Managing Business Rules and Analytics as an Enterprise Asset

Managing Business Rules and Analytics as an Enterprise Asset

Discover Financial Services Business unit of Morgan Stanley Operates the Discover Card brands Largest proprietary credit card network in the United States 50 million Cardmembers 4 million merchant and cash access locations Approximately 14,000 Employees

McKinsey Quarterly: Top 10 Trends to Watch in 2006 Macroeconomic Trends Centers of economic activity will shift profoundly, not just globally, but regionally Action item: Companies need to: Identify competitive strategies (such as customer intimacy or price optimization) that are impossible to duplicate without strong analytic capabilities. Identify how analysis can be integrated into every customer centric decision the enterprise makes. Gartner strategic planning assumption: Service-based enterprises that establish customer analysis as a core competence and a basis for competitive differentiation will have profit margins at least 5% higher than competitors pursing other strategies (0.8 probability). Consumer landscape will change and expand significantly Consumer preferences are becoming more complex and changing more rapidly, experiencing a wider array of interactions, and becoming more demanding. Action item: Build an integrated understanding of the customer by understanding what is likely to drive the evolution of the customer s wants and need. Gartner SPA: By 2015, an inability to understand changes in customer and channel preferences and usage will be the biggest source of customer dissatisfaction, resulting in churn rates at least 10% higher than for enterprises that are more responsive to changes in customer expectations and preferences (0.6 probability). Source: Gartner, Inc., published: March 29, 2006/ ID Number G00138137 Source: McKinsey Quarterly, Article Ten Trends to watch in 2006, published January 13, 2006

McKinsey Quarterly: Top Ten Trends to Watch in 2006 Societal and Environmental Trends Technology connectivity will transform the way we live and interact Customers will increasingly choose how they would like to interact with organizations and will only do business with companies that meet their interaction needs. Action item: Companies need to build integrated insight into the customer s use of channels. Integrate channel usage data that tracks the customer during each interaction, and ask Why did the customer use this channel? Gartner SPA: By 2010, the ability to track customer interactions will be replaced as a competitive differentiator by an organization s ability to understand why customers choose a particular means of interaction. This knowledge will enable organizations to be least 20 % more successful in guiding customers to use their preferred channel (0.6 probability). Source: Gartner, Inc., published: March 29, 2006/ ID Number G00138137 Source: McKinsey Quarterly, Article Ten Trends to watch in 2006, published January 13, 2006

McKinsey Quarterly: Top Ten Trends to Watch in 2006 Business and Industry Trends Management will go from art to science ubiquitous access to information is changing the economics of knowledge Today s business leaders are adopting algorithmic decision-making techniques and using highly sophisticated software to run their organizations. The ability to generate interesting analysis is useless without the ability to deploy that insight to drive business change. The growing availability of information within enterprises will require considerable focus on the ability to use that information, rather than creating write-only data warehouses or ivory towers of customer insight, safely removed from any chance of their use. Action Item: Ensure that the mechanism to effectively deploy and use the analysis exists before the analysis is performed. Ensure that users understand the process that created the analysis, including assumptions and implications. Gartner SPA: Through 2010 more analysis will be done and remain unused in enterprises than is requested and never done (0.7 probability) Source: Gartner, Inc., published: March 29, 2006/ ID Number G00138137 Source: McKinsey Quarterly, Article Ten Trends to watch in 2006, published January 13, 2006

Advanced Decisioning Environment Automate the process with business rules management Increase decision precision with predictive analytics Optimize strategies with predictive analytics Innovative decision technology Innovative new data/attributes Homogeneity in quality Capabilities We are here today Account Level Economic Logic, NPV Total Customer Relationship Level NPV Optimization Statistical Models Data Mining Criteria/ Rule-based Decisions Time

Business Drivers Need a higher return from previous infrastructure investments CRM systems, data warehouses, business intelligence, Web Harder and harder to eliminate costs using technology Movement to automate higher level decision tasks Push for in-stream decisions or straight through processing Increasing business decision complexity Globalization of operations Policy and regulatory pressure M&A activity Competitive pressure for more sophisticated decisions Accelerating pace of business change Shorter windows of competitive advantage Speed of business is outpacing speed of IT to react

Business Expectations Define key business definitions such as customer definitions in a centrally located business controlled repository for common consistent customer treatment. Right techniques on the right customers in the collections environment. Increase use of models and risk strategies across the enterprise. Specialized models for refined segmentation. Multiple scores in decision making. Coordinated efforts implemented across the enterprise as well as decrease time to market for models and business rules deployment. Bottom Line Revenue generation Loss reduction Allow for the use of sophisticated analytical techniques such as neural network technology, optimization, rules management as a component in a real-time customer environment. Take advantage of known information about CMs to improve decision making at key customer touch points; Drive additional revenue from low risk CMs / transactors, and reduce our exposure from high risk CMs. This information could be internal information, credit bureau data, demographic data, or predictive models.

Advanced Decisioning Environment EDM Overview D E V E L O P M E N T & M A N A G E M E N T T R A N S A C T I O N E X E C U T I O N C H A N N E L S DECISION SERVICE Call Center Business Intelligence Analytic Development Modeler Tools Analytic Services Decision Request O P E R A T I O N A L S Y S T E M S ERP CRM Web Email Telemarketing Business User Tools Rule & Model Repository Configuration & Deployment Services IT Tools Rules Services Data Services Decision BILLING SCM Direct Mail Store/ Branch Kiosk/ ATM Rules Management Field Business User Tools Decision Analysis Reporting & Adaptive Analytics Services External Data Source(s) Data Warehouse Operational Data Store Customer Behavior and Strategy Performance

Benefits Across the Decision Management Cycle Strategy Evaluation Development time Dependence on BT resources Maintenance costs Easier Management Re-use Quality / validation Analytic methods Time to Deploy Development $ Business impact Fewer steps to production Analytic Modeling EDM Strategy Design and Deployment Decision Quality Consistency Compliance Areas of Deployment Business Impact Strategy Flexibility Development $ Closed Loop Continual Improvement

Decisioning Infrastructure Other CPB IMS Statements ACAPS Credit Bureau Data Benefits Provides one source of truth for all customer information available for decisioning Supports current and future decision management tools and methods Scoring and Rules may be separated increasing the flexibility available to the business Balance and control Blaze Extract(s) EDS Profile Data Sets (Master) Strata Extract(s) SAS Strata Blaze MB S C O RI N G S T R A TE Blaze (Batch) Confidential For Internal Use Only Blaze (Batch) G Y Risk System Management

EDM Component: Model Builder Executable Model Code Operational System 1 Models and Strategies Operational System 2 Models and Strategies FICO-built Models Transaction Score, FALCON Fraud Model, Merchant Fraud Model, etc. Model Builder Model Builder Modeling Data EDS Insight SPDS Other Executable Business Benefits Innovative decision technology Innovative new data / attributes Homogeneity in quality Models and Strategies Operational System 3 Blaze Models and Strategies Operational System 4 Executable Model Builder is used to access analytical data and construct models Model Builder output is sent to Blaze and the model is integrated with strategies and then deployed into production Model Builder can also produce executable code that can go directly to production systems (e.g., Transaction Score)

EDM Component: Blaze Blaze generates the Portfolio Risk Score (PRS) and executes other models from Model Builder after cycle processing Oracle DB (Operational Reporting) EDS Blaze uses model scores and other data to decision credit line and cash line actions, early account management decisions manage high risk accounts Blaze sends score, credit and cash line action and force to collections non-mons to CPB A record of all decisions made by Blaze are kept for operational reporting and future analytics User Insight (Analytical Reporting) CPB Downstream Systems Reporting Data Non-Mons to CPB (Scores & Actions) Account Memos Vision Blaze Input File Blaze Unix Adverse Action Reasons File AA Job (Blaze HLQ) Monthly Summary Reasons File (Tape) AA Letters PLU

Enterprise-Wide Touch Points Acquisition Post Screen Processing Strategy Evaluation Authorizations Management Customer Solicitation Management Portfolio Scoring Customer Contact Strategy Disputes Processing EDM Early Account Management Credit Line Management Marketing Eligibility Portfolio Pricing Analytic Modeling Strategy Design and Deployment Collections and Recovery Fraud Detection and Management Closed Loop Continual Improvement

Challenges / Lessons Learned Data infrastructure / product integration Competing project priorities / project management Simulation environment prior to production Plan for ongoing expansion of architecture Properly trained resources Senior management commitment