The Enterprise Data Management Organization
In working with both commercial companies and Federal and State agencies, I’ve found that many organizations share a common structural deficit: they don’t have a centralized data management organization that reports either to the business lines or a ‘neutral’ organization (e.g. Finance, Risk, etc.). This demonstrates lack of commitment to improve management of the data assets; in effect, not walking the ‘data culture’ talk.
Referring to the Data Management Maturity (DMM) Model for perspective, it was developed specifically to target the two biggest challenges of managing data as a critical asset: (1) to engender increased engagement from the lines of business, those who create and manage the data generated by their business processes; and (2) to help organizations rapidly develop a successful enterprise data management program.
With that emphasis, the Data Management Function – a vital Process Area in the DMM’s Data Management Strategy category, containing specific practices that demonstrate capabilities – naturally receives a lot of attention when assessing the current state of data management in an organization.
It’s all about progress. Is the organization committed to enhancing its management of data, and is it willing to enhance the org chart to do so? If not, the saying “If you do what you’ve always done, you’ll get what you’ve always gotten,” applies.
In many organizations, the de jure (or de facto) data management organization (DMO) is a group within Information Technology. That is likely to result in the following situations:
- The data management staff is assigned to discrete application development projects, one after another – applied knowledge but no path to improved data awareness in the business.
- Even ‘enterprise data’ projects, like Master Data Management, may be run by one line of business, risking data completeness and accuracy.
- The staff does not have the mandate to work with governance groups, depriving data stewards of important knowledge about sound data practices.
- The staff works on quality issues, metadata, or standards on their own time, creating the ‘hero syndrome’ leading to burnout.
- There is no central location or mandate for policies, processes, and standards which apply to the entire organization.
- The lines of business may value an individual’s work, but do not perceive the data management organization as an equal partner because IT is viewed as a service organization.
- The staff does not have the mandate or resources to educate the business data experts in their responsibilities.
Those are just a few examples. There are many more lost opportunities that may result from the DMO’s placement solely within IT. The importance of the Data Management Function (implemented as the DMO) cannot be overstated. The organization’s data assets, without counterbalancing forces, tend to increase in number, complexity and costs.
Persistent data products affecting, and used by, the entire organization – such as the metadata repository, authoritative data source designations, the business glossary, the enterprise data model, data policies, processes, and standards – need a designated organization to be the shepherd, advocate, and custodian. In addition, the decisions and activities generated by data governance groups need to be monitored and integrated through a permanent function, accountable to the organization. To quote Benjamin Franklin, “We must hang together, or we shall assuredly all hang separately.”
A DMO can provide many benefits to the organization and its value increases over time. Examples include:
- Creates increased organizational cohesion by leveraging stakeholders’ knowledge and experience to develop and approve enterprise policies, processes, standards, and guidelines.
- Lowers costs through efficiencies in executing standardized data management activities.
- Achieves data quality improvements through sound approaches and methods for activities that combine to: define shared business terms; develop data requirements, integrate data, define consistent metadata, develop consistent quality rules, etc.
- Accelerates business value by improving the data layer and metadata for accurate reporting and advanced analytics.
In terms of the DMM, the lack of an effective DMO prevents the organization from implementing enterprise-wide sound practices needed to achieve Level 3 (Managed). Conversely, a well-planned and staffed DMO accelerates and integrates progress in data initiatives and increased business line engagement. To assist organizations in accelerating their progress, we propose beginning with a ‘core’ or ‘Starter’ DMO.
If the organization has the will, this structural deficit – no data management organization – can be rapidly addressed. A DMO weaves a cohesive tapestry of roles, responsibilities, and assignments which benefit all business lines. The ‘starter DMO’ offered here has been successfully implemented in multiple organizations and has also been applied to the launch of new Chief Data Officer organizations.
And another consideration, data governance – governance groups that are the complement to the data management organization, representing implementation of formalized agreements and decision-making across business lines. I’ve observed that most organizations lacking a DMO also have yet to establish effective data governance. The starter DMO organization structure, which can be implemented in any company or agency, can greatly assist in launching governance. If governance is already implemented, the DMO will enhance its activities, serve as ‘integration glue,’ and facilitate achieving its objectives.
The assumption underlying the ‘starter DMO’ is that the organization either has or is about to engage in strategic planning for its (nascent or existing) EDM program. The core organization is intended to be plug and play; it functions as a backbone to anchor the EDM program’s key persistent products and as a permanent advocate for sound approaches and practices. The bullet points indicate key tasks to be led by each individual position in the DMO.
Data Management Organization
Let’s unpack this diagram further, with role descriptions and corresponding responsibilities:
- DMO Director – This individual needs to be knowledgeable about the overall business functions of the organization, have significant data knowledge of key corporate applications and data stores, and possess established relationships with business stakeholders. In addition, the Director should have experience in project management, a history of driving to results, and a track record of facilitating collaboration among disparate groups. The Director will: facilitate and track action items for the executive governance committee; facilitate creation of the data management strategy; develop a data management communications plan; drive the creation of policies, standards, and processes; develop and report on data management metrics; institute a data quality program; and lead creation and maintenance of the DMO web site and process asset library (data management work products, such as policies, processes, standards, guidelines, and training).
- Data Governance Lead – This individual needs to be knowledgeable about the major data groupings (subject areas) supporting the business lines, have excellent communication skills, experience in planning and running meetings, and the confidence to encourage individuals to respond to requests. If governance is not yet implemented, this individual will initially be responsible for developing the data governance charters for the executive governance committee and the data stewards group(s). The Data Governance Lead will facilitate the data steward group meetings, be the central point of contact for stewards to raise issues for decision, and track decisions and action items. By applying the timing, scope and schedule specified by the Data Management Strategy, the governance lead will facilitate the definition of business terms for the business glossary and facilitate planning the metadata strategy with data governance.
- Data Quality Lead – This individual needs to have understanding and previous experience with data quality elements, such as data profiling concepts and plans, data quality dimensions, and facilitating development of quality rules. Responsible for facilitating creation of: the data quality strategy; sound data quality practices; selection of recommended toolsets; and developing data quality policies, processes, standards and reusable templates. The data quality lead works with data governance to stand up working groups for priority projects, establishing data quality processes, and assisting in development of quality rules. In addition, the lead will develop and conduct training for business data experts in the application of data quality dimensions, and work with governance to establish RACI for quality functions.
- Policy and Standards Lead – This individual needs to have experience in developing business processes and understand data management disciplines. Responsible for facilitating analysis of current best practices employed by existing programs and projects (e.g., discovery for best practices and templates); developing a data management policy based on the data management strategy and data quality strategy; definition and documentation of data management processes and work products, approved by governance; development of compliance for those processes; collaborating with IT to ensure that processes harmonize with the systems development life cycle; and development of training for processes. Overall, the policy and standards lead will run point for building a body of work that supports current and future management of the data assets, leveraging projects, and program-level efforts.
Starting from the Beginning
If an organization seeks to implement a core DMO, the following is an initial set of sequential activities that can be adapted as needed:
- Establish the DMO with core staff, reporting to an actively engaged executive.
- Determine the scope of enterprise data that the DMO will address.
- Identify a pilot program for implementation.
- If a governance group for the pilot program is not in place, create a charter and stand it up.
- Determine the most important data management needs for the pilot program.
- Create processes to support this program (e.g., data quality, business glossary, metadata, data standards).
- Obtain and analyze best practice examples across the organization – policies and processes addressing governance, metadata, data quality, etc.
- Refine and apply these to the pilot program.
- Develop a plan to extend to the organization in phases, informed by business priorities and the data management strategy.
This approach ensures that progress will be rapid, because the pilot program provides urgency and the DMO is employing its skills, knowledge, and improved practices to deliver direct benefits. Over time, the DMO can gradually extend its leadership and scope enterprise-wide.
 ‘Enterprise data’ typically refers to data shared by multiple business lines. It may also include: 1) data critical for important business processes, 2) data critical for analytical insights, and 3) data required for regulatory reporting. Defining the scope of enterprise data is addressed in the DMM through the Data Management Strategy Process Area.
 Note: IT was the origin of the data management function, which started out centered around data architecture, data standards and compliance, branching out to classical data warehousing, metadata and data quality capabilities. Over the decades, however, IT has lost ground in its perceived stature and has rarely succeeded in becoming an equal partner with the business. Since the imperative is to succeed in gaining commitment and agreements from the business lines about data, the DMO benefits from being aligned with them in its organizational placement.
 A pilot program/project for launching EDM processes should have the following characteristics: 1) it is an organizational priority; 2) it is about to start or in the beginning phases; 3) it has an active executive sponsor; 4) time and budget allow for establishing sound practices which can later be refined, abstracted and applied to the entire organization. Examples of suitable programs may include: redesign of an enterprise data warehouse; implementation of master data management; redesign of a key operational system with multiple suppliers and consumers.