White Paper: Common Trouble Spots When Implementing Data Governance
Overview: Data Governance is complicated. Implementing a program for any size organization can create challenges that are new and expose weaknesses in a company’s systems or processes. Although these challenges are unique to each company, there are some common themes. Each of these problem areas can be solved either by engaging an outside company to provide expertise or by cultivating that expertise within a few key individuals, which we’ll call Mentors, within the organization.
Problem Areas: Several key problem areas may become evident within the first 18 months of implementing a data governance program. These fall under the following topics: Education of the Stewards; Having the Right People in the Right Roles; Roles & Responsibilities; Stakeholder Engagement; and Data Quality. Stewards must also serve in an advisory role and be the point of contact for communications.
Education of the Stewards: The Data Steward role is often new for the company. It is a new way of thinking. It requires looking beyond where we create and manage data and understanding how the data is consumed. End-to-end thinking is surprisingly uncommon in today’s environment of specialization. Functions within an organization are often so specialized that the people doing them may have little or no exposure to how their actions affect other areas of the company. The Data Stewards need to be given access and empowerment to look across boundaries to ensure that information is shepherded across systems and teams. Most Stewards will need education on this new perspective as well as some guidance to navigate the waters while learning the skills.
Right People in the Right Roles: Understanding and defining what a Data Steward’s role should be is often a struggle. Each of the Data Stewards has a different level of understanding of the organization’s leadership model, data, processes, and disciplines needed to support the governance activities across the boundaries of an organization. Stewards who have job titles of Vice President or Director are generally comfortable and familiar with the role of the Steering Committee and how to provide information to that group (although what to report may still be a mystery), but they may be far outside of their comfort level when asked to make decisions about a data model or glossary term. Stewards with the job title of Analyst or Manager may be very comfortable performing the review of Glossary terms, data quality work, or the specifics of how data is managed, but they may not have the background to ensure proper Stakeholder Engagement and Communication is occurring. Stewards of any job title may be unfamiliar with data modeling, information models, and technical integrations.
Data Steward activities are needed in at least three varieties:
-
Strategic direction at the 30,000-foot and above level: These are the Sr. Mangers and above who understand the strategies and initiatives occurring in their operational area. These Stewards can make decisions and give approvals for initiatives and efforts.
-
Hands-on-keyboard-level: These are primarily Analysts, with some supervisors and Program Managers mixed in. These Stewards understand the data as it is used and managed at a tactical level. They can make decisions and give approvals at the data element level.
-
Technically savvy business Subject Matter Experts who can interpret and advise on the structures and artifacts, such as data models and Source to Target Documents, ensuring that data becomes usable information and is available when and where needed.
Very few organizations set up their program to include all three levels of Data Steward. What this causes is gaps in some business areas or domains that don’t have all three levels represented and the opposite problem where too many representatives in an area are named stewards to avoid gaps. Both problems could easily be solved if we identify the right people, create an active engagement model, and apply some project management discipline to the process.
Roles and Responsibilities: The specific tasks that a Steward must perform is often relatively new and unpracticed, and their knowledge is driven, in large part, by the specific involvement with a project. Beyond representation in projects, Data Stewards are accountable for things such as approving changes to the Code Map document, data model changes in the systems that hold their work, population and/or modification of a Data Lineage tool, “tuning” and working tasks in a Master Data Management software in a data warehouse, and resolution of confusion or conflicts about the method of solving for “new” things.
The method of review, documentation of rules, agreements or decisions, and the engagement model with other business unit SME’s is often not well defined. Putting in place a model to support the Stewards will make a positive difference in understanding and managing the end-to-end data flows and processes. Having a standard format and location to hold and manage documents will also be a part of defining the roles and responsibilities.
Stakeholder Engagement: Almost always, Data Stewards are expected to represent business units and organizations that are beyond the unit where they perform their “day job.” For example, a “Membership” (i.e. client, patient, account holder) Steward may work for the department that enrolls new individual members, but may have little understanding of the work done by the team that enrolls groups. Stewards must engage and inform the stakeholders and SMEs from those adjacent business areas, as appropriate. A solution may be to have a Steward from each group, but that begins to increase the overhead costs of time and complexity.
This illustrates the need for the Data Stewards to have an appropriate communication plan with these stakeholder groups. Many of the groups without a Steward may be completely unaware that the Data Governance organization exists and don’t know what decisions may be made on their behalf.
Data Quality: Where, What and How to measure and improve data quality in a Domain is complicated. In addition to the need to cross-over between different business units, just getting access to the data can be challenging. Defining what is important to measure is something the Stewards need to own. Looking at data in new ways and doing so consistently across Domains forces us to challenge our perspective and look up and out with considerations that may be very different from our “day job.” Ultimately, the quality of a company’s data and consistency of information has a direct impact on the experience of customers and employees and, ultimately, on the bottom line financials. A role of the Data Steward is to help organizations avoid being penny-wise and pound-foolish by looking for synergies and other opportunities to leverage best practices.
Advisory Role: Data Stewards should serve in an advisory capacity in enterprise and strategic initiatives. These items may be Operational in nature and would be appropriately raised and discussed in the Steward Council meetings. Other initiatives are within strategic portfolio and project planning. Initiatives such as the following are good candidates for the Data Stewards to engage around and help coordinate information and decision making:
-
Data Marts
-
Automation
-
End-to-end data flows
-
Glossaries and Data Models
Understanding and evaluating the initiatives, providing input on level of effort and level of risk, identifying opportunities for economies of scale, and advising on the “right thing to do” for the Enterprise and Data Domains, not just business units, will help grow the efficiency of operations.
Point of Contact and Communication: Many problems are solved in silos because it is hard to define the right points of contact. The Data Stewards should understand who in their Domain does what and should be able to serve as initial points of contact to triage questions to the correct stakeholder. They should also be communicating general Domain news back to each other and the business unit SMEs/stakeholders. This is a critical step that is often overlooked.
Solutions: Whether your organization is facing one, many, or all of these trouble spots, the recommended course of action is to provide mentors to the Stewards. These mentors could be from an outside consulting firm or a selected number of Stewards could be sent for training. Either way, having people with the understanding of the Who, What, Where, and How of Data Governance that can serve as mentors will elevate the experience of the Stewards and greatly improve the outcomes of the program. Further, establishing several Governance Leads to serve as guides toward a functional model and have some oversight of the activities occurring will help fill gaps and enable smoother operations.
Summary: Data Governance is a complex topic and becoming a Data Steward is not an easy thing to do. By understanding the problem areas and putting a support structure in place, you can minimize negative impact on the individual Stewards while improving the overall functionality of the role.