Express Scripts Enterprise Data Warehouse Joe Federer, Senior Director EXPRESS SCRIPTS A bit about the Journey • • • • • Introduction Business Story Implementation – SuperBowl Sunday Enterprise Logical Data Model Summary Who Said It??? “Our health is far behind the progress of medical science…Proper medical care is so expensive that it is out of reach of the great majority of our citizens.” • • • • • George W. Bush William J. Clinton Lyndon B. Johnson John F. Kennedy Harry S. Truman State of the Union address - 1949 Demographics Are Crucial The Population is… • growing • aging • getting heavier Who Is Express Scripts? Contact Centers • Arizona • Colorado • Georgia • Michigan • Minnesota • Missouri • New Mexico • New York • Pennsylvania Home Delivery • Arizona • Missouri • New Mexico • Pennsylvania Corporate Headquarters: St. Louis, Missouri Revenue Growth with Acquisitions $18,000 $16,000 $14,000 Specialty Platform Millions $12,000 Labor/ NE/ SG&A $10,000 DIVERSIFIED ® $8,000 $6,000 $4,000 Large Employer/ Mail Managed Care / Rebates / SG&A Middle Market $2,000 $0 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 Teradata Implementation - Progression Pharmacy Reference Adjudication DB-2 RS-6000 SP-2 8 nodes AIX 4.3 DB-2 4 TB / 1.8 TB Data Marts Teradata New Enterprise Logical Data Model Client Ad-Hoc Data Marts Silo Silo Sales Silo Silo Ops Silo Silo Call Ctr Silo Silo Silo Prod Mgt Silo Claim Silo Silo Silo Silo Silo On-going Silo Mktg July 2006 P690 AIX 5.1 UDB 2 TB /800 GB April 2006 RS-6000 SP-2 1 nodes AIX 5.1 DB-2 200 GB / 50 GB Client Portal Fin RS-6000 SP-2 2 nodes AIX 5.1 DB-2 800 GB / 200 GB April 2006 Q1 2007 Operational DW DB-2 2 TB Decision Support Feb 2006 s t a g e “Tall and Skinny” to EDW Goal State Current Cl ai M ms G em L/ Fi ber na nc Be e n P Ho re efit m sc r e i De ber liv er y Dr ug Sa l Co es n Sp t ac ec t ia lty s er it er ry ug es act lty e f c m l e ai mb an e n c rib live Dr Sa nt cia l o n e C B es De C pe M /Fi r S L e P G m Ho %of of Subject Subject Area Built % Built Out Out “Tall and Skinny” to EDW t1 t2 t3 t5 t4 Specialty GL/Finance Claims Contact Member Benefit Home Delivery Prescriber Drug Sales Goal State Architecture – EDW Operational Data Repository Home Delivery Systems CICS Agents Anchor WAN Home Delivery Pharmacy Web Servers CORBA Integration Services Compass Stratus • New Integration Technologies to provide a consistent mechanism for application integration and data inquiry services (ESB/SOA). • High availability in the Data Warehouse Teradata, Informatica, Business Objects, and Web environments. • Active feeds from operational systems into the warehouse via consistent integration technologies. • Inquiry services reference a single data source in many cases. • Operational Business Intelligence is enabled by loading transactions into the EDW in near real-time. Member Portal Enterprise Eligibility Enterprise Service Bus Retail Pharmacy Informatica Outbound / Express Contact Web Servers Teradata Business Objects Data Warehouse and Reporting Operational BI · · · · Member Internal Analytics Product Management Provider Relations Finance BACS Trend Central · Clients · CPMs Patient Care Advocate Client Enterprise Logical Data Model And Now… • A shameless sales pitch for a career at Express Scripts… Our Work Matters • We serve patients and clients • We align our interest with clients • We add value to help us grow • We make this a positive place to work Information Technology • We develop great people • We deliver world-class system performance & reliability • We build solid platforms and architecture • We deliver innovative business solutions • We develop solid strategic plans We are developing top talent for tomorrow’s leaders. 14 How to Apply: • Visit www.express-scripts.com • Click on Careers under “Our Company” • Click on Opportunities at Express Scripts • Select a location or “All” in the Location drop-down list • Select “Intern Opportunities” in the Job Category drop down list • Fill out the application online Watch for Internship job descriptions online after Oct. 16th Conclusions • • • • Good Roadmap Fostered Executive Buy-in Continue to Deliver on Strategy And… consider Express Scripts… BACKUP Desired Situation • Data Warehouse as an enterprise information asset. • Business driven implementation. • Re-use and leverage of intellectual property. • Phased delivery for managing risk and maximizing ROI. • Efficiency in deployment. • Extensibility for new applications and data content. • Consistent enforcement of privacy and security policies. An Analogy Imagine that you are building a city... How would you go about doing it? An Analogy First Method: Construct neighborhoods of the city as required on an as-needed (ROI justified) basis. • Each neighborhood is constructed based on the needs of its constituency and pre-existing neighborhood infrastructure. • Each constituency pays only for its neighborhood construction costs. • As the city evolves and expands, invest in appropriate re-engineering to meet new requirements. An Analogy Example Result: Boston. • Numerous one-way streets and dead ends. • Difficult to navigate. • Non-extensible without major costs. – e.g., “The Big Dig” • New neighborhoods not well integrated. An Analogy Second Method: Invest in an up-front city plan as a “blueprint” for city design. • Requires infrastructure investment prior to constructing first neighborhood. • Funded by “city tax” across all constituents. • City plan provides high level framework, not the detailed design. • Each neighborhood construction project adheres to the city plan (blueprint). • Each constituency pays for its detailed design and construction costs. • Extensibility built into the city plan. An Analogy Example Result: Toronto. • Investment and vision applied prior to actual construction. • Extremely “user friendly” city. • Extensibility of infrastructure anticipated in city design and overall cost is minimized. – e.g., Toronto subway. • New neighborhoods easily added without straining infrastructure. An Analogy The Enterprise Data Model for your Data Warehouse is like the City Plan for designing a Metropolis. Efficient and extensible deployment of a Data Warehouse requires an Enterprise Data Model. Note that disciplined adherence to blueprint is a prerequisite to success. Implementation and Deployment There is a spectrum of implementation approaches for a data warehouse – each with a different set of investment tradeoffs. 1. Independent PDM (Physical Data Model) for each domain of analytic applications (e.g., marketing, pricing, medical management, etc.). 2. Incremental deployment of enterprise PDM with reengineering appropriate to each phase of delivery. 3. Phased deployment of PDM within an enterprise LDM framework. 25 Implementation and Deployment The Enterprise Logical Data Model provides a framework for building the 80% once rather than once for each domain of analytic applications. • Estimated cost for independent multi-mart approach is 70% higher than for phased implementation using an Enterprise LDM framework. • Estimated cost for incremental deployment of enterprise PDM without Enterprise LDM is 20-30% higher due to re-engineering requirements in each phase. 26 Implementation and Deployment The blueprint provided by an Enterprise Logical Data Model allows extension of data content in the warehouse via enhancement to data domains and data relationships... ...without re-engineering the data model for each new application or data source. 27 Implementation and Deployment Avoid “big bang” approach to data modeling: • Goal is to create a conceptual data model to act as blueprint for the enterprise. • Do not get caught up in detailed sub-typing, attribution, or data typing for entities. • Complete conceptual model and opportunity analysis within four months of start date. Funding Model • High-level design of conceptual ELDM performed across business units and funded as infrastructure investment as pre-requisite to data warehouse construction. • Detailed design and population of ELDM is justified by specific ROI opportunities. • Conceptual ELDM deliverable is focused on what is necessary for extensible delivery of the data warehouse solution. • Detailed design of ELDM is delivered incrementally with each ROI opportunity. Think Big, Start Small Enterprise Paradigm Member Reporting Business Value IT Development Member (150) Department Paradigm “The Business User” Member Churn Reporting Churn Growth Churn Growth Customer Growth Customer Customer Member Billing Billing (150) Billing “The IT Provider” “The Business User” Business Value IT Development “The IT Provider” Think Big, Start Small Enterprise Paradigm Member Reporting + Medical Management Business Value IT Development Member (150) Department Paradigm “The Business User” (120) Member Churn Reporting Churn Growth Churn Growth Customer Growth Customer Customer Medical Management Member Billing Billing (150) Billing Member + Claims “The IT Provider” “The Business User” Business Value IT Development (260) “The IT Provider” Think Big, Start Small Member Reporting + Medical Management + Financial Analysis Optimized cost structure Store once, use many Single view of enterprise “The Business User” Enterprise Paradigm Business Value IT Development Member (150) Department Paradigm Inforce Churn Reporting Churn Growth Churn Growth Customer Growth Customer Customer Member Billing Billing (150) Billing (120) Medical Management “The IT Provider” (110) Financial Analysis “The Business User” Business Value IT Development Member + Claims Member + Claims + Revenue (260) (380) “The IT Provider”