This phrase has been attributed to several people including: Ben Franklin, Winston Churchill, and Abraham Lincoln. Despite whoever said it first, the phrase is as true for the EDM Program as for any other ambitious endeavor.
Many organizations excel in planning, executing, and monitoring specific projects, which can be anecdotally illustrated by the number of individuals in the IT and data management industry who have achieved Project Management Professional (PMP) certification. However, organizations often falter in planning for large programs, in particular, those with enterprise-wide scope.
Since an organization’s data will exist (and continue to grow) as long as it is in business, the enterprise data management (EDM) program, unsurprisingly, must be crafted to address the anticipated future state of the organization. The US Army, or countries such as Switzerland, do expect to exist virtually forever, and therefore engage in decades-long planning. Your organization may address shorter time horizons, such as three or five years.
In this column, we’re going to consider the primary dimensions and functional consequences of ‘planning failure’ and apply them to the EDM program and its key components. Let’s refer to the rows in the figure below, created by Delorese Ambrose, a professor at Carnegie Mellon, who founded Ambrose Consulting and Training to help Fortune 500 companies in strategy, change management, and communications.
- Row 1, lacking Vision, represents an organization that has no clearly articulated business strategy. For example, aiming at ‘30% revenue growth per year’ is a goal, but if that’s all the Board can come up with, there’s no path to ‘what, exactly’ or ‘how, exactly’ the growth will be achieved. The overall result is Confusion. (e.g., Why are we doing this? We don’t know.)
- Row 2, lacking Skills, represents a failure to determine what staff needs to know to successfully implement the vision. If the organization has not considered the knowledge that must be applied, then there will be unreasonable expectations and, correspondingly, a lack of training. Understandably, this leads to Anxiety on the part of the staff. (e.g., They expect me to do what? I have no idea how to accomplish this.)
- Row 3, lacking Incentives, represents that the organization has not thought through what it will take to motivate staff in order to achieve the vision, AKA, has not answered the question ‘What’s in it for me?’ at the organizational, business line, and individual levels. The result is Gradual Change, slow progress, and little enthusiasm for action.
- Row 4, lacking Resources, shows that the organization has not sufficiently considered the level of effort that is needed to accomplish its goals and objectives. This issue is endemic in the data management industry, for instance, how many chief Data Officers have zero staff? Far too many. The result is Frustration caused by too few staff trying to accomplish too much. Frustration is a motivation-killer and frequently a primary cause of staff attrition.
- Row 5, lacking a Plan, is the primary subject of this column. It represents an organization that has not engaged in specifying HOW it intends to accomplish the vision. This is almost a guarantee of False Starts, for example, many organizations repeatedly fail in their attempts to launch data governance. This is inefficient, expensive, demotivating, and confusing – both measurable progress and staff enthusiasm are negatively affected.
- Row 6, all components operational, is the best assurance that the organization’s vision will become a reality. This is where we all aim to be – positive Change.
I’ve observed that organizations tend to lack concrete plans for data management programs and activities beyond the project level. The only common exception to this typical situation is when projects interact, or are dependent on, other projects – for example, the initial data population for a new CRM system would likely be migrating from an existing CRM system and probably also obtaining data from other sources – with the level of expenditure and visibility of this or a similar implementation, the organization tends to plan carefully to ensure that this project is successful. But data as a critical component of infrastructure is often neglected – so let’s change that.
The greater the scope of the proposed effort, the more importance should be given to planning. In recent columns I’ve emphasized the importance of Vision, the logical starting point for setting goals and objectives. Quick reminder: the organization should be able to answer the question “What does good look like?” This question can be applied to the overall business strategy, to the EDM program, to its component data management activities and projects, and finally, to the creation of meaningful metrics at all levels.
Since the data is forever, the EDM program must be designed to be in place forever. This is not to claim that the approach is written in stone and never changes. In fact, the overall program will certainly be modified, re-planned, adapted, enhanced, and extended over time. However, there are baseline plans that need to be created, since the organization should:
- Know what it wants to achieve – e.g. why does it want to manage its data better, and what the benefits and results are expected to be
- Know what it will DO to achieve results – e.g., what are the elements that need to be put in place – policies, standards, governance bodies, defined processes, resources, etc.
- Know what its priorities are – e.g. within the scope of the EDM program, what data domains are highest priority, whether to address major problems, achieve a result critical to the organization’s business strategy, or both
- Estimate realistic durations – e.g. the organization can’t fully populate a metadata repository, or fully trace the lineage of all its core data, in a short period of time.
Frequently, organizations are overwhelmed by the magnitude of change that is required to evolve effective data management capabilities. Using the ‘quarterly profits’ analogy[1], there is often pressure to achieve results quickly, the equivalent of starting to code a system before the functional requirements are approved.
One example of this tendency, occurrences found in many organizations, would be an executive demanding that (the unlucky leader selected) ‘Set up data governance – now!’ This approach to establishing data governance has led to many failures, due to (no surprise) lack of vision, lack of incentives, lack of skills, lack of resources, and lack of a plan. For example, I’ve heard many statements like these: ‘we tried to launch governance three times and it failed,’ ‘the stick is wearing out,’ ‘no one knows what they’re supposed to be doing,’ and so on.
In deciding to commit time and effort to developing a plan, there is no substitute for active senior management engagement. The sponsoring leader must communicate the importance of the planning effort, because success will require engagement from all relevant stakeholders.
The list below highlights types of plans that have proven to be useful for establishing and sustaining the EDM program:
- Strategy – A strategy is, simply, a plan of action and policies designed to achieve an overall result. For an enterprise-wide data management program, an organization benefits from creating at least four strategies – we’ll discuss these below:
- Enterprise Data Management Strategy (DMS)
- Data Quality Strategy (DQS)
- Metadata Strategy
- Communications Strategy
- Sequence Plan – a critical supplement to a strategy, describing the order of activities that must be accomplished for a strategy to succeed – also referred to as a ‘Roadmap’
- Charter – a document which establishes a formal group within an organization, describing its purpose, scope, authorities, and responsibilities
Plans, Plans, and More Plans
Let’s say your organization lacks a reasonable set of clear, comprehensive EDM plans. You can use the tables below, representing initial outlines, to inform your planning efforts. Each describes a minimum set of topics that should be addressed. The actual table of contents will be tailored to your organization – based on where it is now and what it intends to achieve.
Data Management Strategy – Scope: The overall EDM program
| Topic | Description |
| Vision | Vision statement – the overall aspiration of the organization for the enterprise data management program. |
| Goals | The business goals – what the EDM program intends to achieve for the organization, and its alignment with the business strategy. |
| Objectives | What key accomplishments the organization commits to implement to realize each stated goal. |
| Guiding Principles / Framework | What fundamental assumptions the organization wants to establish and embed in the EDM program, and what data management framework will provide a baseline for evaluating the program[2]. |
| Data Scope | A description of what data the organization intends to address within the scope of the EDM program – that is, what it has determined to be ‘enterprise data’. |
| Challenges / Issues | A high-level description of current gaps in the way data is managed, and persistent problems that the program will address. |
| Program Scope | The primary capabilities that the EDM program will establish and implement, such as data governance, metadata management, data quality, data lineage, data requirements, asset library, etc. These should follow from the goals, objectives, and current challenges. |
| Leadership and Organization | The organization (existing or new) that is responsible for the EDM program, and the executive to which it reports. |
| Resources | What roles, and how many staff will initially be applied to establish this program and its components. Should include a RACI chart. |
| Governance | What governance bodies will be established to assure mutual decision-making about enterprise data involving all relevant stakeholders. |
| Metrics | What initial metrics the organization will establish and monitor to measure progress against goals and objectives. |
| Compliance | How the organization will ensure that once standard processes and policies are established, they are followed. |
| Sequence Plan | A high-level sequence plan for program implementation – recommended is a 1-2-page diagram and brief text description (see sample diagram below). The activities/elements of the sequence plan will then be decomposed into program and project plans following the publication of the strategy. |
Data Quality Strategy[3] – Scope: Data quality disciplines applicable to the entire organization
| Topic | Description |
| Vision and Rationale | What the organization is committing to achieve for improved data quality, and why it is needed. |
| Goals and Objectives | Specific goals and objectives for the data quality program, aligned with the DMS. |
| Business Benefits | What benefits can be expected for the major business lines from implementing a data quality program, as well as benefits for key capabilities, such as analytics, a data lake, etc. |
| Program Scope | Describes the major data groupings (subject areas) included in the program. |
| Program Elements | Describes the policies, processes, standards and guidelines that will be created through the program. |
| Technologies | Describes the tools (or tool selection processes) which will be acquired to support data quality disciplines, e.g., profiling, address standardization, quality rules for shared data, metadata tools, etc. Specific tools can be proposed. |
| Roles and Resources | Describes the staff roles that are actively engaged in the program activities, and proposes initial resources. |
| Governance | How the quality program will engage with governance groups, data working groups, etc. |
| Training | Describes the roles for which training will be provided, including for business and technology staff. |
| Sequence Plan | The sequence plan for the DQS will have several tracks – capabilities to be established, subject area prioritization, key pilot projects, organizational rollout, and training. |
Metadata Strategy – Scope: Organization-wide determination of what information to capture and stores about ‘enterprise data[4]’
| Topic | Description |
| Vision and Rationale | A vision statement addressing the aspirational state of knowledge management about the organization’s data assets, why metadata is necessary to achieve business goals. |
| Goals and Objectives | Goals and objectives for the metadata program, aligned with the EDM Strategy. |
| Metadata Categories | A summary of the primary metadata categories with examples of their use in the organization – business, technical, operational, process, etc.[5] |
| Metadata Scope | Describes the high priority metadata categories for which metadata needs to be captured and stored and lists the enterprise data subject areas for which metadata will be developed. It may also specify key projects for which metadata capabilities will be enabled, e.g. the new Customer Master, the redesigned data warehouse, the new data lake, etc. |
| Metadata Program | Describe program elements. Metadata is often a layer of activity applied to many projects across the enterprise, but creating a comprehensive set of metadata for enterprise data is a sustained program in its own right. |
| Metadata Repository | Describe the toolset to be used (or the selection process) for the organization’s metadata repository, and the primary features it must support for search, editing, reporting, etc. |
| Meta-Model | Include a high-level description of the categories, sub-categories, and properties that the metadata repository must support. |
| Metadata Sources | List the primary existing sources of metadata which will be consolidated in the metadata repository. |
| Leadership and Organization | Determine the organization that will be responsible for acquiring, implementing, and maintaining the metadata repository, and developing standards, and the executive to which the organization reports. |
| Governance | Describe the role of governance bodies in proposing, capturing, and modifying metadata; engagement in the metadata policy; and permissioning decisions. |
| Compliance | Describe the compliance process, once established, and what organization will assure that it is followed |
| Metrics | Describes the initial set of metrics that will be applied to measure implementation progress and stakeholder satisfaction (e.g., search success survey, etc.). |
| Sequence Plan | The sequence plan for the metadata strategy should include: technology implementation; repository population segments by priority; sequence of category and feature enhancements; responsibility rollout by organization; and training. |
Communications Strategy – Scope: A comprehensive plan for how EDM program communications will be conducted and managed. It is important because the value of data management is still challenging to sell internally in many organizations. Clear, consistent, informative communications is a fundamental requirement to educate staff and demonstrate the accomplishments and value of the EDM program.
| Topic | Description |
| Vision and Rationale | Vision statement that addresses what the organization wants to achieve from effective EDM program communications, and why it is needed. |
| Goals and Objectives | Goals and objectives for EDM communications, such as informing stakeholders about program progress, highlighting business benefits achieved through the program, etc. |
| Audiences | Description of the internal audience, including executives, business leaders, data stewards, data management organization, information technology, etc. The overall audience is essentially anyone engaged in program activities or benefitting from program activities. The strategy should describe the primary information needs of each audience. |
| Communication Channels | Describes the channels of communications that will be utilized, such as the EDM program web site, emails, IMs, in-person meetings, town halls, documents, presentations, etc. |
| Communication Types | Describes the types of communication required to meet communication goals, such as information notices, progress dashboards, emails, case studies, videos, computer-based training, etc. |
| Communications Frequency | Describes the frequency of regular communications. Useful to include a three-dimensional chart depicting communication types and frequencies by audience type. |
| Sequence Plan | The sequence plan includes key activities that establish and grow the communications landscape over time, which may include summary presentations, e.g. unveiling the data management strategy or the data quality strategy, posting the objectives dashboard, a push notice about data awareness training, etc. |
Charters – Scope: Typical charters that are likely to be created in an EDM program include those initiatives that require a formal group to be established and operationalized, for example, for an Executive Data Governance body, a Data Stewards body, a Data Owners body, and a Data Management Organization. The need for these bodies is established in the EDM Strategy and when they will be formed is included in the accompanying sequence plan. A charter is required, rather than a project plan, because these groups are intended to be permanent program elements. While the mission, tasks, and participants may change over time, the groups will be active over the life of the EDM program. The example outline below will require tailoring for the Data Management Organization, because its responsibilities are intended to be full-time job functions.
| Topic | Description |
| Vision and Rationale | Vision statement that addresses what the organization wants to achieve by establishing this group, the purpose of instituting governance, and why it is important to the organization. Guiding principles, aligned with the EDM Strategy, may be referenced as well. |
| Scope | Describes the scope addressed by the group, which may include the entire scope of enterprise data or any segment thereof, e.g., governance body for the design of a Customer Master data hub. |
| Goals and Objectives | The overall goals for the group, which may include support data-driven decision-making, improving data quality, prioritizing and resolving critical data issues, establishing data ownership, etc., and specific objectives, for instance, a focus on implementation of a data lake. |
| Framework and Organization Interaction | A diagram and explanatory text for the proposed structure, indicating its reporting chain, interactions with other groups and business lines, and key areas of focus. For data governance, this will typically be a multi-leveled structure. Each of the governance bodies should have its own charter (or section, in a combined charter), reflecting the decision scope, activities and authorities of that group. |
| Composition | Describes the members of the group by role (and possibly by name), for example, a chairman and one representative from each of 8 business lines, and a RACI chart showing responsibilities of the members. |
| Decision Practices | Describes how mutual decisions will be made, e.g. how many individuals constitute a quorum, the voting scheme, when decisions are escalated, etc. |
| Communications and Frequency | Describes how often the group meets, either in person or via conference meeting, and how communications among members will be accomplished. For instance, a vendor product may be used for workflow management. |
| Reporting and Metrics | Describes the type and frequency of reports that the group provides to its stakeholders, and initial metrics, which may include counts of decisions made and issues resolved, customer satisfaction, progress on objectives, etc. |
| Sequence Plan | The sequence plan for a charter is a separate document, indicating the initial areas of focus, with their order and duration estimated. Governance is most successful and effective if it is primarily focused on a series of projects, while sustaining functions and decisions are also being conducted. This also demonstrates to executives that the members time is providing value to the organization. |
Sequence Plan – Scope: A high-level depiction of activities / projects, arranged by order and duration, should accompany all the strategies above. A diagram with accompanying text is sufficient for a sequence plan. The elements of the sequence plan are decomposed into detailed project plans when they are initiated. An example sequence plan diagram, in this case for a Data Quality Strategy, is provided below.
Sample Sequence Plan for a Data Quality Strategy
Organizations also often do not give enough attention to data architecture plans, which are important because the expense is significant, and the purpose is to guide implementation. The two plan types below are often missing in a review of EDM work products.
- Transition Plan – A data architecture sequence plan that details what components (data stores, systems, technologies, etc.) exist now and how the organization will move to the desired target state addresses data store components to be acquired, built, consolidated, or retired, showing affected data stores and when transformation activities will occur. Typically accompanies target (or ‘ To Be’) data architecture documentation. By the time the organization develops transition plan, the rationale and business case have been approved and the business case justification is formulated.
- Integration Plan – An extension and decomposition of the transition plan with enough detail for implementation purposes, describing when and how the data stores (or sets) being integrated, will be combined into a new data structure. For each transformation effort addressed in the Transition Plan, content will address specific data sources, and the steps and sequence of their integration into the target data store(s).
A few weeks of planning will save months and months of effort and expense. IBM’s software engineering effort chart, published in the early 80s, showed the relative cost of adding major system features in various stages of the software development life cycle – their estimate of relative effort/cost was 10% added to the project in the Requirements phase, rising to 90% in the Development phase. While the disparity may not be as great for EDM program components, failure to plan does negatively affect the progress and success of the program. As a data management professional, you can encourage your organization to engage in planning, to their great benefit. In addition, leading a successful planning effort enhances your reputation and is a strong qualification for more senior positions.
[1] My favorite go-to shorthand phrase, referring to publicly traded companies which devote considerable energy to achieving favorable quarterly results, captured in unaudited SEC 10-Q filings. The demands of the markets, which punish negative short-term results, are frequently blamed for lack of adequate strategic planning.
[2] I suggest the Data Management Maturity (DMM)SM Model as the foundational framework, since it was designed as a measurement instrument based on the typical path of data management capability enablement.
[3] I recommend that the Data Quality Strategy be created following the EDM Strategy, as a further refinement and articulation of overall quality objectives. It is often a great assist to first conduct a high-level analysis of the major quality issues plaguing the business lines – the major issues will drive many of the DQS activities. The Rationale section can refer to the problem analysis.
[4] ‘Enterprise data’ is the scope of data that will be defined, governed and controlled through the EDM program. Typically, it includes: highly shared data, such as master data or reference data; important shared data, for example, Customer, Product, and Sales data; personally identifiable information; data needed for analytics; and data required for regulatory purposes. Each organization must define the boundaries of its enterprise data.
[5] Recommended to include this ‘what is metadata’ section, because metadata is complex, and executives often bounce off the concepts.
