This is the second Post about the influence of Target Levels. The realm of the first post was limited to the Segment Designer. This post will focus on the Load of Segments to Siebel in conjunction with the Load List Format and compare this behaviour again with the Segment Designer.
Target Level and Segment Designer
As already described in the previous Post, the Segment designer allows counting the target level members according to the segmentation criteria of the segment, the count of each block and the the gross count is displayed, hence indicating the members which are part of the segment and will be loaded into the campaign as members once the campaign is loaded.
The below SQL statement was generated by the Oracle BI Server for the Segment Designer to count the Target Level Members within the Business Analytics Warehouse:
Target Level and Campaign Load
The Oracle BI server will write a text file to the Siebel Shared File system and Siebel will load the text file using a EAI Workflow to the Campaign Membership (S_CAMP_CON Table).
To provide the text file in the expected format by the Siebel EAI Workflow, the Oracle BI server provides the Load List Formats designer to provide the means of loading the individual members to the Campaign Membership Table so that they are participating within the Campaign.
Once the Campaign is loaded Siebel sends a SOAP request to the Oracle BI server to execute the segment and generate the text file with the Campaign Members using the Write List Job (visible with the Marketing Job View):
When comparing the SQL statements executed from the Segment designer with the SQL Statement for the Campaign load it can be seen that they are not the same. This is partially due to the different nature since within the Segment designer, the User is interested in the count, hence a SELECT COUNT(*)while the load requires the individual records, hence no aggregation.
Even though the statements are not the same, the logic (in terms of underlying Tables and Joins) for the Load List format (plus underlying RPD Metadata definition) is matching the logic of the Segmentation Catalogue. Thus, the generated text file and the segment count have the same count of Members (# of Records) for the Segment:
To summarize: The Target Level links the Segmentation catalogue with the Load and Export List Formats, but the generation of SQL statements is independent for the areas and only depends on the metadata definition on which they are based:
Since the generated statements for the Segmentation Designer (based on the Segmentation Catalogue) is independent of the generated statement for the Campaign Load (based on the Load List Format) it’s possible to create a Scenario in which both do not provide the same results (count for Segment and Load) even though both are linked to the same Target Level. This will be the focus of the next article.
For some scenarios this might be desirable since the Campaign Load Format is aware of the Campaign (it is loaded to) and the Treatment (Channel), thus it would be possible to perform certain additional checks e.g. does the Campaign Member have a valid Address for the Channel. However, this additional “segmentation” logic makes it very hard for a Marketing End User to compare the results of the Segment Designer with the actual Members loaded to the Campaign.
Launching the Campaign
When launching the Campaign the impact of the Target level is also very minor. As described before The Export List format definition is also based on the BI data model and determines the actual SQL generated on the BI System (Data Warehouse). In order to only generate the respective Members of the Campaign the BI System references the Siebel Transactional Campaign Membership Table (S_CAMP_CON) by passing Parameters from Siebel to the Oracle BI System to identify the respective Campaign via the Campaign Id and Load Number.