Presentation on theme: "Successfully Deploying an Off-the-Shelf Data Warehouse William Goedicke Copyright William Goedicke,"— Presentation transcript:
Successfully Deploying an Off-the-Shelf Data Warehouse William Goedicke http://www.goedsole.com:8080 firstname.lastname@example.org Copyright William Goedicke, 2003. This work is the intellectual property of the author. Permission is granted for this material to be shared for non-commercial, educational purposes, provided that this copyright appears on the reproduced materials and notice is given that the copying is by permission of the author. To disseminate otherwise or to republish requires written permission from the author.
Agenda A New Kind of Data Warehouse Differences between Analytical and Transactional Systems An Implementation Project Plan
Data Warehousing Benefits Single Source with Shared Metadata for Consistent Analysis across Constituencies Consolidate Historical Transactions Increase Productivity of Analytical Work and therefore Overall Performance
Data Warehousing Problems and Their Causes Intractable Data Conversion Issues Massive Project Overruns Why? – Architecture Complexities – Modeling Complexities – Sophisticated Subject Matter Expertise Requirements
Historical Framework Personal Computer Local Area Network Relational Database ERP Off the shelf Data Warehouses
Off-the-Shelf Data Warehouses Delivered by ERP Vendor Well Understood Data Model Inclusive of All Critical Transactions An Architecture (CIF) Denormalized and Star Schema Data Models Data Mart Cubes Integrated, Pre-Configured Analytical Tools Less Opportunity for “Analysis Paralysis”
Analytics vs. Transactions (Shifting Gears) Transactional Systems Analytical Systems # of UsersLotsFew User’s RankAll OverHigh Process Rigidity HighNone
Transactional System Project Plan Acquire Executive Sponsor Define Scope Prototype Basic Setup Fit / Gap Detailed Setup Customization Conversion User Acceptance Test Training Go Live
An Alternate Plan for Analytical Systems StepDelta Identify Executive Sponsor As Important as Ever You Need ALL the Data Develop ScopeBased on Strategic Goals not Process
An Alternate Plan for Analytical Systems PrototypeNo Product Selection, Demonstration is Later Basic SetupIt’s not Basic Fit / GapNA
An Alternate Plan for Analytical Systems Detailed SetupThat’s Done CustomizationIs Both Avoidable and Essential ConversionDramatically Diminished
An Alternate Plan for Analytical Systems User Acceptance Test includes Training There is no Go Live Switch. We’re Analyzing; We Are Live!
Yikes! Our Dirty Little Secret – No User Involvement During Initial Deployment The Answers Have Changed and People Look Foolish in Important Ways New Facts are Discovered; Some are not Complimentary
The Brandeis Example IT Mgmt/CFO Sponsorship Form Tiny Stealth Team Research Institutions's Strategic Goals Pick 1 or 2 Low-Hanging Fruit Develop Modest Working Prototype Present to Executive Acquire Political Capital Enlist Middle Mgm't to Operationalize Prototype Disperse Ownership Evangelize
Conclusions ERP Implementations Allow for Data Warehousing “Products” in Lieu of Projects. This Dramatically Reduces the Skill Requirements. Unbounded Requirements and Small, Highly Skilled Customer Base Allows for Streamlined Implementation Project.
Questions? (and web site plug) William Goedicke http://www.goedsole.com:8080 email@example.com