Reliable Business Data Implementing A Successful Data Governance Strategy with Enterprise Modeling Standards All Rights Reserved
Welcome! Let Me Introduce Myself Marcie Barkin Goodwin President & CEO Axis Software Designs mbg@axisboulder.com www.axisboulder.com All Rights Reserved page 2
A Brief Bio of Marcie Axis Software Designs a consulting and product development company specializing in enterprise modeling standards, procedures and best practices. Senior Enterprise Architect, providing modeling infrastructure consulting services and data modeling training to private industry and government agencies for over 25 years. Designer of EM-SOS! Enterprise Modeling Set of Standards Partner Sandhill Consultants All Rights Reserved page 3
Agenda The Big Picture Data Governance What Is It & Why Do We Need It? The Cost of Unreliable Business Data Data Governance in Modeling Best Practices DAMA DMBOK/DMM Perspective A Strategy for Reliability Modeling Best Practices Properties of Reliable Data Standardized Data Design Practices & Oversight Communication & Metadata Integrity (Who Can You Trust?) Reusability (Saving Time & Money) Reduce Redundancy (Say Again?) Reliable Business Data in Action Best Practices - The Essentials Strategy, Standards and Procedures Implementing a Standards & Best Practices Framework We Know It s Important, But Who Has Time? All Rights Reserved page 5
The Big Picture Data Governance In Modeling It s astonishing, in this world, how things don t turn out at all the way you expect them to. Agatha Christie As in life, as in our IT & Modeling environments Enter Governance. IT Governance, including Data Governance, is a philosophy of accountability, substantiated by a structured plan to guide IT processes towards accurate and accessible information. All Rights Reserved page 6
Data Governance What Is It? Data Governance is the formal orchestration of people, processes, and technology to enable an organization to leverage data as an enterprise asset. MDM Institute Information Governance is the specification of decision rights and an accountability framework to encourage desirable behavior in the valuation, creation, storage, use, archival and deletion of information. It includes the processes, roles, standards and metrics that ensure the effective and efficient use of information in enabling an organization to achieve its goals. Gartner All Rights Reserved page 7
Data Governance Why Do We Need It? Because generally speaking, the management and quality of our data is sorely lacking. The costs of managing disparate IT assets is expensive and the work efforts are cumbersome Data processes aren t consistent and standardized, thus creating different report results Regulatory compliance - Sarbanes-Oxley, etc.: Where management has outsourced functions to third party service provider(s), management maintains a responsibility to assess the controls over the outsourced operations. AND On average, American companies believe that 25% of their data is inaccurate. Survey: Experian Quality Data All Rights Reserved page 8
Unreliable Business Data Rarely do organizations design from an enterprise perspective, and as a result they sub optimize solutions by allowing diverse standards, naming conventions, codes, etc. This silo approach to data and modeling efforts results in data inconsistencies which have a deleterious effect on the overall understanding, accessibility, integration and reliability, or integrity of the data. And other problems exist as well All Rights Reserved page 9
Unreliable Business Data (Cont.) Tendency to clone multiple, non-standardized versions of data structures resulting in: Unreliable answers from data analysis Variable data structures that are non-reusable, inconsistent, inflexible and expensive to fix Lack of enforcement exists for: Naming standards and class words Business dictionaries & abbreviations Overruns of Application Development schedules and budgets Lack of Business to Data Traceability Difficulties with Data Integration & Interoperability All Rights Reserved page 10
The Cost of Unreliable Business Data Regulatory Public Health & Safety Sept 9, 2010 San Bruno, CA Economic Disaster The absence of well documented standards, common business term definitions, and an agreed-to format for exchange makes it almost impossible for parties to understand each other. All Rights Reserved page 11
Do These Issues Look Familiar? Same object name has different definitions and usage Expensive conflicts in processes Inaccurate and risky decision making Siloed metadata (turf wars) Unnecessary duplication and cost due to ambiguity in meaning and usage Same object definition applied to different names Higher rates of redundancy and data errors Inconsistent meaning across lines of business Reduced value of BI Reduced opportunity for re-use Increased cost and time to market, decreased quality and value Change and growth management difficult to impossible Lost opportunity, increased cost of doing business All Rights Reserved page 12
Data Governance in Modeling Best Practices Data modeling Best Practices support the objectives of data governance and produce reliable business data. Develop Roles and Responsibilities including data custodians Develop standards for model and model object naming Develop standardized procedures regarding how models are to be created, organized, stored, changed, archived, backed up, versioned and migrated through a model development life cycle Develop audit procedures that ensure ongoing compliance with standards, business rules and/or government regulations A Best Practices strategy prevents the risk of creating incorrect data which may result in critical harm in the real world. All Rights Reserved page 13
Blueprints - DAMA DMBOK & CMMI DMM The DAMA-DMBOK2 Guide Knowledge Area Wheel Data Management Maturity Model All Rights Reserved page 14
Properties of Reliable Data For data to be reliable, it must be consistent and usable across all time, space and function of the owning organization. It can be characterized by some or all of the following properties: Predictable Measureable Enforceable Manageable Portable Reusable Consistency of: Meaning Usage Form (Datatype, etc.) Interface Results All Rights Reserved page 15
Standardized Data Design Practices & Oversight A modeling strategy which incorporates an enterprise perspective is key to ensuring the reliability of business data. Strategies are determined by clarifying where you are, and determining where you want to be Strategies are not one size fits all, but they do share the critical goals of: Communication the all-important foundation Metadata Integrity The naked truth is always better than the best-dressed lie. Ann Landers Reusability - The reuse of model objects, brings consistency to model structures Non-Redundancy - Redundancy is one of the most common causes of inefficiency & data inaccuracy. All Rights Reserved page 16
Reliable Business Data in Action When an organization manages its data as a corporate asset, and makes modeling best practices a part of its data governance efforts, the focus turns towards the business value of the data, not just the IT requirements. The business drivers for managing data as a corporate asset are: Competitiveness Regulatory compliance Realizing the benefit of investments in BI and other information-based initiatives Risk mitigation All Rights Reserved page 17
Reliable Business Data in Action (Cont.) The keys to successful best practices modeling implementation are: Management support of a data governance program, including an integral best practices approach to modeling Development of a modeling strategy that builds governance and quality into the process Involvement by both the business side as well as IT Resistance to creating or accepting silos of standards and procedures Development of a Model Management Framework to support the modeling strategy All Rights Reserved page 18
A Modeling Framework Must support: Strategic goals & policies of the organization Standards which are extensible across the enterprise Processes & Data Modeling Life Cycle to ensure repeatable, efficient data design and outcomes Data Governance Roles & Responsibilities Must include: Audit documentation for changes in processes & content Metrics Reports to measure improvement in the data design process (reuse, decreased redundancy) Must be: Easily accessible, maintained & enforced Multi-user environment with associated permissions All Rights Reserved page 19
The Challenges We know we need it, but who has the time? Organizations generally understand the need for modeling standards and procedures, but face several obstacles in getting the job done: There isn t a plan for developing a Framework In-house resources are already working full time Resources may not have infrastructure knowledge or writing skills Internal politics or silo groups can t agree on what should be implemented All Rights Reserved page 20
A Modeling Framework Buy vs Build? DMBOK/DMM A great blueprint of what is needed The details need to be defined Buy vs Build Time Money Expertise (quality of the result) Consistency across the organization Acceptance and buy-in Business IT All Rights Reserved page 21
Critical Framework Content List Standards Classword Domain Datatype List Color Themes Data Object Definitions Guidelines Library Structure & Naming Model Repository Profiles Physical Model Object Naming User Defined Properties Procedures Administrative Change Control Data Object Promotion Issue Resolution Logical DM Submission & Deliverables Model Backup Model Change Control Model Creation & Population New User Model Repository Notification & Update Physical DM Submission & Deliverables Stewardship Synchronization Versioning All Rights Reserved page 23
Critical Framework Content List Blueprints for Enterprise Model Development Building an Enterprise Data Model Data Model Life Cycle DMLC Top Level DMLC Detailed View Shared Data Object Creation Agile (Iterative) Data Architecture -Development Integration DMLC Definitions Model Quality Reports Basic Model Metrics & Reuse Glossaries Glossary Templates Template Usage Forms Administrative Change Control Form Logical Data Model Check List Model Change Control Form New User Repository Profile Form Physical Data Model Check List All Rights Reserved page 24
Implementing a Standards & Best Practices Framework Key Steps Understanding Where you are Gap Analysis Assessment Using Assessment to define a strategy Compare to known Best Practices guidance & solutions The Decision: Buy vs Build Collaborate on a Tactical Plan Engage the stakeholders Integrate with overarching Data Governance & Management initiatives Develop a Proof of Concept Prototype/Pilot Deploy, Check & Adjust (If necessary) Communicate, Educate & Align throughout the Implementation (especially between Business & IT) All Rights Reserved page 25
Benefits of A Buy Solution Supports Data Governance principles & requirements Supports any Data Model Life Cycle methodology (Waterfall, Agile, etc.) Solves the problem of requiring in-house resources to analyze & write necessary standards documentation Compatible with the company s data modeling solution Implemented on the company s intranet for accessibility All Rights Reserved page 26
Standard Classwords implemented using ERwin Domains with inheritance All Rights Reserved page 29
DMM Framework Mapping B u s i n e s s V a l u e 1 Performed Ad Hoc localized Standards 2 Managed Top Level Data Management Life Cycle (DMLC) Naming Standards Automated Templates Standards Training 3 Defined Detailed cross functional DMLC Standard processes and milestones Standardized health checks Maturity 4 Measured Product and process metrics Change management and process improvement based on metrics 5 Optimized Adaptive DMLC (Agile) Global deployment of EM-SOS! Data Object Collections and Repository Global Shared Data Objects and Reuse All Rights Reserved page 30
Modeling Environment Assessment Axis offers a comprehensive assessment incorporating a strategy which centers on the reuse and nonredundancy of data structures within an organization. This short term consulting engagement analyzes the customer s current modeling environment and delivers a Findings and Recommendations document delivering analysis and specific recommendations for streamlining and improving the modeling infrastructure. You can start by conducting a Standards Self Check at: http://www.sandhillconsultants.com/emsos.asp All Rights Reserved page 31
Contact Information Mbg@axisboulder.com 303-981-6525 For information, please contact: Robert Lutton Robert.Lutton@SandhillConsultants.com 416-219-4504 All Rights Reserved page 32