EDM Council General Meeting

1

• EDM Council General Meeting
• New York CityFebruary 16, 2006

2

• Implementation Best Practices Working Group
• Status Report on Progress
• February 16, 2006

3
Alignment of Objectives
EDM Council Mission Create value to its
membership by providing a senior forum to share
information on the business strategies and
practical realities of implementing effective
solutions to manage data across the enterprise.
Alignment of Objectives
Implementation Best Practices Working Group
Objective Produce research and provide guidance
on EDM implementations to reduce the risks and
improve the returns associated with enterprise
data management initiatives
What Works and What Doesnt Work
Deliverables Leverage experience of EDM Council
members to produce and publish a series of
practical implementation guides to serve as
reference tools for Council members.
4
EDM Implementation Lifecycle
Establish project scope and boundaries. Develop
business case. Define internal versus external
resource considerations. Design phased project
plan.
Analyze and capture data control/management
problems to address. Establish architectural
foundation. Identify risk elements. Refine
business case and project plan.
Develop, integrate and test all components and
EDM platform features.

• Inception

Develop implementation plan and manage user
migration from existing environment to the EDM
platform. Often includes several iterations.
Governance
Governance is the single most important
requirement for EDM success. Effective
cross-functional governance and business
commitment to the importance of data management
is required to achieve organizational alignment
and long-term results.
Elaboration
Implement the EDM platform and integrate into
daily work processes of users. Initiate
measurement and documentation of benefits.
Develop continuous system and business process
improvement activity.
Construction
Transition
Operate
5
10 Commandments of EDM Implementation

• Governance top-down sponsorship, adequate
  funding and business unit buy-in as unwavering
  mandate.
• Clarity of Objectives well defined end vision,
  clear and thorough requirements analysis, broad
  and sustainable end-user engagement, open lines
  of communication.
• Business Ownership business requirements as
  driver, IT as enabler.
• Strategic Leader a strong Chief Data Officer,
  CDO, to hold the reins and fully empowered to
  (and accountable for) achieving EDM objectives.
• Balanced Team joint business and IT project
  managers and team members with sufficient
  staffing and knowledge about data.
• Holistic Business Case covering enterprise-wide
  interests and incorporating

• data quality, timeliness, linkages, and process
  improvements
• Recognize Complexities understand data and
  process dependencies associated with linking
  front, middle and back office requirements across
  lines of business.
• Adhere to Core Policies/Procedures including
  data model consistency, business rules and data
  quality stewardship. Business applications adapt
  to the model, not the other way around.
• Phased Implementation iterative, realistic and
  disciplined approach to defining project
  milestones. Phased migration with clear and
  incremental ROI for stakeholders. Dont promise
  what you cant deliver.
• Testing, Training and Internal Marketing process
  change is like a new religion, hard to convert.

6
Key Findings Governance
Effective governance is the most important
component of success
Considerations
What Works
What Doesnt Work

• Viewing EDM implementation as a technology issue
  rather than as a business problem
• Failure to get key business unit stakeholder
  buy-in and participation
• Having stakeholders feel alienated
• Taking a line of business or functional silo
  orientation
• Underestimating the logistical challenges related
  to management by committee
• Failure to communicate on where project stands
  against budget projection
• Not establishing an organization that is
  responsible for data quality and cleansing

• Importance of data management as a core building
  block for doing business
• Overall project ownership, areas of
  responsibility and lines of reporting
• Scope of governance model to establish
  priorities, manage conflicts, promote consensus
  and define the rules of engagement
• Levels of governance (i.e. one for overall EDM
  project management, one for subject matter
  decisions)
• Balance between team empowerment and executive
  control
• Relationship to core stakeholders and
  compatibility with other initiatives
• How to promote involvement without
  decision-making paralysis
• Role of communication as a key for integration
  success

• Empowering a CDO as a single point of contact to
  reconcile internal business unit data conflicts
• Utilizing strong project managers guiding
  delivery teams
• Obtaining representation by front and back office
  functions as well as by business units and IT
• Creating practical policies and structures for
  data ownership and business unit data stewardship
  
• Using an architecture review board to ensure
  alignment between strategy and capabilities
• Charging reference data teams with responsibility
  for data model consistency, data integrity and
  resolution of data discrepancies
• Maintaining strict change management policies
  into production
• Creating and implementing detailed service level
  agreements

7
Key Findings Inception
Mobilizing business and IT stakeholders is
necessary for buy-in and funding
Considerations
What Works
What Doesnt Work

• Senior management mandate for downstream systems
  to use central reference databases
• Top down sponsorship and active involvement by
  corporate leadership
• Flexible central reference data teams to work
  with timescales of user systems migrating to use
  of central database
• Clear understanding of EDM as enabler of other
  applications rather than end solution
• Balanced business and IT involvement with strong
  front office representation and backing

• Use of workflow technology to enhance control and
  promote business process automation
• Data integrity and data access throughout the
  transaction chain
• Identification of stakeholders, requirements
  methodology process, level of buy-in and extent
  of involvement
• Reuse of common data elements and agreeing to the
  use of standards to shorten development efforts
  and to provide early deliverables to business
  units
• Determination of the level of resources and
  extent of internal expertise required for EDM
  implementation

• Operating without a precise statement of criteria
  for measuring success
• Scope creep and lack of clearly defined roadmap
  to achieve end vision
• Failure to achieve balance between near-term
  tactical deliverables and long-term benefit
• Misalignment between IT and business objectives
• Underestimating the complexity in raising initial
  financial sponsorship and securing ongoing
  funding allocations
• Failure to understand the sustainability of EDM
  initiative in competition with other internal
  priorities
• Underestimating the challenges in scheduling IT
  work across multiple downstream systems
• Outsourcing project without a strong framework
  plan

8
Key Findings Elaboration
Must engage the right people to clearly define
requirements
Considerations
What Works
What Doesnt Work

• Realistic and disciplined approach to defining
  project phases
• Strong incremental ROI for business units
• Balanced business and IT involvement with strong
  front-office representation
• Concise documentation and substantive
  communication
• Strong methodology to boost confidence
• Honesty about commitments
• Phased migration approach with ongoing support of
  multiple databases through the transition
• Separating the program into smaller targeted
  stand-alone projects
• An ideal mix of internal, consultant and third
  party involvement

• Failure to recognize the importance of ensuring
  added value for front office applications
• Misalignment between IT and business objectives
  and complexity of engineering solutions
• Allowing one group to be dominant in defining
  scope and objectives (and allowing other groups
  to lay quiet)
• Expecting business units to share architecture as
  well as the costs of IT development
• Simplifying operational characteristics and
  resource requirements related to supporting
  multiple business units
• Underestimating the difficulty in reconciling
  multiple legacy systems
• Lack of staff resources and underestimating the
  risks of rework from parallel development
• Failing to fully understand business unit
  requirements and how users relate to both data
  and systems

• Scope, planning milestones, governance structure,
  implementation timeframes and budget
• Current state analysis with key stakeholders
  (CRM, credit risk, compliance, operations,
  finance, etc.)
• Integrity of data model and understanding of data
  dependencies throughout enterprise
• Impact on data acquisition, cleansing, storage,
  processing, distribution to downstream systems
  and access milestones
• Level of business case rigor including costs and
  benefits of various phases
• Boundaries of new and existing architecture
  including middleware requirements and interface
  design
• Phase and interim operation models for
  implementation
• Validation of software RFP and defining best
  practice implementation and operating plans

9
Key Findings Construction
Building and rolling out the platform in phased
increments is key
Considerations
What Works
What Doesnt Work

• Lack of centralized governance
• Failure to apply regional feeds and hierarchical
  rules to counter-party processing
• Limited knowledge transfer of components
  developed by external parties
• Failure to adequately design mapping and domain
  table storage
• Over engineering the EDM solution in complexity
  and detail
• Over engineering the application construction
  process

• Understanding of user requirements (when
  application is needed, business functionality, ST
  vs. LT objectives) prioritized against budget
• Definition of core business rules to be
  incorporated into validation and data loading
  process
• Definition of testing and implementation
  strategies
• Appropriate mix of internal and external
  personnel and process (skills transfer and
  effective change management)
• Use of third party vendors in key portions of
  overall program
• Reuse of functionality (assume 80 of
  functionality can be leveraged)
• Subsequent data population efforts after initial
  project phase

• Ability to demonstrate incremental and measurable
  progress to key stakeholders through iterative
  releases
• Line up early adopters and maintain upfront
  constituencies
• Normalization of definitions, attributes and
  field names for applications precision and ease
  of use
• Ample test environments, product release version
  control and strong change management procedures
• Use of regional teams to address local data feeds
  and mapping to local standards
• Offshore resources for testing
• Identification of specialized data requirements
  and unique applications requests
• Separation of GUI development from backend
  mapping, enrichment and validation

10
Key Findings Transition
Strong processes are needed to sell migration and
support end user adoption
Considerations
What Works
What Doesnt Work

• Business continuity planning on a global basis
  and quality of MIS
• Knowledgeable personnel to recognize how the
  output should appear and identify subtle data
  issues
• Documentation, user guides, key operating
  procedures and training
• Use of governance to prioritize operational users

• Allowing sponsors to be uninvolved
• Underestimating the level of regional variance
  and resistance to new process
• Minimizing the level of assistance required to
  help users understand an entirely new process
• Underestimating the amount of MIS breaks and
  local IT support required
• Insufficient guidance and Toolkits for adopters
• Postponing training to the last minute
• Failure to market product to users (alienation)
• Adopting a hot fix mentality
• Underestimating the importance of version control
  and scope of regression testing required

• Constant coordination with users and clear
  methodology to obtain user feedback and
  incorporate into future releases
• Strong end user experience and data knowledge for
  those that interface with clients
• End-to-end process definition and UAT with key
  hub regions (site visits, process walk-through,
  crib sheets, live telephone support, local
  presence)
• Vendor SLA definition and management
• Tight control over change and implementation
  plans
• One-by-one migration of downstream systems
  (phased implementation)
• Run new data sources in parallel and compare data
  (switch off old data sources when content is in
  sync)
• Web-based user documentation and user support
  post initial training
• Final UAT of central tool using original data
  sources

11
Key Findings Operate
Clear operational model needed to support new
processes and future enhancements
Considerations
What Works
What Doesnt Work

• Building procedures to ensure that the
  operational model is as well constructed as the
  data model
• Continuing to roll out enhancements, refine and
  add business rules and improving validation
  checking
• Implementing sign off process on changes and
  major releases
• Establishing a governance program for small
  changes
• Establishing a workstation for overrides
• Maintaining error logs and reports
• Creation of off-shore support team

• Data model integrity and enforcement of data
  quality issues
• Metrics that are needed to determine EDM return
  on investment (i.e. cycle time, new account
  creation, account data elements, overall data
  quality, process replication, processing
  timeframes)
• Establishment and management of ongoing data
  ownership and stewardship
• Coordination of new applications requirements
  from users and change control
• Global maintenance and support
• SLA process for managing vendors (i.e. feed
  updates, product enhancements, de-bugging)

• Sponsor/steering committee satisfaction as gauge
  of success
• Poor closure process (lessons learned, budget,
  benefits documentation)
• Pushing for rapid change causing instability and
  data errors
• Sporadic development after the initial phase is
  implemented
• Underestimating the data clean up requirements
• Underestimating the degree of difficulty in
  establishing a go live date across all global
  operations
• Over validation of data causing conflicts in data
  ownership
• Underestimating the importance of scalability
  requirements
• Underestimating the impact of local variation in
  data and data usage (regional conflicts)

12
Working Group Next Steps

• Expand depth of data collection on current menu
  of implementation issues by issue, domain area,
  role and function
• Extension of research process to cover
• Governance structures, data stewardship
  approaches and communications alternatives
• Internal versus external staffing (mix of buy,
  build, partner and outsource)
• Vendor selection and management
• Strategies to introduce and expand EDM solutions
• Practical transition management
• Data validation and quality control processes
• Implementation metrics and benchmarks
• Development of EDM infrastructure
• Repository of EDM best practices and mechanism
  for ongoing contribution
• Network of experts and online contact lists
• Web site functionality (message board)