Stanford Human Resources (HR) has approved the creation of a project independent data stewardship team to provide a renewed focus on the effective management of critical HR institutional data assets. In addition to the stewardship team, the Policy and Process Committee has accepted the responsibility for the executive aspects of data governance in the HR subject area. Further details can be found in linked presentation: Project Independent Human Resources Data Stewardship
Attendees: Mario Acquesta (University HR), Rana Glasgal (University HR), Matt Hoying (University Data Governance), Cindy Martin (University HR)
The purpose of this meeting was to discuss the scope and direction of a project-independent data stewardship (DS) effort around the Human Resources (HR) data subject area. This effort will ultimately include stakeholders from each of the schools and VP areas and is not intended to be a strictly University HR (UHR) effort.
Data Stewardship can be defined as the formalization of accountability for the definition, usage and quality standards of specific data assets within a defined organizational scope. (This definition, along with more information on Data Governance and its relation to Data Stewardship, can be found in our first DG at Stanford newsletter: http://dg.stanford.edu/wp-content/uploads/2011/11/DG-News001.pdf.)
Many of the responsibilities and activities associated with data stewardship occur today, both as part of day-to-day operations and formal projects. What we want to do is document and refine these processes, clearly define the roles, and assign responsibility and accountability to specific individuals in order to assure that we are consistently managing key data. These activities can be divided up into six primary categories: Metadata, Administration, Data Quality (DQ), Audit, Technical, and Support. While the DQ activities could be distributed among Metadata, Audit, Technical, and Support rather than forming a separate category, here they are combined to ensure that there is a concerted focus on DQ.
- Metadata includes the activities around documenting data, instances of data and relationships between data.
- Administration includes the prioritization of data stewardship and data quality activities and the definition, execution and enforcement of data policy.
- Data Quality includes activities related to identifying and analyzing data quality problems and developing metrics, thresholds and remediation strategies to guarantee fit-for-use data.
- Audit includes the ongoing operational activities that compare data policy and data standards with implementation.
- Technical includes the design, development and maintenance of the technical infrastructure (both hardware and software) that enable efficient and effective data stewardship.
- Support includes training and communication activities that ensure the consistent understanding and implementation of data policies and standards.
As it would be unreasonable to expect that we would be able to implement all of the tasks and roles associated with DS immediately, our focus is on selecting the activities that will give us the most benefit with the least effort. The following diagram and chart were used to discuss the relationship between DS activity categories and some of the representative activities within each category: http://dg.stanford.edu/wp-content/uploads/2012/06/DG-Stewardship-Matrix.xlsx (first two tabs).
The beginning of a program like this is often one of the most difficult phases as participants are still learning the purpose and procedures related to each of these activities. Fortunately, through the HR Metrics Dashboard-Phase I project (lessoned learned can be found at http://dg.stanford.edu/?p=497), we were able to get some experience with the activities around defining data (http://dg.stanford.edu/?tag=data-definitions), assessing data gaps and making a formal request for change (RFC) (http://dg.stanford.edu/?p=465).
All attendees agreed that formalization of data stewardship with the HR area should be pursued, as this class of activities potentially has significant value across the university. There will probably need to be two separate teams: one in charge of the general data policies and scoping activities, and a second responsible for developing the specific definitions and data quality metrics. The initial effort will focus on a specific business function where the data is relatively narrow in scope and there is a clear connection between data quality and business impact. This will allow us to demonstrate value and establish baseline policies that can be later applied much more broadly.
A follow up meeting has been scheduled for Thursday, June 21st at 1:00 PM PDT to finalize primary and secondary tasks and goals, business function of focus and the composition of the identified teams.
Data stewardship can be approached in many different ways based on the specifics of the environment in which it is implemented. Factors such as resource availability, program scope and company culture are significant factors in the design of the data steward role. In most cases, the responsibilities (and accountabilities) associated with effective data stewardship necessitate the definition of many complementary roles such as, Data Custodian, Data Trustee, Coordinator, Operational Data Steward, Data Analyst, Metadata Custodian etc.
Linked below is a non-comprehensive list of activities associated with data stewardship. Not all of the responsibilities have to be fulfilled in every case but each should be considered in the design of a stewardship program.
Please contact Matt Hoying with any questions, comments or suggestions.