Data Governance and Management
Regulated Industries Research Reference Gartner s research to illustrate the ubiquity of the data quality challenge. $100M+ $50M+ $10M+ $5M+ Don't Know Estimated annual costs for data quality issues Source: Findings From Primary Research Study: Data Quality Issues Create Significant Cost, Yet Often Go Unmeasured (G00206869)
Business Strategies Call for Revenue Growth, Reduced Costs and New Connections With Customers Business Strategies Ranking of business strategies CIOs selected as one of their top three in 2012. Ranking 2012 2011 2010 2009 2008 Increasing enterprise growth 1 1 * * * Attracting and retaining new customers 2 2 5 4 2 Reducing enterprise costs 3 3 2 2 5 Creating new products or services (innovation) 4 4 6 8 3 Delivering operational results 5 9 * * * Improving efficiency 6 8 * * * Improving profitability (margins) 7 21 * * * Attracting and retaining the workforce 8 12 4 3 6 Improving marketing and sales effectiveness 9 18 * * * Expanding into new markets and geographies 10 11 13 10 4 Improving governance, compliance, risk, security 11 10 11 12 14 Implementing finance and management controls 12 25 * * * Improving business processes 13 5 1 1 1 * Not an option that year 2
Governance Foundations Legal charter Amendments and regulations Data governance Principles Policies Standards Guidelines
Who Owns the Data?
Business Leaders are Data Stewards
IT Leaders are Data Custodians
The Goldilocks Principle of Governance Policies are too restrictive Policies are too loose
Find the Burning Bridges 9
Translate Tech Lingo into the Business Language Businesspeople and IT people each have their own perceptions but data governance is about business outcomes Tech Lingo We improved data quality by remediating CR #I8-123487 that occurred due to the contradictory data in the authorization message. The issue in fields 18.1 and 22.2 combined was uncovered by the data quality monitor and consequent steps of the DMO together with Region West Business Language $1,825,000 is the annual recovered transaction value delivered jointly by the Data Governance Program and Western Region 10
Value recovered ($,000) Build a Compelling Business Case 850 800 IQ Cumulative Value Net of Cumulative Costs 750 700 650 600 550 500 450 400 350 300 250 200 150 100 50 - (50) 0 5 10 15 20 25 30 35 No. calendar quarters since governance launch 11
Report the Business Value of the Program 12
Think Big but Start Small Establish a tactical team Choose one problem - One concept: homeowner, corporate codes - Three systems: one source, 2 targets Articulate the business s requirements Develop a business case Design the CDO program Establish controls 13
The 3-Ms : Models, Metadata, Measures Standards for Models & Architectures Standards for Data & Metadata Behaviors Statistical Process Control (6 Sigma) Standards for Analysis of Data
Build Targeted, Sparse Data Models
Critical Competencies Strong commitment to data management Business executive support seems quite strong Critical data elements are known Senior business executives care CIO is engaged Data quality is reasonably good Talented technical people in the program Program elements are properly defined
Critical Deficiencies We ve been working on this since 199n leads to early despair and the let s go to the next page of the report syndrome Progress reports contain details of peoples jobs not their business accomplishments The program design is tactical whereas it should be strategic The program is IT, not business, focused The Steering Committee and Working Groups are not focused on the right things policy and monitoring
Critical Deficiencies The program is tackling far too many data elements don t boil a lake or an ocean No business rationale or explanation for actions. Line of business managers not properly engaged - Need senior business data stewards - Need senior IT data custodians Management and personnel changes result in strategy and implementation discontinuities Decision-making authorities for data and metadata are not articulated clearly
Critical Deficiencies Metadata is NOT an executive concern Data architecture is critical and not on track No master data management program. Analytics is disconnected from the program. Too many data owners pick one per concept Improper data integration model BI infrastructure disconnected from the program Education and training program missing Data models are too complex
MDM and the Art of Motorcycle Maintenance 20 Master Data Management (MDM) is an IT process in which business takes leadership to specify, improve, and maintain current, reliable, accurate, consistent, and valid lists of the enterprise s critical information its Master Data
Complete Decision Process Flows
Foundations of Data Governance Legal Framework Knowledge Domain v. IT Responsibiliti es Strategy Roles Management Preferences Behaviors Funding Architecture Write policies and standards for master data design Chief data architect must approve new database schema Use only an approved data modeling tool Data governance council may veto project funding Business Strategy, Risks, Goals, and Priorities Infrastructure Applications & Services Establish a master data program for the enterprise Chief data officer is responsible for all data services Services must use data services platform (DSP) DBMS licenses require 5 year of amortization Develop data services platform (DSP) capability Data management organization (DMO) delivers data architectures Applications must access data using the DSP DSP costs will be allocated across all projects IT Portfolio Management Data will be treated as a strategic asset Enterprise architecture (EA) delivers conceptual data models Use software as a service (SaaS) for email servers Master data must be used for all services
Data Quality Management Understand master data and important data fields Apply Data Quality best practices Measure Data Quality
Develop a Master Data Error Alert Process Data from source Monitor raises a data error alert Is the root cause internal or external? No Work with IT on solution Yes Work with trading partners on a solution plan Monitor benefits No Partner supports a solution plan? Yes Assist the trading partner s IT group with problem resolution 25
Build a Data Management Architecture
A Data Governance Roadmap Build a vision for success raise awareness Survey data creators, consumers, and custodians Analyze selected data element and systems Learn the compelling stories repeat them often Choose a problem and correct it Learn from failures and roll-out governance 27
Data Governance and Management Build a roadmap for successful data management. - Think big, but start small. - Get buy-in from all stakeholders. - Formalize the data governance processes and teams. Find one important problem and fix it. - Assign stewardship and custodianship responsibilities. - Determine the value of the solution. Advance and improve more data and systems Keep the focus on the business while fixing data.
Recommendations Reorganize the program Assign data decision-making and accountability to line business and IT managers Adapt data governance to the business culture. Remember that behavioral change forces people to automatically erect roadblocks. Remove roadblocks to program success. Execute data governance programs using: Rationalize data sources for remediation Build data definitions conceptual, logical, and physical Create quality metrics and measure business value 29