Data Professional Introspective: Piloting the Plane

An organization’s Enterprise Data Management Strategy, which establishes an EDM Program, requires a corresponding sequence plan (aka, ‘roadmap’) describing initiatives and estimated timeline of the programs and projects it will undertake according to priorities aligned with the business strategy.

I’m in the roadmap business. A DMM Assessment provides detailed scoring, specific improvements by process area, and a series of initiatives (projects) over a multi-year duration, organized and sequenced into topic tracks (e.g. data architecture, data governance, etc.). The overall goal of the roadmap is to enable rapid EDM capability building while reaping short and long-term business value.

This post features Guests Bob Seiner, Todd Henley, Jeff Wolkove, and Sully.

Some proposed initiatives may have dependencies, and some can be accomplished independently. Prioritization reflects many factors and sources of input: affirmations of workshop participants and interviewees, the snapshot score against the DMM’s practice statements, the information firehose of detailed examples, problems, aspirations, resource constraints, organizational factors, the funding model, and the current presence (or absence of) a data management organization and operationalized data governance.

In crafting the best path for an organization to achieve rapid EDM progress, there are two primary considerations:

  1. What the organization wants to become, that is, what it needs to be, do, build, or have to realize its data management goals, and
  2. Determining the shortest journey to fix persistent data problems.[1]

When these foundational considerations are integrated with planned and in-flight initiatives, viola! An EDM roadmap, which can be expanded into an EDM Strategy with principles, rationale, resources, roles, budget, leadership designations, and integrated planned accomplishments.

However, it has become apparent to me that when explaining the rationale for recommending ‘do these initiatives, in this order’ as the Best Way Forward, I’m operating under an assumption about the organization’s program and project management capabilities, which is that the organization has the structure, culture, resources and leadership needed to Get Things Done.[2]

Mea culpa here. As we often do with core beliefs, I’ve reflexively assumed that every organization can plan, execute and integrate complexities and dependencies – of tasks, work products and organizational / resource factors. Probably from many years spent focusing on enterprise level programs.

As Chesley Sullenberger (Sully the heroic pilot) states, “In so many areas of life, you need to be a long-term optimist but a short-term realist.” While recommended initiatives may indeed be actionable, practical and achievable, proven by successful outcomes in numerous organizations, the context of the roadmap assumes “You can do this!”

And as oft-quoted Peter Drucker said, “Efficiency is doing things right; effectiveness is doing the right things.” You can tell an organization what to DO, based on detailed facts, analysis, and collective consensus, but if it hasn’t previously accomplished similar work, such as an IT Strategy or Enterprise Architecture transition plan at the enterprise level, or succeeded at data-centric efforts, such as implementing a metadata repository or a master data management solution on the project level, then adopting a strategic approach, thoughtful planning and an efficient way of working may be elusive.

This column addresses some key factors that organizations should consider to improve their management of EDM programs and projects:

  • Accountability
  • Leadership
  • Culture
  • Domains
  • Education
  • Funding
  • Creativity

Let’s start with observations from helping clients and students to plan, structure, implement and perform data governance, from Bob Seiner of KIK Consulting, Publisher of – “Data is not going to manage itself” and “Someone has to be in charge.

Click the image to see it larger.

The first statement goes without saying – it’s the reason why data professional positions continue to multiply, now that more organizations are clued into this fact. We’ll circle back to it in the context of Domains. The second statement leads us into a discussion of Accountability. The organization needs to decide WHO is going to be accountable for execution, control, and results, for both project and program level initiatives.

At the enterprise level, organizations are advised to ensure that a senior executive accepts the role of overall EDM program accountability. An important aspect of this role is to effectively communicate the EDM Strategy, plans, and milestones with executive peers and update milestone reporting regularly, because broad visibility is essential for sustained success. An enterprise Project Management Office (EPMO), if operationalized, can provide the methods, skills, and resources to ensure that the organization, mechanics of coordination and accountability designations across the EDM program are on track and reported.

Key initiatives within the EDM program, for example, building an authoritative data source through consolidation of multiple shared data stores, need a strong business owner (e.g., a program sponsor) and a strong IT program manager. Because there are many stakeholders and actors, milestone, task and resource coordination require active monitoring, adjustments and issue resolution. A certified Project Management Professional (PMP) or an individual with extensive project management experience should assist the business owner and IT program manager to allocate resources, structure, and schedule tasks.

At the project and task level, for example, creating a Business Glossary for customer master data, there are also multiple actors: data stewards, the business data experts they consult, the data management staff capturing the output, possibly the metadata repository staff, etc. Once again, a skilled project management resource is advised to produce the schedule, formalize assignments, and help the team monitor progress.

Even at the data working group level, for example, a small group of knowledgeable staff combining their data expertise to analyze data quality issues their business line cares about, someone should be selected as the facilitator, the group should agree on task objectives, internal deadlines, the criteria for stating ‘we are done,’ how results will be reported and to whom recommendations will be presented.

Remember, staff time is the most precious organizational resource. Executives and managers in most organizations have now ‘gotten religion’ about the criticality of data, and they are the ones to authorize staff time. Let’s make sure they get regular doses of ‘answered prayer’ to strengthen their faith – e.g., the customer data IS improving, the analytics teams ARE happy with the new well-documented authoritative data source, etc. All data-related efforts should be communicated, successes promoted, and participants recognized – if this is neglected, funding won’t be as available and the desired “data culture” will be very slow to blossom. Act like you’re from Missouri, the “Show Me” state.

Leadership for EDM at all levels comes with extra challenges, because typically there are numerous data producers and consumers, therefore many stakeholders who have an interest (and an opinion). If an organization has operationalized multileveled data governance, this challenge can be addressed more easily as areas of influence and decision-making authorities have been established, but whatever the case, the essential first step in leading data management programs and initiatives at every level is to identify relevant stakeholders (where ‘relevant’ equates to ‘having a major role’ versus ‘all end users’).

But first, the essential ingredient – an EDM program executive needs to be a visible and active advocate for the transformation, and the anointed messenger, publicizing the cornucopia of benefits that the organization will realize from well-organized, accessible, high quality governed data assets. As Todd Henley, head of Enterprise Data Governance for American Electric Power has discovered, “You need a champion, more than one if you can get them.”

Executive opinion shapes your reality, so you need to employ opening-night-level persuasive powers to make the case and achieve 1) enthusiastic adoption and 2) achieve active and convincing executive sponsorship. Without that, EDM efforts will languish and you’ll be stuck with a few impassioned preachers without a congregation. Therefore, nurture any embers of interest, make hard-hitting business cases, perform detailed analysis to demonstrate ROI, and essentially, do whatever it takes to encourage the sponsor and their peers to catch on fire.

Let’s say an organization is about to launch a program-level initiative, modernization of a product data warehouse; data producers might be Product Research and Product Development, and consumers might be Sales, Marketing, Inventory, eCommerce, etc., – lots of internal customers, right? An immediate task is to determine:

  • Who should be in the working sessions to define and approve product master data, product release data, product availability data, etc.
  • When the program plan should schedule activation of these resources
  • What the scope of their input is likely to be.

It’s advised to first identify all the data producers, and then engage the data consumers.[3]

Click the image to see it larger.

A recommended approach for a similarly scoped effort would be to construct a RACI matrix including both the individuals making up the ‘Core Team’ (producers) and the ‘Extended Team’ (consumers and other key stakeholders). To the extent that data stewards are already recognized, resource selection will be simplified. If not, you’ll have to parse through the org chart and send individual requests to managers to grant permission for staff time. It may also be useful to outline an operating (or, interaction) model, illustrating the RACI roles with respect to the program. Questions you can pose to determine who should be on the extended team include “Does this person have approval authority at any level of decision-making?” “Is this person important for an in-scope business process?” or on the downside, “Who will yell loudly if they’re not happy?”  

What if you’re working with multiple organizations? How do you get them to work together and agree? Let’s hear from Jeff Wolkove, Arizona Strategic Enterprise Technology, the data management leader for the State of Arizona, who has established a State-wide EDM program. Without State-wide mandates pertaining to EDM, he had to find another path, and concluded ‘Driving change through implementing policies is key.’ He convened an inter-agency governance group of interested parties in multiple agencies, and they drafted a comprehensive Data Governance Organization Policy featuring recommended governance levels, role descriptions, activities and a corresponding data management function. Policy statements referenced best practices from the DMM and the DMBOK. After an agency-wide comment period, the policy was refined, published and promoted. As established policy, compliance measures are included in agency audits – now there’s a stick to complement the carrots. He has since championed a Data Governance Operations Policy, an Interoperability Policy, and a Data Technology Policy. A Data Quality Policy is currently undergoing comments.

Another EDM leadership challenge is management of contractor staff. Remember that Julius Caesar’s army included large numbers of mercenaries, and Rome employed mercenary soldiers for over a thousand years. You need to lead and motivate contractors, even more important if your organization has a high percentage of contract staff working on data projects. If you hold the attitude that you can simply demand extra hours from the contract staff, you risk substandard performance. Don’t fall into the trap of simply relying on their professional integrity if you want sterling results. Sully stated “In the cultures of some companies, management depends heavily on the innate goodness and professionalism of its employees to constantly compensate for systemic deficiencies, chronic understaffing, and substandard subcontractors.” Hold the friendly fire (which kills you just as dead), be encouraging, and recognize stellar performance by both staff and contractors frequently.

Complexity, complexity, complexity – another big EDM management challenge. How do we address that? Through Education. We’ll use the product data warehouse example to illustrate the many connections, applying a series of questions that need to be answered, and aligned them with data management processes in the DMM as an organizing framework:

  • Have we defined and engaged all relevant stakeholders? (Governance Management)
  • Have we planned and organized the information and approvals needed from stakeholders? (Governance Management)
  • Have we defined the product data scope clearly? (Architectural Approach)
  • Are all business terms within the scope defined and approved? (Business Glossary)
  • Have we defined the metadata to be captured, stored, and made accessible? (Metadata Management)
  • Have we defined use cases and reporting requirements? (Data Requirements Definition)
  • Have we defined the data representations for the harmonized, normalized data set? (Architectural Standards, Data Integration)
  • Have we mapped the data sources to the new warehouse structures, and decided how we’re going to archive aged data over time? (Data Lifecycle Management)
  • Do we have a plan for assuring data quality in the new warehouse? (Data Quality Strategy)
  • Have we profiled the data that will populate the warehouse? (Data Profiling)
  • Have the producers and consumers identified quality issues, quality targets and thresholds, and defined quality rules? (Data Quality Assessment)
  • Do we have an efficient plan for cleansing the data? (Data Cleansing)
  • Have we defined and evaluated all relevant data sources? (Provider Management, Data Integration)
  • [We could go on, but you get the idea – many data management disciplines must be performed and integrated, and most of them are dependent on others].

This exercise should demonstrate the need to educate your management and participating stakeholders on the interconnected disciplines required for an organization to build trusted, easily integrated and accessible data assets. For example, if a new data steward isn’t aware of the importance of defining and approving the in-scope business terms, which will be codified for years in the new warehouse, you just can’t expect them to engage actively – that is, make it their business to be collaborative, to wrestle with refined distinctions, and to represent their business line’s data interests fully.[4]

Training should be tiered, for example ‘data awareness’ for everyone, data steward training for people putting their shoulder to the wheel, domain leader training for the domain owners and business sponsors. It can be instructor-led, computer-based, or even a series of informal short videos. It’s important to provide students with the broader context as well as role-specific education.

Which brings us to Culture. Organizations which have carefully fostered a collaborative way of working have an advantage in motivating staff, sustaining participation and achieving their EDM goals. A few selected culture categories are pertinent here:

  • People-Oriented – employees’ job satisfaction is a high priority, staff members are empowered, and usually respond accordingly with loyalty and dedication
  • Team-Oriented – employees often work in teams and intersecting teams, creating resilient relationships within working groups
  • Hierarchical – employees need to conform to uniformity of process, reporting and decision structure, creating stability but less encouragement, more red tape, and less agility, and innovation.

The first two culture types have a stronger foundation for agreement, consensus, and group actions. For example, a large financial organization with a self-described “consensus culture” was able to operationalize and expand data governance without much fuss. Employees understood the importance of their data and agreed that both producers and consumers should be engaged in their major data warehouse transformation initiative. Their governance has become highly effective and is embedded across the organization.

For data governance to succeed in its three main functions – building, nurturing[5], and controlling the data assets – you have to find a way to win the hearts and mind of each individual participant, as well as lead, facilitate, and reward collaborative analysis and decisions.

Hierarchical cultures may have difficulties in assembling core teams and extended teams, may not have recognized data stewards and data owners, and may incur decision delays as agreements work their way up the management chain. If the prevailing culture is hierarchical, the organization should attempt to assure that its data management organization and governance groups overcome this norm by active executive encouragement of the team approach and empowering more peer group decisions about data (And immediately stamp out any ‘shoot the messenger’ attitudes).

Dear to an enterprise data architect’s heart is the concept of Domains. Categorizing, organizing, and governing data by domain is an enormous achievement for an organization, turning the organization’s customary thinking from systems-orientation to data-orientation. This is analogous in degree of difficulty to executing a hairpin turn on the Gerald Ford aircraft carrier (100,000 tons). Massive – huge – an outstanding accomplishment![6]

Domains are depicted in an organization’s or business area’s conceptual data model. We can think of domains as primary subject areas / data groupings, information about which is essential for performing core business processes. For example, for Human Resources, Employee is a key domain. In the average organization of any size – there are hundreds, perhaps thousands of data elements about employees – demographic information, start and end dates, family members, benefits, salary, position, reviews, etc. – and probably many system data stores and repositories in which Employee data resides.

Success IS possible. One large bank categorized the data in 5000+ data stores into more than 15 domains, aligned application data stores with those domains, and assigned Domain Leaders, designated accountable and responsible individuals, with the mission of directing staff to analyze and harmonize the data for construction of authoritative data stores, refine and approve the designs through data governance, and identify specific applications within the domain for consolidation, migration or retirement. If they can do it, so can your organization. Take the big leap, jump off the 30-foot diving board, win the big fights!

I’m going to repeat Bob’s comment – “Data is not going to manage itself.” A commitment to managing and governing data by domain is not only a key success factor for EDM programs, it solves many problems in the management of EDM programs and projects. In our Product data warehouse example, the program manager would have a much easier time identifying the domain owners, data stewards and physical custodians for the data sets; the Core Team and Extended Team participants and organization of these groups would also be logically assigned.

Funding for data management programs and projects has historically been a challenge, insofar as data wasn’t viewed as a critical infrastructure asset. Although this has improved in general, some organizations have really stepped up and made big changes to realize their future.  As Bill Gates stated, “How you gather, manage, and use information will determine whether you win or lose.”

What I observe is that although there is often lip service paid to the principle that data is extremely important, organizations may not yet be willing to take the next step – funding a centralized data management organization, just as they now provide non-discretionary funding for Finance and Human Resources – essential functions that will always be performed for the life of the organization.

An organization’s funding model can be a big inhibitor of EDM funding; if the technology and data budget is only allocated by, for example, two categories, ‘capitalized’ and ‘maintenance,’ data programs and projects begin, execute, and end. There is no organizational unit responsible for pushing for reusability, persistent products are insufficiently maintained (e.g., the metadata repository, the data quality toolset, etc.) governance participants are not recognized, and there is no easy way to determine who should lead, who is accountable, who is responsible, etc. In addition, there is little recognition of the short and long-term value that knowledgeable data professionals provide, hence they are always attempting to herd cats, chasing down unbelievers in the cafeteria, and incurring undue stress in their under-resourced and under-funded attempts to Get Things Done.

Organizations, the future is catching up with you – your steady-state spend on data is constantly rising and overwhelming innovation, and your analytics suffer from undocumented, unavailable, poor quality data. New technologies can maybe save you from drowning, but they cannot sweep you back to the shore. E.g., you can purchase an AI platform, but if the data is terrible, you’ve wasted your money and it will fail of its promise. Open the check book, rescue your data from the shelter, and adopt into a loving home (your own revitalized organization).

But what if, despite your efforts to persuade, influence, beseech and beg, there is still little or no funding to accomplish vital data management programs and projects? For instance, many organizations hire a Chief Data Officer because their peers have done so – and often this individual is highly accountable yet has no staff and no line management authority. Candidates, don’t be tempted to accept this position, you’re likely to become a scapegoat, hence the often-repeated pattern ‘Joe, you’ve been here two months now, why is our data still bad?’

If you’re stuck with ‘EDM on a shoestring’ that’s when you need to dig deep and free your inner Creativity. As Jeff Wolkove learned, suceeding in substantive efforts across the State of Arizona despite lean budgets, “Don’t lament what you don’t have – work with what you do have.”

To be innovative, take advantage of situations without much funding when you’re still expected to accomplish something meaningful. You need to learn to function effectively within your sphere of influence.

This is where moral persuasion of colleagues and peers takes center stage. In Jeff’s example, his annual EDM conferences established community networking and raised interest, which enabled him to establish an EDM Steering Committee. Sponsored sessions of instructor-led training equipped the nucleus of data champions from many agencies to understand EDM as a whole and apply key concepts and project plans to their own organizations. Hence, this coalition of the willing was motivated to participate in drafting policies, provide detailed comments, and analyze comments to refine and approve the final policies applicable across the State.

Leveraging communications best practices, Bob Seiner coaches his client organizations to develop data governance communication plans in three areas: Orientation to raise awareness; Onboarding to convey skills, methods, roles and activities learning; and Ongoing to track decisions and achievements. Similarly, the DMM’s Communications process area outlines essential components calls for an EDM Communications Strategy with roles, frequencies, and channels defined, applicable when launching a program and important to ensure ongoing engagement and interest, because the data is forever, and you will always need sustained active participation from your stakeholders and governance participants.

And finally, let’s not forget the tried and true foot-in-the-door effort, the pilot project. For example, with only a person or two, you can conduct a data quality pilot and create a process and templates that can easily be reused for any data store across the organization. See my TDAN column “Boot-Strapping with Jet Packs: Accelerating Enterprise Data Quality” for a step by step project that delivers tactical improvements while setting the stage for organization-wide use. From small beginnings, you can accomplish great things. As the iconic blues guitarist Mississippi John Hurt urged, “Make it sound like a whole band!”

[1] The inextricable polarity of ‘aspirations/problems,’ e.g. ‘get me to the promised land and heal my ills.’

[2] Mitigating this tendency is the galvanizing effect of a DMM Assessment; it most often spurs action and commensurate funding and resources.

[3] There are many examples of how complex this can be. For instance, for one customer master data implementation, the business sponsor didn’t have any idea how many systems created the customer [yes, there were multiple creates] or modified customer data. The final tally was seven producers and twelve consumers – think of all the knowledge workers that represents, each with their own use cases, rules, representations and values. Yikes!

[4] EDM training is so important. I’d love to expand on this topic here, and I know Bob, Jeff and Todd would too, but I’m going to sacrifice to avoid an excessive word count. A future column on EDM education is on the horizon.

[5] Aka, ‘sustaining,’ in a caring and devoted manner.

[6] In terms of the DMM, an organization cannot achieve Level 3 – Defined if it has not made this commitment.

Share this post

Melanie Mecca

Melanie Mecca

Melanie Mecca, CEO of DataWise Inc., an Enterprise Data Management Council Partner and Authorized Instructor for DCAM certification courses, is the world’s most experienced evaluator of enterprise data management programs. Her expertise in evaluation, design, and implementation of data management programs has empowered clients in all industries to accelerate their success. As ISACA/CMMI Institute’s Director of Data Management, she was managing author of the DMM and has led 35+ Assessments, resulting in customized roadmaps and rapid capability implementation. DataWise provides instructor-led courses leading to DCAM certification, as well as for data stewardship and the proven Assessment method that she developed. DataWise offers a suite of eLearning courses for organizations aimed at elevating the data culture and providing practical skills to a large number of staff. Broad stakeholder education is the key to data management excellence! Visit to learn more.

scroll to top