an aggregating key figure, which is aggregated by summation and the Group By (grouping criterion)
indicator is not selected in the data source, it appears as follows in high-volume segmentation:
A segment that is defined as a filter with a Score
<= 6, would only find the business partner with the ID 48.
Assuming that no filter has been defined for this key figure, the distribution chart for the Score
would distribute the values in the Score
using the range 6–17. This implies: Minimum value of Score
= 6, maximum value of Score
is 17.
This is because in this case the values in the Score
are aggregated with reference to the Year
and Product Category
.
If the Score
is a non-aggregating key figure, it appears as follows in high-volume segmentation:
A segment that is defined as a filter with a Score
<= 6, would only find the business partners with the IDs 33, 48 and 56.
Assuming that no filter has been defined for this key figure, the distribution chart for the Score
would distribute the values in the Score
using the range 2– 9. This implies: Minimum value of Score
= 2, maximum value of Score
is 9.
This is because in this case the values in the Score
are not aggregated.
As a result of this, InfoProviders that contain non-aggregating key figures (such as Query as InfoProvider
with a calculated key figure), contain as few characteristics as possible.This means they should only contain the characteristics that are used to distinguish these non-aggregating key figures. In the example above if the Score
is a non-aggregating key figure and this is only to be broken down for each Business Partner ID
and Year
for business reasons, a corresponding InfoProvider should not contain any characteristics other than the Business Partner ID
and Year
. If additionally, in a different business scenario, the Score
is only to be broken down for each Business Partner ID
, a second InfoProvider should be created that does not contain any characteristics other than the Business Partner ID
.
Activate attributes
Select the attributes that you need for your attribute list.
Define filter for attributes
Place the cursor over the relevant attribute and choose Create Filter
. Enter the filter criterion and a description for the filter. Depending on the type of attribute, you can either select from predefined values or enter your own values.
Activate TREX fast find for filters
By selecting the TREX Index
checkbox, you can define that TREX fast find is used to determine the distribution for this attribute value.
The attribute data is then loaded to SAP NetWeaver Search and Classification (TREX). The system reads the distribution of attribute values to the business partners from the TREX index; the distribution does not have to be recalculated each time you start a count.
You can define the TREX index refresh frequency according to your requirements. To manually start the initial data transfer, choose Refresh TREX
. You should schedule value distribution updates in a background job. You can find the date of the last update in the attribute properties.
For more information, see TREX Fast Find for Attribute Distribution in Segmentation.
Note
If you activate the TREX fast find for an attribute, and this attribute can have multiple values simultaneously (for example, with the attributes 'Hobbies' or 'Children's Age'), you must select the Multiple Values
checkbox in the attribute properties, otherwise the system always uses only a single value of this attribute. If the attribute is a marketing attribute, the system selects the Multiple Values
checkbox automatically. However, for attributes from other data sources, you must select the checkbox manually.
Change attribute properties
You can change the attribute description, type, or how the attribute is displayed in segmentation, for example. For attributes deriving from BW InfoCubes or from InfoSets, you can also add filter conditions with values that should be valid for the key figure to be read. In the attribute context menu, choose Edit Filter Conditions
and enter the variable name and the corresponding values that you want to apply. For example, you could specify a certain period and sales office for the key figure "Turnover".
Under Properties
, you can specify an update interval for your attributes. This enables you to store the determined attribute distributions in a cache for a freely definable period of time, thus improving performance. Note that you have to schedule the report CRM_MKTTG_CACHE_UPDATE
for this, as a periodic job (transaction SM36
). For more information, see Customizing for Customer Relationship Management at .
If you have also installed SAP NetWeaver Search and Classification (TREX), and activated the fast find for the attribute, specify the Update Interval
in days. In this case, you must schedule the report CRM_MKTTG_TREX_UPDATE
as a periodic job.
You have created a data source, and a related attribute list that is now available within the segmentation for modelling.
Choose Segment Builder View
to display a preview of the attribute list as it appears in graphical modelling. Ensure that the attributes in the attribute list are arranged in a logical order.