It’s been a while since I’ve written about the goals and objectives of an Enterprise Data Management (EDM) Program and how organizations achieve and sustain significant improvements in fundamental data management disciplines in their journey to Great Data.
In this series of columns, we’re going to explore “What Good Looks Like.” We’ll summarize the collective accomplishments of high-capability organizations, with examples of how they’ve built their program and challenges overcome. (Names withheld to protect the superstars from the paparazzi).
And we’ll provide suggested approaches and some work product outlines to help your organization follow in their footsteps. Without further ado, let’s start at the top, with the EDM Program and EDM Strategy.
The EDM Program
Data management activities have continuously been occurring in multiple functions and business processes across an organization since the dawn of the information age. They are typically centered on information technology projects delivering solutions, but historically, they have been project-based.
Regardless of the knowledge, experience and professional excellence of the implementation teams, processes, standards, templates and guidelines produced by even the most successful projects are frequently not standardized and extended for enterprise-wide adoption. Much like galaxies separated by the vast empty spaces between, or bright flashlights in a forest wilderness, progress is local and is not leveraged.
To consider just one example of how this typical situation inhibits attaining organizational goals: Most organizations have data dictionaries (at least a few instances) reflecting the content of specific data stores. Even if some subject areas, entity types or attributes are shared by multiple business areas, usually there is no attempt to harmonize data names, descriptions, datatypes or values across other databases that may capture and store the same data.
Therefore, when (hopefully when, not if) an organization realizes that inconsistent use of business terms is hampering internal communication and reporting or needs to undertake a redesign or integration of shared repositories, they face the challenge of where to begin, who to involve, and how to gain agreements for a business glossary, starting with multiple data dictionaries representing the same data. This is inefficient, effort-intensive and messy – and definitely to be avoided if possible.
Whether the primary driver is a major technology transformation, compliance with regulatory reporting, high maintenance costs, a want to gain competitive advantage through predictive analytics, or another strategic business priority, eventually, most organizations come to the conclusion that they have to ‘get serious’ about their data.
Voila! The birth of the EDM Program – “The infrastructure, programs, projects, and staff resources that an organization commits to for the purpose of establishing and following sound data management capabilities for its prioritized data assets.”
The diagram above, which you may remember from previous columns, highlights the three EDM Program elements[1] that are important for developing an Enterprise Data Management Strategy.
Since it’s not a great idea to rush blindly into the wilderness without a map, preparation and planning is strongly advised. Where do we start? Well, with Step One – determining the size of the territory to be navigated, analogous to ‘Do I need a State map, a County map, or a National Park map?’
Know Thy Enterprise Data
It’s critical to clearly define the scope of an EDM Program. A first consideration might be answering a simple question – how big is your budget, relative to your desire to achieve high priority business goals? The EDM Program is a commitment to evolve data management processes to the extent necessary to have ‘Great Data’ corresponding to what’s critical to achieve your business goals.
But not all data is ‘enterprise data,’ and the organization needs to define what the boundaries are. If the scope is not clarified, the EDM Strategy, budget, staff resources, and supporting technologies can’t be fully determined. This would not only be a big hindrance to effective planning, but could derail the entire effort, e.g., the sin of over-promising and under-delivering, most hated by executive management. Many organizations I’ve worked with hadn’t previously considered this decision. If your organization hasn’t yet decided on its program scope, we can start with this diagram. The four segments characterize the data assets that should be taken into account in defining scope.
- Master & Reference Data: One vitally important segment of data assets to manage well are those containing ‘highly shared’ data. In most organizations, highly shared elements include Master data (identifiers and core descriptive attributes for key entity types, such as ‘Customer’ or ‘Product’) and Reference data (consistently referenced by multiple business processes, such as country codes, interest rates, etc.). Your organization needs to agree about which key entity types are highly shared – that’s part 1 of the scope.
- Mandatory Reporting: Another segment originates from reporting requirements for industry, national, international, or state regulations. For example, there are many financial industry regulators, including the Federal Reserve, Office of the Currency Controller, Federal Deposit Insurance Corporation, Securities and Exchange Commission, Financial Industry Regulatory Authority, the National Credit Union Administration, etc. Each of these agencies and organizations mandate standards and reporting. Since in the end, a financial organization has a survival-level need to stay in business, data used to generate regulatory reports is part 2 of the scope.
- Critical Business Processes: Business processes in the ‘lines of production’ (integral to the mission and business functions of the organization) produce and manage data that is vital for the organization’s operations. At least a portion of this data is likely to be shared by more than one business area. For example, if a bank provides auto or home loans, the details of the loan application and supporting documents may not be shared, but the amounts, delinquencies, and other financial data are aggregated for reporting, and individual loan information may also be shared (e.g., Marketing and Sales for campaigns, for Customer Service operations, etc.), since they are customers of the bank. Therefore, this data should be part 3 of the scope.
In addition, for the business processes most vital to the organization, most of the data produced and managed from end to end may be included in the scope. For example, an inmate management system for a prison system, the beneficiary payments system for a pension plan administrator, etc.
- Other Shared Data: This segment may have a different makeup for every organization, but typically includes data, such as Employee and Location, that is produced and managed for key supporting processes, some of which is shared. An analysis of this segment should complete part 4 of the enterprise data scope.
Determination of scope is the foundational decision that allows the organization to create an EDM Strategy. As a part of the EDM Strategy, the organization will identify data Domains – evolving towards management of data independent from systems – and the analysis of enterprise data is the high-level outline that is decomposed in the EDM Strategy to define Domains and describe the roles of Domain Stewards and Data Stewards.
The EDM Strategy
What does GOOD look like for an EDM Strategy?
High-capability organizations have made the commitment and effort to think through what they want from their data, assess the current state of data management practices, determine the desired end state, and set out a Grand Plan for how the vision will be achieved.
What does your organization have to do? First of all, launching the effort to create one – it takes some ‘ooomph’ to get off the ground. “Ya had’ta wanna!”[2] – interpreted in the light of business, we’re talking about drive, commitment, and collective intent – the will to achieve Great Data. Otherwise, why bother to spin up the EDM Program at all?
Many organizations face significant challenges in marshalling the motivation, energy, and executive commitment to create an EDM Strategy. For the first challenge, defining enterprise data, we’ve offered an approach. Other typical challenges include:
- Lack of a clear business strategy – what do we want to be when we grow up?
- Difficulty associating business problems with data problems – every business process depends on data, or as desert Bedouins instructed in selecting a mount, “No foot, no horse.”
- Lack of operationalized data governance – who is responsible for what data, and how do they work together for mutual decision-making?
- Lack of organization-wide perspective / functions – lacking a focus on business process improvement, no enterprise architecture, no target data architecture, etc.
- Lack of executive commitment – to managing data as a critical asset. Executives have to get religion, or lack of attention will dim the arrival celebration when it is published and result in lack of follow-through.
- Relying on contractors – sure, you can get help, but the EDM Strategy should be largely home-grown – muscles are muscles, and yours are yours.
These challenges are discussed in detail in my previous TDAN columns “Failing to Plan is Planning to Fail” and “Why Your Organization Can’t Create an EDM Strategy.” When they are faced and overcome, the organization has won its way through to create a consensus strategic Vision – As Plato set forth, The Good, The True and The Beautiful (and as Aristotle expounded, “the primary transcendent properties of Being.”)
The EDM Strategy establishes (or expands) the EDM Program, which is forever, because your data is forever and you’ll always have to manage it. Organizations that have developed a comprehensive, viable EDM Strategy have accomplished the following in creating the strategy and developing the content:
- Collaborative creation with business lines and key stakeholders
- A clear set of business aspirations that the strategy enables – rationale for the program
- Identification of current gaps in processes and practices
- Identification of priority data management disciplines, policies and processes to implement (e.g., Metadata, Data Quality, Target Data Architecture)
- A short development time-frame – e.g., 90 days, not nine months
- A data governance structure for oversight and building of data management processes and persistent products
- Identification of data management policy requirements for development
- A comprehensive sequence plan of 2-5 years duration
- Review and approval by executive data governance
- Broad communication and promotion across the organization
- Broad education about the EDM Strategy for all relevant stakeholders.
The last three bullet points are important for developing a ‘data-aware culture.’ Executive approval and promotion of EDM Strategy is like planting a large flag on a hilltop; everyone can see it and the direction is clear. In the end, every stakeholder who defines, produces or uses data must row together to move the boat along swiftly and smoothly.
Note – An EDM Strategy may be combined with one or more related planning components, such as a Data Technology Strategy and a Target Data Architecture, to create an overall Enterprise Data Strategy. However, this approach will greatly increase the duration and level of effort, which will cause enthusiasm, motivation and impact to dissipate. Since the EDM Program is a significant organizational transformation, in and of itself, creating the EDM Strategy is essentially gearing up for a big change to the status quo, and deserves focused energy and attention.
What the EDM Strategy Should Include
The EDM strategy defines the overall framework of the EDM program. It is the formal rationale, outline of elements, and implementation guidance for the program, collaboratively developed and approved by all relevant business lines. It addresses what the EDM Program is designed to achieve, capabilities to be developed over a specified period, and the scope to which they will be applied, accompanied by initiatives and key milestones in an accompanying sequence plan (Roadmap). A data management strategy should include, at a minimum:
- A vision statement (described above – aka, ‘data Nirvana’ for the organization), with an overall summary and the business aspirations for data assets that achieving the vision would enable.
- Core operating principles, such as ‘minimize redundant data,’ ‘data first design,’ ‘rationalize before build,’ etc.
- Program goals, aligned with the organization’s business goals.
- Defined objectives to achieve EDM Program goals.
- Scope of enterprise data assets – (addressed above) – high-level list of Domains which are the focus of the EDM Program.
- Major gaps – summary of the current state of data assets and management practices,[3] and the negative impact they cause in achieving business goals and objectives
- EDM process scope – the disciplines, policies, and data management business processes that are required to attain business goals and remediate gaps (e.g., business glossary, data profiling, data lineage, etc.)[4] EDM processes include corresponding work products, such as policies, standards, and guidance.
- Business benefits – these should be described
- Satisfy use cases – e.g., predictive analysis of potential Product sales based on season, geographic area, economic factors, demographic trends, etc.
- Improvements – e.g., to Customer Service, Regulatory Compliance, Product Development, etc.
- Tangible benefits – e.g., minimize maintenance costs, reduce quality defects causing delays in closing the books, ROI for speeding up development of new products, etc.
- Priorities – how priorities – both Domain and EDM processes – are determined and what factors are involved – e.g., dependencies, perceived business value, alignment to strategic initiatives, and level of effort.
- Alignment – The EDM Strategy development process will allow you to create a strategy alignment chart (a trick of the trade.)
Why? To ensure that your approach has been comprehensive, and to give yourself a leg up on convincing executives to support the EDM Program. They will ask ‘What does the EDM Program do for us? and you will be able to answer – ‘This.’ [The rows may be 1:1, M:1, or M:M (the handy Merge Cells feature)].
From this chart, you can then narrow your focus to develop the Elevator Pitch presentation, which you will need to deliver, in various forms, not only to get the program and funding approved, but to keep up momentum over several years as you check off accomplishments and trace the resulting benefits.
• Governance structure – a high-level description of governance roles, governance bodies, and how they interact (or refer to your governance model if operationalized) – engagement of business data representatives to define data, build, enhance, and control the data assets.
- Staff resources – estimate of resources required, and new positions to be filled – for example, the Data Management Organization, a Chief Data Officer – whatever the size and scope of the program requires.
- Metrics – how will you know you’re achieving program objectives? An initial high-level set of progress and component process metrics should be set out in the strategy.[5]
- Benchmarking – what method and measurement instrument the organization will adopt to achieve an objective measure of capability development and implementation.
- And last, but definitely not least – a high-level sequence plan, 2-5 years in duration, showing the major initiatives and taking into account in-flight and planned major projects – e.g., implementing a vendor CRM system, implementing a data lake, etc. For the sequence plan, I’ve found that a one-page chevron diagram across the selected timeline is most effective, with text descriptions following, and decomposed into supplementary charts for additional detail about major initiatives.
Managing the EDM Strategy Initiative
Once you convene the EDM Strategy working group, which should include senior data experts from the business lines which produce and use enterprise data, reach consensus on an aggressive timeline and manage your agenda and feedback sessions on the topics outlined, which will keep participants focused on the vision and elements of the strategy.
If there are persistent disagreements, (e.g., ‘My LOB wants Customer Accounts to be the first priority,’ ‘No, my LOG wants Product to be the first priority’) summarize the issue, send the champions back to their business lines for feedback, and if necessary, escalate the issue.
Aim to develop a draft for review by the executive governance body within 90 days from the first working group meeting. Remember that perfection is not the goal, building consensus and enthusiasm is more important.
The EDM Strategy must be able to evolve as the needs of the organization change, so once the first version is approved, plan to review it at least once a year. Take your program metrics seriously – they satisfy multiple requirements: progress checks for the working groups and initiative teams (‘We really are getting somewhere’); maintaining executive interest (‘Look at that, the Glossary Working group defined and approved 200 loan business terms in the last quarter.’); and encouraging the lines of business to be patient while the organization manages to enterprise priorities.
When
milestones and objectives are achieved, trumpet your successes, recognize every
participant, and relate your success story far and wide. The next Almost Heaven
column will discuss essential supporting processes for the EDM Program that
flow from the EDM Strategy, including communication, business cases, and
funding.
[1] A previous TDAN.com column – “Piloting the Plane – EDM Program and Project Management”
[2] George Carlin, on mortal sin: “Ya had’ta WANNA! In fact, WANNA was a sin all by itself. “Thou Shalt Not WANNA”. If you woke up in the morning and said, “I’m going down to 42nd street and commit a mortal sin!” Save your car fare; you did it, man!”
[3] You won’t be surprised that I recommend conducting an EDM Assessment based on the Data Management Maturity (DMM)SM Model as the quickest method to precisely pinpoint the state of EDM practices.
[4] See the Data Management Maturity Model’s list of 25 process areas, and the Knowledge Areas in the Data Management Body of Knowledge to ensure completeness – note that the DMM focuses on fundamental EDM practices, while the DMBOK also includes solution areas (e.g. content management).
[5] In a subsequent column, we’ll discuss data management metrics – why you need them, how to formulate them, and how to report on them.
