Dambaru Jena Senior Principal Hewlett-Packard (HP)
Agenda Introduction Master Data Management (MDM) Data Governance (DG) Data Quality (DQ) Architecture & Best Practices Q&A Appendix Additional Slides MDM and Data Governance: Two sides of a coin 2
Introduction MDM and Data Governance (DG): Two Sides of a Coin MDM & DG is like Chicken or Egg Which one comes First? Should you focus on Data Governance first and then MDM or Vice Versa? You can t build MDM without proper Data Governance and You can t have a meaningful DG without focusing on Master Data Management (MDM) & Data Quality MDM and Data Governance: Two sides of a coin 3
What is master data management? Master data is the information that describes core business entities such as customers, products, locations, suppliers,. Master data typically is non-transactional data, shared by several applications, static in nature. Master data management is the required organizations, processes and tools to ensure that every master data element Is captured accurately & consistently thus enabling master data quality Is stored in a way that guarantees integrity and a single place of reference Is made available to those who need it, whenever they need it, both internally and externally. MDM and Data Governance: Two sides of a coin 4
Types of Master Data Management (MDM) Enterprise Data Transactional Data Operational MDM Master Data Analytical MDM Analytical Data Operational MDM Operational MDM describing how MDM is used to run the business (e.g., how business operations users create and manage new and old master data) Analytical MDM Analytical MDM describing how MDM is used to measure the business (e.g., how business users prepare data for use in reporting and analytics, such as BI) Enterprise MDM Combination of both Operational and Analytical MDM Enterprise MDM MDM and Data Governance: Two sides of a coin 5
What are the Benefits MDM? Single version of the truth One single reference for critical master data, across geographies and business units Business intelligence Allows true global reporting: customer volume, product volume, key customer accounts, performance of customers/products Spent analysis, economy of scale Operational efficiencies Data captured correctly and consistently according to data standards Improves business process accuracy (e.g. workflows) Cost saving in data maintenance (less duplicates, less rework, consolidation of organizations) Application integration Single version of the truth shared internally in heterogeneous system landscape Bi-directional data synchronization between operational sources and MDM Compliance Standards documented, applied and enforced Traceability of changes Role-based security MDM and Data Governance: Two sides of a coin 6
Why are MDM projects so hard? Difficulty in obtaining Executive Buy-in Cross-enterprise scope and impact causes angst Difficulty to properly scope and execute the project Heavy reliance on strong business/it partnership Extent of Organization changes required to succeed e.g. Data Governance, Data Quality, Business Process The magic mix: a business-driven initiative with IT firmly behind it; commitment from both sides is critical MDM and Data Governance: Two sides of a coin 7
How to sell MDM to Key Executives? Chief Executive Officer (CEO) Master Data Management will: Unlock the power of internal synergy within your company, across operation companies Enable true global decision making Facilitate the integration of new acquisitions Chief Information Officer (CIO) Master Data Management will: Provide the building blocks for your Services Oriented Architecture (SOA) Decrease your Total Cost of Ownership Accelerate the ROI on the ERP & BI implementations you have invested in Chief Finance Officer (CFO) Master Data Management will: Streamline the investments across the enterprise Improve your purchasing power Reduce the total budget by implementing once BU/Functional Executives Master Data Management will: Improve the data quality of your core entities leading to increased customer satisfaction and thus competitive differentiation MDM and Data Governance: Two sides of a coin 8
MDM and Data Governance: Two sides of a coin 9
What is Data Governance? Data Governance is defined as the processes, policies, standards, organization, and technologies required to manage and ensure decision rights, ownership & accountability of data in an organization MDM and Data Governance: Two sides of a coin 10
Data Governance Framework Scope Data stewardship Data quality Master data management Metadata Business rules Data security Prioritization People A core and community group with governance responsibilities. Scope Support People Business Data Business Need/Process Processes Technology Processes Data quality metrics and reporting, updating master data, ongoing monitoring and profiling, stewardship, define privacy and security for roles, conduct audits, Technology Use or deploy technology to support processes such as business quality auditing Support Provide ongoing problem resolution for issues. Education/ Training Education/Training Provide information and ensure that the processes 11 MDM and Data Governance: Two sides of a coin 11
Data Stewardship Data Governance Data Governance Organization Structure The Data and Business Intelligence organization will lead the transformation, putting the building blocks in place to enable the organization to derive value from data assets Enterprise Divisions & Programs DQ Analyst Executive Governance Council Data Sponsors (Program Steering Committee) Business Stewards Lead Data Stewards Business Stewards Subject Area Oriented Data Stewardship Teams Business Data Stewards & Other SMEs DQ Analyst Data Lead Governance Office Data Steward IT Business intelligence Master Data Management Information Management Consumer & Producers Data Governance Scope Data Quality Mgmt Data Architecture Mgmt Data Operations Mgmt Reference & Master Data Data Development Data Security Mgmt Data Warehouse & BI Meta Data Mgmt Document & Content Mgmt MDM and Data Governance: Two sides of a coin 12
Estimated Annual Cost of Data Quality (DQ) DQ is integral part of MDM and DG Solution Framework. 4% spent more than $100 millions per yer 50+% customer spends $0.5 million or higher on DQ Issues every year 35% don t know how much they spent on DQ MDM and Data Governance: Two sides of a coin 13
Data Quality Improvement/Certification Lifecycle Match & Consolidate data Profile data to assess, analyze, and report on data quality issues Standardize and cleanse the data using DQ Tools Enrich data using 3 rd part data MDM and Data Governance: Two sides of a coin 14
Dimensions of Data Quality Data quality is about more than just being correct; it is about understanding what dimensions of data quality are relevant to the business. The following dimensions can be used to gauge how much of a chosen data subject area gets cleaned Dimension Description Importance/Relevance Data Correctness Data Definition Consistency Data Completeness Data Format Referential Integrity Data Reasonableness Data Duplication Business Rule Relevance Represents the truth of data and certifies that systems are reporting consistent and accurate information Refers to a common perception of data and conveys whether one can truly and clearly understand what the data implies Alludes to the presence or absence of data in addition to the data s ability to answer relevant business questions Describes the appropriate level of detail in the data that a field is supposed to follow; i.e. data type, width, mask, etc. Describes inspections for foreign key violations, orphaned entities, broken relationships, and discontinuities Verifies whether data is a valid value and within acceptable ranges or other tolerances based on prior trends Refers to the presence of redundant data. This rule maintains a single version of data to avoid contradictory analyses Describes the relevance of domain-specific business rules, constraints and cross-field validation Incorrect data can lead to problems including improper territory alignments, difficulties in prescriber-to-script reconciliation, and erroneous customer information Businesses which have numerous definitions for the same data can cause undue confusion in its interpretation leading to contradictory reporting and analysis While complete data sets allow seamless analysis, incomplete data sets can cause difficulties if data is missing or inadequate Data which lacks proper and consistent formatting leads to discontinuities, confusion, and difficulties that need to be reconciled before the data can be analyzed If broken linkages exist between complementary systems, the systems will have difficulty making the proper connections and relationships required during reporting Without defined tolerances, businesses are unable to identify outliers in the data Multiple versions of data cause unnecessary confusion and control issues Lack of consistent business rules can lead to miscommunication between parties and erroneous data in the system MDM and Data Governance: Two sides of a coin 15
MDM and Data Governance: Two sides of a coin 16
MDM Implementation Models Three Implementation Models 1. Registry Model 2. Repository/Hub Model 3. Hybrid Model DB1 DB2 DB3 Registry Model Links the IDs from various sources to a global master record. Only stores the IDs in MDM. Keeps the data in the sources. Real Time Rationalization Key Identity Attributes Batch EDW DB4 Batch Repository/Hub Pulls the source records in real-time and creates a consolidated master record in the MDM. The Master record is then published to the Source Systems. Hybrid Combination of Registry and Hub Architecture depending on the source App1 Rationalization App1 Rationalization App2 App3 Real Time MDM Hub Batch EDW DB2 DB3 DB4 Real Time Batch MDM Hub Batch EDW MDM and Data Governance: Two sides of a coin 17
MDM Solution Conceptual Framework Stewardship process Metadata Distribution request New/changed master data request Quality Improvement Initiative Operations & quality monitoring Meta data shopping Impact analysis Operations monitoring Quality investigations Master data services MDM governance & stewardship Acquisition & authoring Real-time/near Batch Change capture Metadata services Suppliers Suppliers Authoritative sources End-user authoring Master data management services Master data quality services Administration services Distribution Services Acquisition & Authoring Services Consumers Maintenance services Distribution Master data quality Validation Audit, balance & control Quality tracking Historical data management Storage services Schema services Mapping/alignment services Hierarchy management Self-service (pull) interface Publishing (push) interface Messaging-oriented interface Administration and maintenance Change management Security management Operational support Process monitoring Performance management Stewardship workflow MDM and Data Governance: Two sides of a coin 18
ETL / EAI Brand Manager Brand Category Product Family Product Name Product Saleable Product Packed Product Product Classification Class Group Sub Group Parent Company Customer Delivery Point (Outlet) Location Type Customer Nace Classification NACE Section NACE Group NACE Class MDM Technical Reference Architecture Enterprise Master Data Management / Service Oriented Architecture (SOA) Source systems Web-based master data management Business modeling Harmonized master data BP-Shell Sales Volume Proceeds Distribution cost Manufacture cost ODS/ ERP Transaction master data Metadata EDW SFDC CRM Master Data Management Hubs Portals D&B Metadata Reference data Business definitions Incomplete, invalid, unauthorized data Region Country District Manager Target Cust Partners Account Grp Account Contract Geo Address Del Point Industry Sector Size!! Target Enrichment Region Country District Manager Target Cust Partners Account Grp Contract Account Geo Address Del Point Industry Sector Target Size Complete, enriched, valid, authorized data User community Information Governance & Data Stewardship Data Model / Metadata Architecture MDM and Data Governance: Two sides of a coin 19
MDM and Data Governance: Two sides of a coin 20
Best Practices - Implementation approach Vision & Strategy: where are we going? Implementation: how do we get there? Maintenance & Support: how do we sustain success? Assessment & Roadmap Project Lifecycle Phase N Business enablement Project Lifecycle Maintenance & support Phase 1 Information management Project Lifecycle Phase 0 How do we improve and continually realize business value? Start with vision, roadmap and plan Pragmatically adopt rapid, iterative implementation Build effective maintenance and support MDM and Data Governance: Two sides of a coin 21
MDM - Key Success Factors Business support Provide strong sponsorship to overcome roadblocks Secure IT & business investment (funding, people, systems) Enforce policies to leverage master data as a corporate asset Data Governance Establish Data Governance Organization early Establish ownership from enterprise and business line perspectives Establish clear stewardship roles and responsibilities Data Quality Document Data Quality metrics and processes to resolve DQ Issues Focus on Data Quality Issues as a Part of MDM Implementation Right Architecture Select solution that is based on budget, timeline, business objectives Choose the right vendors and tools MDM and Data Governance: Two sides of a coin 22
Data Governance: Key Success Factor Clarity of purpose Build governance to manage, guide and oversee the BI/MDM Initiatives Insist on active support by C-level executives Clearly defined goals and objectives Focus on Business Needs/ Challenges Think Globally; Act Locally Design governance models and processes for the entire organization. Build and implement against high-impact areas. Governance is more than standards, reporting and prioritization of projects. It is a business function that provides structure by which complex technology programs can be managed to successful completion. Flexibility Communication Recognize that one size does not fit all when it comes to governance. Start small, make it customizable (within guidelines) and people will get a sense of ownership Governance is not static; it must evolve over time Frequent, directed communication will reduce anxiety and provide a mechanism for gauging when to correct the course Consider a change management process to ease transition MDM and Data Governance: Two sides of a coin 23
MDM Gartner Magic Quadrant Key Evaluation Criteria: Ability to Execute Product/Service Overall Viability Sales Execution/Pricing Customer Experience Marketing Execution Completeness of Vision Market Understanding Market Synergy Offering Strategy Vertical/Industry Strategy Sales Strategy Innovation Source: Gartner (October 2010) MDM and Data Governance: Two sides of a coin 24
Contact Information: Dambaru Jena Cell: 630-699-3109 Dambaru.Jena@hp.com MDM and Data Governance: Two sides of a coin 25
MDM and Data Governance: Two sides of a coin 26
Multi-Domain Hub - Oracle Oracle Data Hubs 1. Customer Data Hub 2. Product Data Hub 3. Supplier Data Hub 4. Site Data Hub Oracle's Hub Architecture and Technical Capabilities Presentation Identifier Goes Here 27
360º View of the Customer Contact Vendor Agency Customer Information Customer Party Product Product Detail MDM and Data Governance: Two sides of a coin 28
MDM Key Best Practices 1. Start with the end in mind 2. Scope and frame the project realistically 3. Include Data Quality and metadata standards as part of the discussion 4. Recognize Governance and Data stewardship is an MDM must 5. Implementation must show ongoing value of the project no big-bang approach Use Iterative Implementation Approach Accept that MDM is an evolutionary process MDM and Data Governance: Two sides of a coin 29
Key MDM Trends Market Maturity The MDM market will continue to shift gears from early adopter to mainstream as 95%+ of financial services, communications services, high tech & pharmaceutical/life sciences enterprises actively explore to replace homegrown MDM solutions. Market Consolidation Through 2009-10, mega IT vendors (IBM, ORCL, SAP, & TDC) will dominate the MDM market with niche/best-of-breed vendors (DNB/Purisma, DataFlux, Tibco ) thriving in specific industries & horizontal/corporate applications Data Quality : Data quality is an inherent and important part of any MDM project. Other Drivers Data Governance : The need for cross-enterprise data governance that includes entire master data Business Intelligence : The convergence of MDM & BI is set to accelerate as enterprises leverage MDM concepts in a BI context MDM and Data Governance: Two sides of a coin 30