The Long and Winding Road to Data Governance

Document Sample
The Long and Winding Road to Data Governance Powered By Docstoc
					The Long and Winding Road to Data Governance

Thomas V. Carlock
Tuesday, January 08, 2008
Agenda

ü Introduction – About me
ü Company X - Case Study Facts
ü The Road to CDI Using Data Governance Light
ü The Road to Full Data Governance
ü What’s Working
ü Lessons Learned
ü Q&A




                                                2
Company X - Case Study Facts

 We are the world’s         Leadership in Key Markets             Key Facts
  largest independent
  commercial finance                                             Market Capitalization           $8.6B
  company, leveraging:      § Equipment Lending and Leasing
                                                                 Percentage of Non-US Business   21%
 § a broad set of capital   § Vendor/Dealer Finance
   and advisory solutions                                        Countries with Company X Customers
 § a century of             § Asset Based Lending                                                52
   commitment to our
   customers’ success;      § Credit Protection/Factoring        Employees                       7,000
   and
                                                                 Clients Worldwide               650,000
                            § Small Business Administration
 § a deep understanding
   of the industries we      Lending                             Industries Served               30
   serve
                            § Aerospace and Rail Asset Leasing   Offices                         220

                                                                 Assets Under Management         $70B
                            § Broker Originated Home Equity
                             Financing

                            § Student Lending




                                                                                                         3
The Road To CDI Using Data Governance Light

   Source US
    Source US
                    • Like many large global companies, Company X data is distributed
  Sales System
   Sales System
                    across multiple transaction processing systems
 Rcvble Source
 Rcvble Source      • Company X understands the value and importance of getting our
   Source UK
    Source UK       hands around the consolidated data
 Source Central
  Source Central
                    • Company X recognizes the importance around having guidelines for
 Source Asia (2)
  Source Asia (2)   visibility, sharing, privacy rules and usage of this data
      LALS
       LALS
                    • Variety Of Data Models
    Australia
     Australia
                    • Duplication Of Customer And Address Information
    EDLG
     EDLG
   Warehouse
   Warehouse
                            Who are my customers?
   M&A Data
   M&A Data
                         What have they financed with us?
   Source US
                          What services can I offer them?
    Source US

    Loan IQ
     Loan IQ                Are we at risk anywhere?
     Stucky
      Stucky

   Data Scan
   Data Scan
                    • Company X Business & IT Sponsored the CDI Project

                                                                                        4
The Road To CDI Using Data Governance Light
                     Credit   Finance
                                          Developed a small data governance task force of senior business SME’s
                   Sales         Legal
  Source US
   Source US

 Sales System
  Sales System                            Developed a strategy to achieve CDI and some data governance
Rcvble Source
Rcvble Source

  Source UK
   Source UK         Data Governance      Chose a few key Data Governance processes required to be successful
                     Task Force
Source Central
 Source Central

Source Asia (2)       Data Definitions,          Data             Data             Data             Data
 Source Asia (2)                                                                                                 Oracle
                       Standards &            Profiling &      Transform       Consolidation     Provisioning
    LALS
     LALS                 Metrics            Stewardship      & Cleansing                         & Sharing     Customer
                                                                                                                Data Hub
   Australia
    Australia
   EDLG
    EDLG            Data Governance Processes
  Warehouse
  Warehouse
                                          Goals
                                          Data consolidated in a single physical database
 External Data
  External Data
   i.e. D&B                                § Single consistent enterprise-wide customer data model
    i.e. D&B
                                           § Consolidated & standardized definitions of a customer

                                          Consistent Data Management Pipeline
  Source US
   Source US
                                           § Data quality and standards adhering to your company’s rules
   Loan IQ
    Loan IQ                                § Data can be enriched with external sources
    Stucky
     Stucky
                                          The Customer Data Hub becomes the master data source for Company X
  Data Scan
  Data Scan
                                          Customer information

                                                                                                                    5
Data Definition Process & Benefits




       Benefits
       Benefits
       ••Clear business definitions with rules defined by business people
         Clear business definitions with rules defined by business people
        §§ Easier to communicate corporate needs to local business teams
            Easier to communicate corporate needs to local business teams
        §§ Easier to map corporate definitions to local business system fields
            Easier to map corporate definitions to local business system fields
        §§   Easier to profile to understand potential data quality issues
              Easier to profile to understand potential data quality issues




                                                                                  6
Data Profiling Process & Benefits

                   1
            We begin with locating the data
            values we want to use.




                              Profile the Data
    Define Data
                               Using a Data
      Values
                               Profiling Tool




                                          2
                           Data profiling gives us complete
                           visibility into the quality and content
                           of any data source. Through
                           effective data profiling, we can
                           uncover data anomalies by
                           inspecting the true content, structure
                           and relationship hidden within
                           enterprise data sources.




                                                                     7
Example: Data Profiling Process & Benefits

                    1                                                 3                                                      5
             We begin with locating the data                      Once we have                                         Based on the business
             values we want to use.                               complete the reviewing the                           feedback we will establish data
                                                             data we write a report of our                       quality targets. Identifying root causes
                                                             findings to be reviewed with the                    of data problems and work with the
                                                             data owners of that source.                         business to seek resolution. Signing off
                                                                                                                 will authorize us to go ahead to place it
                                                                                                                 into the Corporate Data Warehouse


    Benefits
     Benefits
    ••A well defined business process ensured consistent steps to achieve profiling tasks
                        business process ensured consistent steps to achieve profiling tasks
       A well defined Profile the Data
     Define Data                                       Confirm
    ••The defined process aallowed us to educate localResults
       The               Using Data
                                        Report Data      business teams on profiling
                                                                   Set Quality
       Values defined process allowed us to educate local business teams on profilingData
                                          Findings                             Monitor
                                Profiling Tool                                                                 Targets
     § Business understood why profiling was important
        Business understood why profiling was important
         §
     § Business understood what was expected of them during this process
      § Business understood what was expected of them during this process
     § Business valued the results, embraced the process and is now actively involved
        Business valued the results, embraced the process and is now actively involved
         §
                              2
     § Data Stewardship is underway!
      § Data Stewardship is underway!                         4                                                                                 6
                                                                                        We have to ensure our                                Results:
                                                                                        findings are accurate so we                          - Consolidated Data
                            Data profiling gives us complete
                                                                                                                                             - Enterprise Definitions
                            visibility into the quality and content                     meet with the business data
                                                                                                                                             - Quality Metrics
                            of any data source. Through                                owners to review the data
                                                                                    anomalies we have found. Together                        - Data Catalogue
                            effective data profiling, we can
                                                                                    we will flush out real data quality                      - Effective Business
                            uncover data anomalies by
                                                                                    problems and determine the best                          Process
                            inspecting the true content, structure
                            and relationship hidden within                          approach to resolving the data issues.
                            enterprise data sources.




                                                                                                                                                               8
Company X’s Data Management Pipeline

                                              Source US
                                               Source US
                                                                Repeatable-Sustainable-Costs Effective                                             Data
                                             Sales System
                                              Sales System
                                                                                                                                                Warehouse
                                                                                                                                                 Or (ODS)
                                            Rcvble Source                                                                      s                Operational
                                                                                                                            ard
                                            Rcvble Source
                                                                                                         on ion
                                                                                                      ati                 nd                    Data Store
                                              Source UK
                                               Source UK                                           lid rmat            Sta
                                                                                                so                  &
                                                                                             on nsfo            lity t
Line of Business && External Data Sources




                                            Source Central
                                                                                            C ra
 Line of Business External Data Sources




                                             Source Central
                                                                                          –             es    ua n
                                            Source Asia (2)                             1    – T Rul ta Q me lity
                                             Source Asia (2)                               2     –        a   ch bi
                                                                                              3      – D Enri usa
                                                                                                  4      – Re
                                                                                                      5     –
                                                                            Customers                     6
                                            Local Warehouse
                                             Local Warehouse                                                                 Oracle
                                                                                                                            Customer      • 3rd Party Sources
                                                                   Data
                                                               Management
                                                                                                                              Data        • Analysis
                                                                            Contracts
                                                                 Pipeline                                                     Hub         • Analytics
                                                                                                                                          • Campaigns
                                              PeopleSoft
                                               PeopleSoft
                                                                                                                                          • Reporting
                                                                                                                            Financial,    • Visible Parent Child
                                                                             Assets                                        Sales, Other   Relationships
                                                                                                                               Data
                                              M&A Data
                                              M&A Data
                                                                                                                                                Downstream
                                            Prospect Files
                                             Prospect Files     (ETL) Extract, Transform & Load Layer                                           Applications




                                                                                                                                                          9
Data Consolidation Process & Benefits


                                                      3 Companies are
    Jones Inc.
       Jax
                      Bob Jones
                        Tempe
                                      The Jones
                                       Canada
                                                      actually a single
      $100K             $100K           $150K
                                                      customer!



                 Jones Inc.       Smith Co.       Gorman Co.   Gupta Inc.   Chen & Co.
                   $350K           $100K            $200K       $300K         $150K




                                                                                         10
Data Consolidation Process & Benefits
D&B Parent Child Trees + Company X Customer Data
                                                                                        Some members of the tree
                                                                    SGG Inc.                          Global and
                                                                                        are already customers,Ultimate: highest
                                                                                                      ranking global member of a
                                                                                        other members we’d like to
                                          Party 1                                                     corporate family
                                                                                        have as customers

                                                                                                    Domestic Ultimate: highest
                                                                                                    ranking member of a corporate
                                                                                                    family within a country
           Party 2 (UK)                Party 3 (US)                  Party 4 (Canada)               Parent: any company that
                                                                                                    owns more than 50% of
                                                                                                    another company
                                                                                                    Subsidiary: more than 50%
                                                                                                    owned by a Parent
 Party 5          Party 6                 Party 7                    Party 8        Party 9
                                                                                   Party 19         Headquarters: any company
                                                                                                    having Branches or Divisions
                                                                                                    Branch: secondary location or
                                                                                                    operation, not a separate legal
                            Party 10                     Party 11                                   entity




               Party 12        Party 13               Party 14         Party 15    Jones Inc.


                                                                                                                             11
Data Consolidation Process & Benefits

The single customer is
            Benefits
             Benefits
                                                                     New
                                    SGG Inc.                        Growth
                larger
a member of aThe CDI Project is delivering VALUE!
               The CDI Project is delivering VALUE!
                                   $1,100m                        Opportunities
organization!                       Corporate

                  Value
                  Value
                                   Headquarters


                  Assures
                  Assures
                  Love
                  Love
                  Understanding
                  Understanding
      20 more
                  Executives Gorman Co. Gupta Inc.
                     Smith Co.
                  Executives
                      $100K        $200K        $300K               50 more
    Branches of
     Smith Co.    $ Continued Funding J
                  $ (Subsidiary)
                      Continued Funding J
                                 (Subsidiary) (Subsidiary)         Branches
                                                                  of Gupta Inc.



    10 more          Jones Inc.                       Chen & Co      3 more
   Subsidiaries        $350K                            $150K       Branches
    buy $20k          (Branch)                         (Branch)     buy $30k


                                                                                  12
The Road To CDI Using Data Governance Light

  To Summarize
  • Developed Small Governance Task Force Using Business SME’s

  • Developed a strategy to achieve CDI and some data governance

  • Chose a few key Data Governance processes required to be successful
   §   Data Definition, Profiling & Stewardship

  • Implemented the CDI project and delivered value in 9 months

  • Results: Opportunity to make a case for a full data governance program




                                                                         13
The Road To Full Governance

   ü Business Sponsors & Data Champions
   ü Selling the Program Strategy to Senior Management
   ü Selecting a Consulting Partner
   ü Understanding Your Companies Readiness Level
   ü Building Out of the Program
      ü Build the Organization
      ü Develop the Program Guide
      ü Develop the Processes
      ü Educate and Communicate
   ü What Worked
   ü Lessons Learned



                                                         14
The Road To Full Governance
Business Sponsors & Data Champions


           Identify an individual within Company X who has the delegated authority of
           the CEO to make things happen.
            §   Corporate Controller, VP of Finance, Head of Compliance, CIO (not alone)

           Identify key people who have the knowledge, expertise and passion to
           become Data Champions for the program. (Culture Changing Agents)
            §   Requires a dedicated Business, Data Governance and IT leads.
            §   They should report directly to the Sponsor(s)
            §   Dedicated, accountable and hands on visionaries and executors of the program
            §   Ensure the program is constantly visible
                 § Volunteer to do presentations at all levels (IT, Mid Level & Senior Mgt., etc…)

                 §   Share information (portals, internal newsletter, blogs, etc…)




                                                                                               15
The Road To Full Governance
Selling the Program Strategy to Senior Management

                                                                              4 Projects totaling $20.4 million
                                                                               4 Projects totaling $20.4 million
 Europe          US East         US West


                             ms
                            aams                                 Project 4
                          Te
                        ss Te
 Canada
                      es
                   inines
                                                        Corporate Dashboard
                                                                                                $6.7 million
                                                                                                 $6.7 million
                 us
                BBus
                                                Project 3
                         Business Unit 1, 3 & 5 Data Warehouse


  Latin
                            Project 2                                                 $10.2 million
                                                                                       $10.2 million
  America
                               Credit Hub




            Project 1                                                $2.1 million
               Tax Hub                                                $2.1 million


                                      $1.4 million
                                       $1.4 million
                                                                                               Technology Teams   16
The Road To Full Governance
Selling the Program Strategy to Senior Management (cont.)

  No Enterprise Data Management Strategy
   §   No Data Governance

  Inaccurate Data
                                                                                What’s wrong with
                                                                                What’s wrong with
   §   Data duplication
                                                                                doing it this way?
                                                                                doing it this way?
   §   Limited or no data quality control

  Data Infrastructure Build Out Would Be Too Expensive
   §   More expensive to build, maintain and extend

  High cost to business units in productivity losses
   §   Large amounts of time spent collecting data for manual integrated view
   §   Large amounts of time spent reconciling disparate results and individual BU data with Corporate Finance

  Business decisions made on incomplete view
   §   Understanding of your company’s position in industry or portfolio incomplete
   §   Time from business event to actionable information not optimum
   §   Wait time for critical IT projects lengthened and costs increased


                                                                                                           17
The Road To Full Governance
Selling the Program Strategy to Senior Management (cont.)

    New Approach: Treat Data as a Corporate Asset
     §   Build comprehensive data management platform and process
     §   Re-work data management component of in flight projects
     §   Build reporting on top of best practice infrastructure
     §   Establish a data governance organization
     §   Establish a dedicated data management group
     §   Socialize data management throughout the corporation

    Advantages of comprehensive approach
     §   Less expensive to build
     §   Less expensive to operate
     §   Less expensive to extend




                                                                    18
The Road To Full Governance
Selecting a Consulting Partner
      Why choose a consulting partner?
       §   Too difficult to go it alone
       §   Requires highly qualified professionals who bring context to the
           engagement
       §   Ensures a 3rd party opinion
       §   No need to reinvent the wheel

      How do you choose the right partner?
       Large Consulting Organizations          vs.      Small Consulting Organizations




                                                                                    19
The Road To Full Governance
Understanding Your Maturity Level or Readiness
 3rd Party Company interviews verify that Company X wants data governance
    ü “Break the silo approach. Get everyone on board to learn to share data.”
    ü “Can't get back to the true source of the data; can only go as far back as the UDB. Legacy systems lack data quality and
      documentation.”             Diverse Opinions
    ü “I am tired of having to go to 14 different people to get information. It’s frustrating!”
    ü “We need a single source of the truth (EDW). Need common reference data group.
    ü “Would love to see standardized definitions where upon data can be pulled from one conformed data set.”
    ü “Create separate Data Governance department. It must be business driven, independent of systems management.”

 Company X data governance maturity level is low across 11 data governance disciplines
    ü Fix: Build a Data Management Program Roadmap

 Company X lacks methodology to support data governance projects
    ü Fix: Align the Program with the IT Governance/PMO methodology

 Company X is project focused
    ü Our culture is to recreate the wheel
        ü “Tons of manual effort and reentry of data.”
        ü “Frustrated with lack of data management capabilities and the many ad hoc data requests. Frustrated with the reporting
           capabilities. Excessive lists of reports generated, many of which contain the same data elements. No horizontal
           collaboration and no rules for accessing data.”
    ü Fix: Change the culture from Project to Program through socialization, communication, center of excellences and data
      stewardship




                                                                                                                             20
The Road To Full Governance
The Data Governance Organization


   Data Management Program – DMP                                            The Program
        Data Governance Executive Sponsorship – DGES                          The Sponsors
           Data Governance Steering Committee – DGSC                                    The Data
                                                                                The Program Decision Rights Team
                 Data Governance Council – DGC & Sub Committees                    The Data Decision Rights Team
                    Data Management Team – DMT
                                                                                     Governance
                                                                                    The Administration Team

                                                                                    Organization
                       Business Team                      IT Team

                     Data Quality                   Corporate Data
                                                    Warehouse design,
                     Data Stewardship               build & operations      Corporate
                     Data profiling                 Data Profiling            Data
                                                                                                                Project 1
                     Definitions, policies,         Extract, transform, &
                                                                            Warehouse                            Project 1
 Local Data
   Local Data        & standards                    load data
  Stewards
    Stewards         Master data & meta                                                                         Project 2
                                                    Data error capture                                           Project 2
                     data management
                          Internal                 Integration
                                              Data Data modeling
                     Acquisition
  Corporate
                           Data                      and
                                                   Data provisioning
                                                                                           Meta
                                                                                                                Project 3
                                                                                                                 Project 3
   Corporate         Quality assurance
Communication
 Communication       Data Sharing              Data Delivery
                                                   Data marts
                                                                                           Data
                     Communication                                                                         Project
                        External                                                                             Project
                                                                                                           Steering
  Customer
                            Data
                     Facilitate Council &                                                                   Steering
   Customer          Committee meetings                                                                   Committee
   Service                                                                       Data                      Committee
    Service         Operational
                     Program Support                                             Marts
                        Data

                                                                                                                       21
The Road To Full Governance
Data Management Program Guide
                                Data Governance Program Guide
                                • Executive Summary
                                • Business Drivers
                                 §   Vision and Mission Statements
                                 §   Program Goals
                                 §   Business Goals
                                 §   Benefits
                                 §   Risks
                                • Guiding Principles
                                 §   Metadata
                                 §   Architecture
                                 §   Data Quality
                                 §   Stewardship
                                 §   Data Access and Delivery
                                 §   Security, Privacy and Compliance
                                • Program Organization, Roles & Responsibilities
                                • Program Processes
                                • Architecture Components
                                • Roadmap Activities                               22
The Road To Full Governance
Data Management Program Guide
Guiding Principle Sample Other principles: Metadata, Architecture, Data Quality, Data Access and Delivery, Security, Privacy and Compliance




                                                                                                                                         23
The Road To Full Governance
Develop The Processes




                              24
The Road To Full Governance
Educate and Communicate
Communication Strategy
• Developed a communication plan
• Built a Program Portal
 §    Links to Program projects
 §    Links to key sites and information
 §    Link to common BI portal
 §   Extensive and well organized
     document libraries
 §   Advanced search capabilities
 §   Customized Content “My Portal”

Education Strategy
• Volunteer to present at business & IT meetings
• Meet in smaller groups, one on one
• Talk about Governance when appropriate
• Be prepared for Elevator Speeches

Future Features
• Have a “Google like” query
engine for meta data
• Data Quality Metrics
• Dashboards


                                                   25
The Road To Full Governance
What’s Worked
 §   Governance organization is alive and active…
      §   Cross teams of business and IT folks are learning to talk the same language.
      §   Teams are defining and standardizing business data.
      §   131 customer, contract, and dealer data elements governed.
      §   32 master data lists are governed. Industry standards and Company X standards.
      §   4 hierarchies are governed.
      §   Reviewing DQ cases and resolving problems at the local source via top down
          approach.
 §   Oracle Customer Hub delivering 360 degree view of customer data.
 §   Governed data improved communication with local teams on data needs.
 §   Data steward program is finding data anomalies and process problems
     during profiling.
 §   Local teams understand data issues and are receptive to fixing data.
 §   Business process issues are being reviewed.
 §   Some people are starting to understand and accept governance.
                                                                                         26
The Road To Full Governance
Lessons Learned

§ Governance      is like marriage…
§ Get   dedicated commitments from sponsor(s) and business champion.
§ Governance      takes more time than you expect.
§ You    can not do this alone… seek professional help!
§ Governance      must be a business driven program and not an IT project.
§ Governance      is not about building a data warehouse or a reporting
  solution.
§ Very   few people understand governance, they understand value.
§ Very   difficult to show business value while building out governance.
§ You  can not short cut governance (get sponsors, do assessment,
  build a program guide, develop the organization, assign data
  champions).

                                                                          27
The Road To Full Governance
Lessons Learned

 §   Develop processes with clear roles & responsibilities and follow them. No
     matter what!
 §   No matter how prepared you think you are, your never fully prepared for all
     the unexpected challenges. There are and will be plenty. Follow the
     process.
 §   If people can not be held accountable don’t expect to get focused people,
     timely or accurate information.
 §   Nothing is perfect and there will always be changes. Accept this and
     minimize rework early by modifying the processes.
 §   Exceptions will be needed, document them and get them approved before
     allowing them to happen.
 §   Your data is bad regardless of what you or others think or say...
 §   You need patience, understanding and nerves of steel!!!
 §   If you believe, you can make this happen… it just takes time.


                                                                             28
Q&A




      29