There is a Segmentation Engine hidden inside OBIEE 188.8.131.52. This is being used extensively with the Siebel Marketing application which is a Oracle’s application offering for campaign management since days of Siebel Analytics.
There is no harm in using Segmentation Engine in standalone mode. One can extend any repository to support segmentation. There is nothing extra need to install to use segmentation engine and it is very useful for SET Analysis.
Oracle Answers vs. SET Analysis
Oracle Answers is terrific, we can use it analyze, respond, predict and plan, it is being enhanced with lots of new features in OBIEE 184.108.40.206 release.
But in Marketing, classification of potential entities (e.g. customers, accounts, contacts, sites etc) by one or more characteristics is the market standard; we would often need COUNT of entities based on what is not there rather than what is there.
For example a software company would want to generate a list of contacts based on following criteria.
“All Contacts at Sites where there are more than 50 seats of A, but no seats of B”
Let us try to generate a list of contacts for the above criteria by using Oracle Answers, it might need four requests, two of them use the ‘Filter Based on Another Request’ feature
#1: Build a list of Sites that have get 50 Seats of A
#2: Build a list of Sites that Own B
#3: Build a Request that lists Sites in #1, but not in #2
#4: Build a list of Contacts at Sites in #3
After creating four requests as mentioned below there are challenges in terms of putting this segment generation process on a dashboard, associated training and figuring out how to download the result and associate these to campaign execution.
In short Oracle Answer is not a right option for SET Analysis. Marketing Segmentation which is a part of OBIEE 220.127.116.11.0 provides the interface, integration and effective options to carry out the SET Analysis.
The SET operations supported and equivalent OBIEE Segment Manager Operations are as follows.
This blog is an attempt to explain the on how to use the segmentation and dashboard to implement SET Analysis within OBIEE 18.104.22.168.0.
Terminology – Segment, Target Levels, Qualified List Items
Segment – Synonymous to a SET. It defines a target set of Contacts, Customers, Sites (or any other entity).
Target Level – It defines the unique target identifier, usually a Natural Key (e.g. INTEGRATION_ID), not the Surrogate Key (ROW_WID)
Qualified List Item – Together with a Target Level, they answer the question: “What do you want to COUNT ()?”. Also used to generate Lists
Conforming Dimension Link – This allows you to cross Subject Areas (i.e. Segment “Installed Sites”, then find the “Contacts” at those sites
List Catalogs – When you want to generate Lists. (You may “count” many things, but you’re only Marketing to People; you don’t send an e-mail to a Product…)
List Format – Customized format for List pull. In other words, “What do you send to the people that generate the campaign?”
The Marketing Manager view (Manage>Marketing Manager) of OBIEE repository shows a predefined target level (customers), Qualified List Item (customers) and List Catalog (A - Sample App).
This section is an attempt to explain a end to end process to define a SET Analysis which includes defining a segment, setting up the link segment output with the dashboard report to get list and reports driven by the segment output.
Predefined Marketing Default Settings
Administrator> Set Marketing Defaults
Set Marketing Defaults, which manage the default setting such as default campaign load format, default global audience for marketing and profile dashboard for the target levels.
Design a Report to be driven by Segment Criteria
A report is designed with customer profile columns such as number, name, income level, city , country etc and a column x20 Cust Key (Segmentation) is set as prompted as a filter for the report. This filter is to make sure that the report should be driven by the segment criteria. A filter view is also included in the report to make sure that the compound view of report also shows the filter condition applied which is same as the definition of the segment criteria.
“All Customers who are members (customer type) from two groups of customers, one group where city is San Francisco. And Customer Home Level is within level 6 to level 9 and other group of customer where revenue is less than 10000 and must be from EMEA region”
New > Segment > Choose Target Level
Define two groups and exclude criteria as follows – using columns from the subject area exposed (A –Sample Sales)
Exclude – Non Members
Once done click ‘Update Counts’ to get the count, it shows the individual count at each level and segment count at bottom with allows to click to go to the Profile Dashboard which is a pre-defined dashboard with couple of reports, this dashboard has been set as a default dashboard for the customer target level as Administration of Marketing Server default settings.
A dashboard with couple of views for customer info is shown below.
Filter View of Compound Report View
This filter view is the defintion of the segment criteria.
I hope you find this information useful. I would like to cover how to define the repository metadata for to support marketing segmentation with appropriate details in future.