Data Stewardship (2014)

CHAPTER 1 Data Stewardship and Data Governance

How They Fit Together

This chapter discusses the definition and deliverables (including policies, procedures, and processes) for a Data Governance program. The structure of the organization is discussed, as well as the roles and responsibilities of participants in the program, including a brief list of stewardship tasks. The key role of Data Stewardship in a Data Governance program is discussed, as well as how Data Stewardship fits into the overall program.

Keywords

Executive Steering Committee; Data Governance Board; Data Governance Program Office; asset; business rule; policy; procedure; process

Introduction

Data has become so vital to the success of almost all organizations (both large and small) that many are attempting to implement a Data Governance program. When done properly, it provides the means to manage the overall collection of data (i.e., the data asset), including the structure, processes, and organization needed to manage key data elements. A vital component of a Data Governance program is Data Stewardship. However, to understand the role that Data Stewardship plays in Data Governance, you need to understand the overall Data Governance program itself, including the purpose, deliverables, roles and responsibilities, and value that Data Governance adds to the company. You also need to understand how Data Stewardship interacts with other aspects of the Data Governance program, and what part Data Stewardship plays in the success of the program.

This chapter defines Data Governance and Data Stewardship, and explains the structure and inner workings of the organization needed to support the effort. It details the responsibilities of each type of participant, including executives, Data Governors, Data Stewards, members of the Data Governance Program Office, and IT support staff. It also provides a “target”—the end result of what a good Data Stewardship effort can and should achieve.

What is Data Governance?

Ask a room full of Data Governance practitioners what Data Governance means, and you’ll probably get as many definitions as there are people. One of the best definitions I have ever run across comes from my friend Gwen Thomas of the Data Governance Institute (DGI):

Data Governance is the exercise of decision making and authority for data-related matters.

It’s a system of decision rights and accountabilities for information-related processes, executed according to agreed-upon models which describe who can take what actions with what information, and when, under what circumstances, using what methods.

The key thing to take away from this definition is that the practice of Data Governance has more to do with establishing the roles and responsibilities about how people manage and make decisions about data than about the data itself. That is, Data Governance—and Data Stewardship—is all about making sure that people are properly organized and do the right things to make their data understood, trusted, of high quality, and, ultimately, suitable and usable for the enterprise’s purposes.

Note

It is not the purpose of this book to fully define and discuss Data Governance. Instead, this book focuses on the practical details of Data Stewardship, which is a necessary component of Data Governance. For an excellent reference on Data Governance, please see Data Governance: How to Design, Deploy, and Sustain an Effective Data Governance Program by John Ladley, published by Morgan Kaufman.

What is Data Stewardship?

Data Stewardship is the operational aspect of Data Governance, where most of the day-to-day work of Data Governance gets done. According to Danette McGilvray:

Data Stewardship is an approach to Data Governance that formalizes accountability for managing information resources on behalf of others and for the best interests of the organization.

As Danette also notes in her book Executing Data Quality Projects (Morgan Kaufmann Publishers, 2008), a steward is someone who manages something on behalf of someone else. In the case of Data Stewardship, the “someone else” is the business function that owns the data, represented by the business function’s representative on the Data Governance Board (Data Governor).

Put another way, Data Stewardship consists of the people, organization, and processes to ensure that the appropriately designated stewards are responsible for the governed data.

Data Stewardship is crucial to the success of Data Governance (which, in turn, is crucial to the success of data management). This is because it is through Data Stewardship (and the Data Stewards) that all of the metadata (definitions, business rules, and more) is collected and documented. In addition, having stewards who are responsible for the data, as well as having a set of procedures that require the stewards to be consulted on decisions about the data they steward, helps to ensure that the decisions are based on knowledge and are made in the best interests of all who use the data. The combination of designated stewards, processes, and a mission to manage data in the best interests of all leads to the data asset being improved in quality, and to it being used to drive competitive advantage and regulatory compliance for the business.

Overall Goals of Data Stewardship

So, what does “good” Data Stewardship look like? That is, what goals should a Data Stewardship program be striving for—what are we trying to achieve? The goals should align with the highest level of maturity (optimized) discussed in Chapter 9 and should include the following:

- A smoothly functioning Data Stewardship Council.

- Policies and procedures are in place, and have become part of the corporate culture.

- Business Data Stewards are designated and participating from every business function that owns data. There is no participation from business functions that do not own data. Data Governance and Data Stewardship includes outside business partners.

- Technical Data Stewards are designated from all enterprise applications, data stores, data warehouses, data marts, and ETL processes.

- Data Stewardship involvement is integrated into enterprise processes, such as project management and system development methodology. Data Stewards are viewed as an integral and necessary part of data management.

- Clearly defined roles and responsibilities for all Data Stewards and ratings for how effectively those roles and responsibilities are performed are built into the steward’s compensation objectives.

- The responsibility of all employees for the management of data is accepted as part of the corporate culture.

- There is executive support and endorsement for Data Stewardship. Executives are publicly supporting the policies and actively promoting adherence to them, as well as the creation of procedures to implement the policies.

- There is clearly defined and recognized value being added by Data Stewardship.

- Key business data elements are identified and defined and business rules determined, and all are linked to physical instances of the data. Where appropriate, the data has been profiled to understand and correct data quality.

- Data Stewardship decisions are clearly documented and published to interested parties using sanctioned communication methods.

- Training for all involved parties (including stewards, project managers, and developers) has been written and is being given on a regular schedule.

- Supporting tools (e.g., a metadata repository, business glossary, central issue log, and data profiling tool) are installed, supported, and in regular use.

- Innovation in maintaining the vision of data quality and remediating data issues is encouraged, as well as creativity and competitive advantage in using high-quality data.

- Management and the Data Governance staff keep abreast of important emerging trends in data management and adapt accordingly.

- Processes and procedures are written, approved, and in use to:

Identify key business data elements

Collect, review, and approve business metadata for the key business data elements

Log, analyze, prioritize, and remediate data and data quality issues

Support projects

Manage domain data

Replace Data Stewards as necessary

Review and analyze data quality improvement opportunities

Document stewardship work done and methods for publication

Moving Data to a Governed State

At its simplest level, the purpose of executing on Data Governance is to move data from an ungoverned state to a governed state. Ungoverned data refers to most of the data that an enterprise has, at least at the beginning of a Data Stewardship effort. It is rarely defined, its quality is unknown, its business rules are nonexistent or conflict with one another, and no one is accountable for the data. Governed data is data that is trusted and understood and for which someone is accountable for both the data itself and for addressing issues about the data.

In addition, fully governed data means that you know all of the following at the data element level:

- The standardized business name of the data element. This is the standard business name by which the data element is called (ideally) everywhere in the company. Where a business unit has a need to call it something else, that alias is documented.

- The standardized business definition of the data element. Just as there should be one standard business name, there should also be only one standard business definition for the data element. Where there is disagreement on the definition, the data owner must either change the definition or the errant data element must be defined and given a new name.

- For calculated or derived data elements, the calculation or derivation rule. The rule needs to be very specific so that there is no confusion about how a quantity or value is derived. As with the standard business definition, if there is disagreement, either the derivation rule must be changed, or a new data element (with a different derivation) must be defined.

- The physical location of the business data element in a database/system. The physical data element is essentially a mapping of the business data element to, for example, a column in a table in a database in a system, or something equivalent.

- The data quality rules, in context. This includes the rules that specify good quality (e.g., format, range, valid values, pattern, and so on), as well as the level of quality needed for each intended use of the data.

- Rules for creating the data element. These are the rules that must be followed before an instance of the data element can be created. When applied at the entity level (e.g., customer, product, policy, etc.) having and following the creation rules can ensure that all the needed data is present and of appropriate quality before the new entity (e.g., customer) is created. These rules are critical when implementing master data management to ensure that “half-baked” data is not created.

- The usage rules for the data element. Usage rules specify what purposes the data can or cannot be used for. For example, the patient’s name in a pharmacy might be fine to use for filling a prescription, but not for generating mailing labels for refill reminders.

- The Data Governor/Owner and Business Data Steward for the data element. These are the individuals who serve as the authority and decision makers about the data—the people who are accountable for the data. It is important to understand that a proposed change to any of the items in this list must be authorized by the Business Data Steward and approved by the Data Governor/Owner. Failing to do so can lead to unforeseen ramifications (see the sidebar “Using Data Incorrectly”).

In addition to determining and documenting the preceding bulleted items at the data element level, governed data is data to which the approved procedures and processes for change management and issue resolution are rigorously applied. That is, even if you know all the metadata for the data elements, if the governing processes and procedures are not used to manage the data, then the data is not “governed.”

Note

Both Data Governor and Data Owner are terms that are used interchangeably in many organizations, and are defined later in this chapter. From here on this book will refer to the middle layer in the Data Governance organization as Data Governors.Business Data Steward is defined in Chapter 2.

Using Data Incorrectly

A fairly well-known story about incorrect data usage comes from the early days of my career. A pharmacy at a major chain drugstore recorded patient names and used that data to fill prescriptions. However, having no place to put certain pieces of key data in the aging system, the pharmacists took to adding various characters to the end of each patient’s name to indicate certain conditions they needed to keep track of. For example, if the patient had no insurance, the pharmacist might append a “$” to the last name. Or, if the prescription was related to a worker compensation claim (the individual needed the medication because he or she had been injured on the job), the pharmacist might append “(WC)” to the last name.

All was well until someone decided to use that data to generate mailing labels for refill reminders. The problem, of course, was that when the mailing labels were generated and the letters sent out, all those character strings were printed on the labels, causing confusion and some angry responses from the customers. While we didn’t have a formal Business Data Steward for that data, there was a well-known expert in the company who was the unofficial steward. Had anyone checked with her first, she would have pointed out the issue and a plan could have been made to improve the data quality to the point where the data could be used correctly.

Take a look at Figure 1.1. At the beginning of a Data Stewardship effort, the bulk of the data is ungoverned, as shown by the outside rectangle. There are many drivers for getting data into the governed state, represented by the inner circle. Among these drivers are building a data warehouse, instituting Master Data Management (MDM), improving data quality, implementing information security, and migrating data into a new system. The role of Data Stewardship in these activities is discussed more fully in Chapter 5.

FIGURE 1.1 There are many drivers for moving data from an ungoverned state to a governed state.

While all of these efforts can be (and have been) attempted without Data Governance and Data Stewardship, they are much more prone to delay, poor-quality results, and, yes, utter failure. That is because without a clear decision maker about the data, a lot of guessing goes on, usually by IT personnel doing their best to figure things out and deliver on a deadline. For example, think about how many times a fruitful discussion has disintegrated into an argument about what a piece of data means, or how it is calculated. With an assigned Data Steward as the documented authority on that data, all you have to do is ask. The data authority can either provide the answer or take responsibility for finding the answer, relieving other people of a responsibility they shouldn’t have had in the first place.

The Three P’s: Policies, Processes, and Procedures

To have effective Data Governance and Data Stewardship, it is necessary to have the three P’s: policies, processes, and procedures. Generating these items is one of the key early deliverables from a Data Governance effort. Policies establish a set of goals and state “this is what we need to do” at the enterprise level. Processes (which can be represented by a process flow diagram such as a swim lane) state what is required to comply with the policies. A process specifies a high-level set of tasks, the flow of the tasks, and who is responsible for completing each task. Finally, procedures describe in detail how exactly to perform the tasks.

An example might help to illustrate the three P’s. The enterprise might have a policy that states that redundant data elements are not permitted. The process would then lay out the tasks and flow—for example, query the Business Data Steward, initiate a search through the business glossary by name, do a definition comparison, resolve potential duplicates, then create a new business data element. The procedures would then state exactly how you perform a search, what amount of overlap in the definition should signal a potential duplicate, how specifically you resolve duplicates, and exactly how a new business data element is created.

Common policies established by a Data Governance effort might include policies for Data Governance itself, data quality, data remediation, communication, and change management. And because policies require a high level of authority to give them “teeth,” they tend to be approved at the executive levels of the company.

Note

It is rare that the executives actually write the policies. Data Governance policies are typically drafted by someone in the Data Governance Program Office, often with input from Data Governors and Data Stewards. The executives are then responsible for approving the policies, and taking the necessary action when policies are violated.

Data Stewards don’t write the policies, but they must be aware of what policies have been written. In many ways, the policies drive what the Data Stewards are responsible for achieving and formalize the goals of Data Stewardship. That is, the processes and procedures are intended to carry out goals established by the policies.

The Data Stewards have a considerable amount to say about processes and procedures. Processes and procedures state how the work actually gets done, including what steps have to be taken and who needs to approve decisions. For example, when a data issue is raised, there is a process (or there should be) to analyze the issue, assign the responsible individuals, remediate (or choose to ignore) the issue, and get approvals for the decisions made from the Data Governors. It is the Data Stewards who do the work, going through the processes and procedures to reach an end result. Thus, it should come as no surprise that to meet the goals of the policy, the processes and procedures need to be designed, tested, and modified as necessary by the Data Stewards.

How Data Stewardship Fits into Data Governance

In a nutshell, Data Stewardship is the operational aspect of an overall Data Governance program, where the actual day-to-day work of governing the enterprise’s data gets done. The Business Data Stewards, working together in the Data Stewardship Council, perform the multitude of tasks detailed in Chapter 2. Quite frankly, without Data Stewardship, Data Governance is just a framework of good intentions that never gets implemented.

By now you are probably getting the feeling that the overall Data Governance effort involves several levels, as well as some support from the organization. And you would be right! There are a number of ways to organize a Data Governance program, but Figure 1.2 shows a fairly simple structure that works well in many organizations. It also illustrates the relationship between various parts of the program.

FIGURE 1.2 A Data Governance program is often arranged as a pyramid, with support from IT and a Data Governance Program Office.

The pyramid structure illustrates both the level of data responsibilities as well as the typical number of participants at each level. A relatively few number of executives provide the support, any necessary culture change, and impetus to drive the program. They may, on occasion, also make far-ranging decisions, such as changing how people are incentivized to drive data quality. The Data Governors represent their business function, appoint the Data Stewards, and make decisions based on the recommendations of the Data Stewards. The largest numbers of participants are the Data Stewards, who understand the use of the data, the impacts of changes to the data, and what rules must apply to the data. The Data Stewards create the metadata (definitions, etc.) and make recommendations for the Data Governors to act on.

Executive Steering Committee

Not surprisingly, the tip of the Data Governance pyramid is occupied by the Executive Steering Committee. I say “not surprisingly” because any program of the scale and potential impact of Data Governance needs executive support and a well-defined escalation path. The Executive Steering Committee:

- Drives cultural changes needed to treat data as an asset and manage it effectively across business-area boundaries. The previous example applies here too—treating data as an asset and making decisions that benefit the enterprise as a whole takes executive support and a careful balancing of priorities. In addition, the culture of the organization may not be immediately conducive to Data Governance. In many organizations, the concept of a Data Governor is foreign, and all decisions are made by consensus. But Data Governance requires that someone be ultimately both responsible and accountable (two different things) for the decisions about data, and that may well mean that consensus is not reached. Sometimes it gets to the point where you have to make some group unhappy to move forward for the greater good. The executives need to communicate that this is acceptable and expected as part of Data Governance.

- Makes necessary changes to the organization and tools as required for effective Data Governance. Organizational changes are necessary to implement Data Governance. A new headcount is usually required, at least for the Data Governance Manager, and other support personnel may be needed as the program matures. In addition, IT tools may need to be added, such as a metadata repository, business glossary, data quality profiling tools, and web-based collaboration software. Adding tools not only requires funding (for licenses and hardware) but the expenditure of efforts to specify requirements, evaluate vendor proposals, and install and maintain the new tools. Typically, organizational changes and early funding requires executive support.

- Creates and promotes the vision for the Data Governance program. With the changes needed to manage a successful Data Governance program implementation, employees will be looking to the executives to state what they want to happen (a vision) and throw their support solidly behind that vision.

- Authorizes the Data Governance Board budget. While not always possible, providing a budget for the Data Governance Board to manage enables that governing body to not only make decisions about important issues, but to fund the decisions that have a high-enough priority to warrant remediation. But in most companies, establishing a budget requires executive approval.

- Balances business priorities with operational needs across the enterprise. There can be conflicts between priorities established by a Data Governance effort and keeping normal operations running. For example, something as “simple” as standardizing a calculation for a derived business element may have wide-ranging impacts to operational systems. And while analyzing those impacts falls to the Business Data Stewards, balancing the benefit of the standardization against the modifications that may be needed to key operational systems is a decision that has to be made at the Data Governance Board level or, if of sufficient impact, by the Executive Steering Committee.

- Approves Data Governance policies. Policies establish a set of rules (and possibly penalties) for the enterprise. Data Governance generates a set of policies, and for these policies to have credibility and impact to the organization, they must be approved by the executives.

- Reviews, evaluates, and reports to executive sponsorship on the Data Governance performance and effectiveness. Ideally, the executive sponsor for Data Governance is a member of the Executive Steering Committee, but if that is not the case, the Executive Steering Committee needs to review and evaluate the progress reports (typically created by the Data Governance Manager and Data Governors) and provide a summary to the executive sponsor.

- Provides advice, direction, counsel, and feedback to the Data Governors (members of the Data Governance Board). There needs to be an open line of communication between the Data Governors and the executives, especially since the Data Governors may need executive support to drive the Data Governance program forward. The form of the feedback can vary, but frequently results from discussions of Executive Steering Committee meeting agenda items that are then relayed back to the Data Governance Board via the chairperson and meeting notes.

- Ensures decisions regarding the data support the strategic direction of the organization. The executives are in the best position to gauge the long-range needs of the organization, and evaluate decisions based on those needs. It may well be that the Data Governors are not even aware of upcoming initiatives, and thus their decisions could have a negative impact that they aren’t even aware of.

- Ensures active participation by the business and IT. The Data Governance Manager (who often chairs the Executive Steering Committee meetings) is in a good position to know whether the business and IT participation is sufficient for Data Governance to be effective. For example, terminations or reorganizations may leave a business function without adequate representation, or a new business area may need to be added to the program. In either case, the executives will need to ensure that the appropriate people are designated to have the necessary duties. Further, if the designated participants are not participating to the level needed, this situation will also need to be dealt with.

- Represents their business function’s direction and views in regard to the adoption and deployment of enterprise policies and practices. The executives represent different areas of the business, and must evaluate proposed policies and practices to ascertain how they affect the business direction. Where there is a conflict, decisions have to be made at the executive level to either change a policy or practice, or adjust a direction.

- Appoints a Data Governor from their business function(s). The executive representing a business function has the authority and the responsibility to appoint the Data Governors who represent them on the Data Governance Board.

- Resolves issues escalated by the Data Governance Board. It is reasonable to expect that most issues related to data usage and management will be made at the level of the Data Governance Board. On occasion, however, it may be necessary to escalate a decision up to the executive level. This can happen when one or more of the Data Governors refuse to budge on an issue, perhaps advocating a “solution” that safeguards the data usage by their own business area to the detriment of the enterprise as a whole. As an example, a frontline application (e.g., for writing policies in an insurance company) owned by one business area may be used to capture only the data needed by that area even though it is capable of recording other data needed by other business areas. If the owning Data Governor refuses to have the data collectors take the extra time to record the additional data, then this issue may need to be escalated.

Where Do Data Governance Participants Come From?

The participants at each level of the Data Governance effort are assigned by the level above, as illustrated by Figure 1.3. This starts with the business sponsor who provides the authority to request cooperation from line-of-business (e.g., sales, marketing, finance, and so on) executives who need to assign the Data Governors who will represent their portion of the business. The Data Governors then assign the Business Data Stewards to represent them in the day-to-day operations of Data Governance. If Project Data Stewards or Domain Data Stewards (explained in Chapter 2) are needed, these are assigned (and potentially hired or contracted) by the Data Governance Manager or Enterprise Data Steward.

FIGURE 1.3 Responsibilities for assigning Data Governance program participants.

On the IT side, there is an IT sponsor who provides the authority to request cooperation from IT application owners to assign Technical Data Stewards.

Data Governance Board

At the middle level in the Data Governance pyramid is the Data Governance Board. The board members (referred to as Data Governors or Data Owners) tend to be the primary decision makers, acting on recommendations from the Data Stewardship Council. The Data Governance Board makes the decisions about changing data usage, making data quality improvements, and (especially) making decisions that require resources (both funding and personnel) to be expended to make a change or remediate an issue or impact. To be more specific, the Data Governors prioritize issues to be remediated (or approve the priorities assigned by the Data Stewards) and assign resources to work on remediating high-priority issues.

However, it is rare that all the recommendations made by the Data Stewards get escalated to the Data Governors for a decision, as that would simply not be practical. In the real world, metadata decisions, such as definitions, derivation rules, and data quality requirements, are determined by the Business Data Steward for the data elements in question, or agreed to by the group of Business Data Stewards who represent the stakeholders of the data elements. For example, though the sales business function may own a specific data element and set the data quality requirements, finance may need to have a say in the required quality because the element is used in a report. Only if the stakeholder Business Data Stewards can’t agree on the quality requirements would this get escalated to the Data Governors. On the other hand, if the agreement on quality required that funds be spent to upgrade the system to prevent data quality degradation, that decision will need to go the Data Governance Board.

The Data Governance Board:

- Has funding authority to spend budget money on data management improvements. The ideal situation is for the Data Governors to have the ability to spend budget money to correct high-priority data management issues. This enables them to correct issues without going through the normally onerous task of getting projects approved and funded. It also makes sense for the Data Governors to have expenditure authority because they are responsible for prioritizing issues and deciding what is important enough to be remediated.

- Prioritizes decisions regarding data to address the most relevant needs of the organization. There will undoubtedly be many issues that get raised about data, and it is up to the Data Governors to prioritize their decisions so that the most important issues are addressed by the limited resources available. Ideally, the Data Governors are in a position to know most of what is going on in the company and act accordingly.

- Reviews, evaluates, and reports to the Executive Steering Committee on Data Governance performance and effectiveness. The Data Governance Board needs to work with the Data Governance Manager to create progress reports for the Executive Steering Committee.

- Ensures that annual performance measures align with Data Governance and business objectives. When a Data Governance program is first starting up, participants are being chosen from the ranks of business and IT to drive it forward. Initially, these participants will have objectives (often tied to variable compensation) that do not include Data Governance deliverables. Unfortunately, if that state of affairs is left unchanged, there is a likelihood that the duties that lead to additional compensation will get prioritized ahead of Data Governance participation. To take care of this, people who are assigned to participate in Data Governance should have their objectives adjusted to reflect the importance of that participation. The Data Governors would make these decisions, typically in conjunction with the executives and the human resources department.

- Reviews and approves Data Governance policies and goals. The Data Governors review the policies and goals laid out by the Data Governance Program Office, and typically play an active part in adjusting the policies and goals to match enterprise priorities. The Data Governors then provide approval, after which the policies are sent to the Executive Steering Committee for final approval.

- Ultimately is accountable for business data use, data quality, and prioritization of issues. As will be discussed later, Business Data Stewards are responsible for making a lot of recommendations about how data will be used, required data quality, and what issues are important enough to be addressed. However, the ultimate accountability for proper use of data rests with the Data Governors, as they have (or should have) the authority to ensure that business data use and data quality are properly determined and documented, and that issues are properly prioritized and acted on.

- Makes strategic and tactical decisions. In a nutshell, the Data Governors are responsible for making decisions about the data. Some of these decisions are strategic—should we replace an aging system that is corrupting data quality and causing regulatory problems? Other decisions are tactical—should we require the frontline customer service employees to record email addresses and mobile phone numbers, even though they don’t need it for their own purposes? Keep in mind, however, that the input for decision making will come largely from the recommendations of the Business Data Stewards.

- Reviews and, where appropriate, approves the recommendations made by members of the Data Stewardship Council. The Business Data Stewards are typically the people who directly feel the pain of poor data usage, quality, or confusion about meaning and rules of data. They are also the people who know the data best (which is one of the criteria for choosing the right steward, as discussed in Chapter 2). However, the stewards are rarely placed highly enough in the organization that they can make decisions that others will pay attention to and execute on. Thus, the stewards need to make recommendations to the Data Governors, who are highly enough placed to make these decisions. However, not all recommendations need to be referred upward to take effect. For example, if a Business Data Steward is designated as the steward for certain data elements, that steward can usually define and assign the business rules without having to have a decision made by the Data Governor.

Note

There are exceptions to the Data Governor and Business Data Steward being different people. This can occur when a highly placed manager also directly uses the data, and thus fills the role of both Data Governor and Business Data Steward. This can be seen in small organizations with flat structures.

- Assigns the Business Data Stewards to the Data Stewardship Council. Data Governors represent a business function, as do Business Data Stewards, so it makes sense for the Data Governor to assign the Business Data Stewards for that function. There may be multiple Business Data Stewards for a given Data Governor because the Business Data Stewards require a more detailed knowledge of the data, as well as direct responsibility for managing the metadata. For example, a Data Governor may be assigned to handle the insurance portion of a business, but will likely require separate Business Data Stewards to handle claims, underwriting operations, and actuarial because no one individual knows and uses the data in each of these areas. One of the more interesting aspects of this responsibility is that the Business Data Stewards do not have to report organizationally to the Data Governor—he or she can pick pretty much anyone in the organization. That may not sit well with the designated steward’s supervisor, but in a proper Data Governance program, the Data Governor has the right to pick the best person for the job. Of course, a balance must be maintained between Data Governance and the other functions of the business!

- Represents all data stakeholders in the Data Governance process. Ensures appropriate representation and participation in Data Governance across the enterprise. It is key for the Data Governors to take the “enterprise view” of managing data properly, and protect the interests of everyone who has a stake in the data. In fact, one of the few causes for escalating an issue to the next level (Executive Steering Committee) is when a Data Governor does not represent all the data stakeholders. In addition, ensuring that there is appropriate representation and participation in Data Governance includes dealing with the following:

Lack of representation from a business area. As the Data Governance effort matures, a time will come when the data being discussed does not have a Data Governor. This can happen as the Data Governance effort expands into a new area of the company. It can also occur when a Data Governor realizes that data he or she is assumed to own is actually part of a different business function. For example, the Finance Data Governor on one engagement realized that the data being discussed was Accounting data, which he did not work with or know much about. He suggested adding a Data Governor and Business Data Steward from Accounting, which was accomplished and filled the gap.

Understaffing of the Business Data Stewards. It is important that the Business Data Stewardship effort be fully staffed and that there is continuity when the stewards change. It may become necessary to replace a Business Data Steward due to normal organizational changes, such as when someone changes assignments or leaves the company. Just as the Data Governor was originally tasked with appointing the Business Data Steward, the Data Governor must also appoint a successor.

- Owns and governs the business processes that produce data. The Data Governors have to take responsibility for the processes that produce the data in their business function. Keep in mind that bad data is typically a symptom of a broken or poorly defined process.

- Identifies and provides data requirements that meet both enterprise business objectives and the objectives of their business function. Balancing the needs of the enterprise (or other business functions) against the needs of the owning business function can be tricky, and requires that the Data Governors work closely together to reach solutions that benefit all the data stakeholders.

- Defines data strategy based on business strategy and requirements. The data strategy must satisfy more than just the needs of the owning organization—it must drive forward the overall enterprise business strategy.

- Communicates concerns and issues about data to the Data Stewardship function. Since the Data Governors are in contact with each other on a regular basis, issues may be brought to them from peers or other concerned parties. In such cases, the concerns and issues must be communicated to the Business Data Stewards to research, validate, and make recommendations on the advisability of a fix or the proposed solution.

Data Stewardship Council

The bottom of the Data Governance pyramid is the Data Stewardship Council, comprised of the Business Data Stewards. We will discuss who these people are, their qualifications, what makes a good steward, and their day-to-day tasks in much more detail throughout the rest of this book. But make no mistake, although they sit at the bottom of the pyramid, they are extremely important—much like the bottom of a real pyramid, they are the foundation of Data Governance. I have seen relatively successful Data Governance efforts occur without the Executive Steering Committee or the Data Governance Board, but I have never witnessed any success without the Business Data Stewards. Stewardship—in the sense of “taking care of something that is not yours”—can and does take place in the absence of a formal Data Governance structure when the employees understand that the quality of their data has an impact on organizational success. And sometimes, when the employees are the kind of people who want things to work right!

IT Support Through the Technical Data Stewards

The primary role of the IT Technical Data Stewards is to provide technical expertise in support of the Data Governance efforts with respect to systems and application impact analysis for proposed changes and data quality issues. IT resources are assigned to this role officially by IT management and are expected to respond to requests for assistance from Data Governance in a timely manner as part of their regular duties. These individuals are often lead programmers, database administrators, and application owners, and their detailed responsibilities will be discussed in much more detail in Chapter 2.

Data Governance Program Office

The Data Governance effort, including documentation, communication, and enforcement, is run by the Data Governance Program Office (DGPO). The Program Office must be staffed with ample resources to do the work, including at least a full-time program manager. Failing to create the DGPO and staff it adequately, or depending on part-time resources with other jobs to fill in, is a sure recipe for failure. So don’t do it.

Note

The person who runs the Data Governance Program Office is referred to as the Data Governance Manager or Chief Data Steward. In this book we will refer to the person as the Data Governance Manager, but either title is common.

The coordination of the Data Governance program is the responsibility of the DGPO, and that includes chairing the three committees already discussed: the Executive Steering Committee, the Data Governance Board, and the Data Stewardship Council. The DGPO is also responsible for making sure everything that Data Governance does is fully documented and made available to all interested parties. Table 1.1 summarizes the responsibilities of the DGPO, and many of these items will be discussed further in the book.

Table 1.1

Data Governance Program Office Responsibilities

The Overall Data Governance Organization

The pyramid is a useful way of conceptually showing the various levels of the Data Governance organization, but it does not show the relationships between all the participants. Figure 1.4 shows how the various individuals come together to participate in Data Governance.

FIGURE 1.4 The business and IT view of the Data Governance organization.

There are some things to notice about this figure. First of all, there is both a business sponsor and an IT sponsor for the Data Governance program. These positions are crucial to a successful program because they provide the key executive sponsorship. The business sponsor is often the COO or equivalent, whereas the IT sponsor is usually either the CIO or a direct report to the CIO. However, these are not hard-and-fast rules. The business sponsor can often be the CFO, as people in finance are used to having to follow strict rules and often feel considerable pain when data is not governed properly. Another alternative is the CDO (Chief Data Officer) if the company is sophisticated enough to have recognized the need for this position and found a suitable individual.

The Data Governance Manager runs the Data Governance Board. The role of Enterprise Data Steward runs the Data Stewardship Council. This role may be filled by the Data Governance Manager in small organizations, although it is often a separate individual who reports to the Data Governance Manager and is part of the DGPO.

Also notice that Technical Data Stewards don’t magically appear—these people need to be assigned and made available by IT managers with sufficient authority to make the assignment.

Figure 1.5 shows a different view (but the same structure as in Figure 1.4) of the organization, including the reporting relationships within the DGPO. The primary difference between the two figures is that Figure 1.5 shows the Executive Steering Committee and individual Data Governors (“DGB Member”) and Business Data Stewards. Additionally, it shows the reporting relationships of the members of the DGPO (Data Governance Manager, Enterprise Data Steward, and Domain Data Stewards). Notice also that the Data Governance Manager is proposed as being a direct report to a highly placed executive.

FIGURE 1.5 Reporting relationships of the overall organization.

Note

The Domain Data Stewards and Project Data Stewards are discussed in Chapter 2.

Summary

As stated earlier, Data Governance is all about how people work together to make recommendations and decisions about data. That is, the people are the most important part of the equation. The Data Stewards represent the largest population of those people, as well as the most knowledgeable about data. They are, therefore, the keystone of the Data Governance program.

The key to a successful Data Governance program is to set up and staff the organization as needed, as well as understand and socialize the roles and responsibilities for the Data Governance participants. One of the most critical parts of the organization are the Data Stewards, who must work together to determine ownership, meaning, and quality requirements for their data. Without a smoothly functioning Data Stewardship Committee, Data Governance cannot produce the important data-related deliverables that are the most important measure of the program’s success.