Siebel Analytics Installation and Configuration Guide > Configuring Metadata for Analytics Applications >

Developing and Deploying Predictive Scores

The Loyalty Management Dashboard and several Siebel Analytics subject areas use customer scores generated from a third-party predictive modeling application. A predictive modeling application uses mathematical models to predict customer behavior. For customer scoring to be made available for analysis in Siebel Analytics, CME metadata is provided which maps these customer scores to dashboards and subject areas.

The following procedure describes the process of developing and deploying these predictive scores.

To develop and deploy predictive scores

1. Generate predictive scores using any third-party predictive modeling application.

   NOTE:  This is performed outside of the Siebel operational application.

2. Integrate the scores into the Siebel transactional database.

   Once this is completed, scores may be viewed in the Siebel operational application by accessing the Accounts > Profiles > Loyalty Profile view.

3. Load the integrated scores into the Siebel Data Warehouse during the extraction, transformation, and loading (ETL) process.
4. After the scores are loaded into the Siebel Data Warehouse, map them to the following Siebel Analytics metadata fields:
   • Churn Score
   • Customer Lifetime Value Score
   • Upsell Score
   • Cross-Sell Score
   • Financial Risk Score

     In conjunction with other associated metadata, these fields are primarily used to populate the Loyalty Management dashboard.