The Chief Data Officer in Higher Education

Key Takeaways

  • The role of chief data officer meets two urgent needs on campus: leading data administration efforts and building analytics capacity to drive decisions with data.
  • Like all valuable assets, data need active, intentional, and coherent management, perhaps best known as data governance.
  • The CDO can and should evangelize the value of data for analytics and decision support, while data governance ensures the best data are available to support an institution's tactical needs and strategic objectives.

Mike Kelly, Chief Data Officer, University of South Carolina

Efforts to collaboratively manage data access started taking root at the University of South Carolina system in 2008. Then, in the run-up to replacing our legacy systems with commercial off-the-shelf solutions, the Data Access Advisory Committee of data stewards was formed to make recommendations to the CIO about data concerns. In short order, members recognized their concerns extended beyond access and broadened their scope to address data administration. They initiated collaborative efforts to develop data definitions and data standards, but iterative efforts faltered when business cycles and urgent concerns vied for members' attention.

In 2012 the committee advised USC Vice President for Information Technology and CIO Bill Hogue that it needed dedicated leadership. Hogue recognized their concerns meshed with an increasing executive need to support and drive decisions with data. The role of strategic information manager was created under the CIO to lead data administration efforts and with an eye to building analytics capacity.

Year One: Definition and Evolution

Just after starting the position in January 2014, the job title was aligned to an emerging role then gaining prominence in business and industry: the chief data officer, or CDO. In the first year I had two primary concerns: defining the CDO role at USC and improving our collaborative data administration efforts.

I started by asking committee members what they needed, then adapted extant CDO models to serve a state flagship institution comprised of eight campuses and an online degree completion program, Palmetto College. All CDO models are rooted in the belief that data is both a valuable business asset and a significant opportunity and risk. Like all valuable assets, data need active, intentional, and coherent management, perhaps best known as data governance.

"Many CIOs now believe that — aside from people — data are the most valuable assets for colleges and universities. That's becoming axiomatic whether you're talking about using data to help make good tactical decisions or using data as a foundation to drive strategic initiatives. Continued investment in describing, evaluating, and managing college and university data ecosystems is imperative."

—Bill Hogue, Vice President for Information Technology and CIO, USC

USC data stewards had been on an evolutionary path since 2008, broadening from data access to data administration and seeking dedicated leadership. Data governance was the next evolutionary step, and the CDO was the obvious leadership role. But any suggestion that the CDO could be a single-handed "data governor" would be misleading.

Why? Because the principles of data stewardship hold that data should be managed by those with daily responsibility for its collection, processing, maintenance, and appropriate use — in short, those who best understand the data and bear the legal, regulatory, and compliance risks pertaining to it. Consequently, data governance requires both concentrated leadership — the CDO — and broad participation by those who "own" the data — the data stewards.

From a data steward standpoint, it's vital to have someone bridging across functional areas with a focus on information governance; data standards; identity, access and permissions management; reporting; etc. With a recent implementation of a new student information system and a pending implementation of a new finance and HR solution, we were in great need of coordination and best practices at a foundational level. It was necessary for "driving the bus" to be someone's role.

—Stacey Bradley, Associate Vice President for Student Affairs and Academic Support, USC

In its first year, the new CDO role became linked with our evolution toward data governance. The CDO is the subject matter expert on data governance principles, policy, and practices, and facilitates activities to address three compelling data challenges: compliance and reporting obligations, data risks, and data quality. The data stewards tap resources to implement data governance decisions and actions in their respective areas.

Data and Information Governance

At USC, a working group of data stewards and I developed a data and information governance framework comprised of five core programs (see figure 1). We envision an executive-level Data and Information Strategy Council that will align data governance to the university's strategic priorities and authorize initiatives and investments to improve institutional data and analytics capabilities.

Figure 1. Proposed data and information governance framework

The first core program, Data Stewardship, concerns practical implementation of policy and strategic directives, management of data-related operations and services, and assuring proper security and privacy of data. A Data Stewardship Council will serve as the coordinating body, and each data steward will assign resources to the other four core programs.

The Data Standards program will be charged with building and publishing an enterprise data glossary for key data elements, documenting data lifecycles to improve change management and communications, and implementing reference data management, as well as providing documentation and training.

The Data Quality and Integrity Assurance program will involve front-line managers who know our enterprise information systems intimately. The program will focus on proactive recognition and resolution of issues (accuracy, completeness, consistency, definition, and timeliness), gathering end-user feedback, implementing change controls, and revising business processes to avoid future data defects.

The Identity and Access Management program will involve front-line managers responsible for establishing and maintaining unit records of students and employees. The program will focus on preventing duplicate records for individuals, preventing accidental merge of records for two or more persons, resolving issues that do occur, specifying functional requirements for institutional credentials, and setting standards for role-based permissions and access controls.

Finally, the Reporting, Analytics, and Decision Support program will serve as a centralized resource for units that conduct assessment, reporting, and analytics initiatives, ensuring their access to the deliverables, knowledge base, and services of the other four core programs.

As CDO, I will be responsible for instantiating and managing the core programs of data and information governance. I have two ends in mind:

  • First, assure high-quality institutional data and information are available to enable leaders to make accurately-informed decisions.
  • Second, assure that institutional data enables continuous improvement of services, academic programs, research, and other compelling university interests.

Tony Shaw noted that the CDO role can be a business-driven role rather than an IT-driven one1; at USC, our data and information governance framework emphasizes non-IT area engagement and derived benefit. Housed in the Division of Information Technology, the CDO can take a fairly objective leadership role that consolidates the interests of multiple data stewards, and can engage colleagues within DoIT to support the data stewards through infrastructure and services that will enable their individual and collective success.

A Word about Analytics

When the position description was written, the USC CDO was expected to have a secondary focus on analytics. Through continuing efforts across the university and a new partnership with IBM, the investment in and expectations for analytics have expanded rapidly. We expect our institutional data will yield new understanding and insight. And we recognize that without the coordinated activities, services, and deliverables derived through assertive data governance, our analytics efforts might occasionally be frustrated.

Potential stumbling blocks to analytics and mandatory reporting include:

  • Inconsistent personal identifiers that inhibit merging records from different data sources
  • Missing data values that cause whole unit records to be dropped from analysis
  • Non-conforming data values that lead researchers to exercise individual judgment, potentially undermining consistency and reliability
  • Inaccurate or incomplete understanding of data elements and their definitions

Our core programs for data and information governance are designed to head off such concerns, to remediate issues that do occur, and to build in feedback loops for continuous improvement of our data and business processes. In doing so, our governance efforts should produce a solid — and vitally necessary — foundation for reporting and analytics.

CDOs: Exclusively Inclusive

The U.S. Senate has often been referred to as "the most exclusive club," numbering 100 members. By comparison, there are roughly 8 CDOs in North American colleges and universities.2 I feel confident that none of us has any interest in remaining an "exclusive" club. Quite the opposite, and here's why:

First, those of us staking our future careers on the CDO title believe a data-centric leadership role is crucial. We believe in the business value of having an individual charged with purposeful management of ever-increasing mountains of data and information. The number of CDOs should and likely will expand.

Second, a CDO knows that success in data governance can only be realized through broad inclusion and positive relationships with functional, operational, and technical colleagues across our campuses. According to Jane Griffin, a principal with Deloitte Consulting LLP, "Effective CDOs are those individuals who possess a balance of technical skill, business knowledge, and people skills to smoothly navigate the technical and political hurdles of shepherding valuable corporate data."3 CDOs know our success is often more about relationships than technology.

Relationship Matters: Stewards and Stakeholders

A CDO is unlikely to "own" any data, and he or she is not likely to have direct responsibility for the business processes that collect or generate data. The CDO should identify and bridge common interests of data owners. And we should build consistent policies, procedures, guidelines, and practices that best serve the data needs of organizational units and the university as a whole.

Often, these owners go by the titles of "data steward" and "data trustee." However, few (if any) bear that as their sole job title. More often, titles of steward or trustee are designations conferred on high-level managers such as the registrar, bursar, and various directors of admissions, financial aid, human resources, and academic credentialing. But a CDO's relationship building must go beyond the top-level managers.

Data stewards often have a number of direct reports, and an effective CDO should develop trust with the stewards to be able to interact with their staff when appropriate. The key activities and daily responsibilities of stewardship may be delegated to multiple staff — team members who can actively contribute data definitions, resolve identity conflicts, research holes and invalid data values in unit records, and explain or revise business processes to improve data quality and integrity.

A CDO will also need strong and trusting relationships with other stakeholders, including institutional research and assessment experts, analytics teams embedded in functional and academic areas, the information security team, director of internal audit, risk manager, compliance officer, institutional review board director, general counsel, IT operations, and academic research leadership.

Working with stakeholders and data stewards alike, a CDO should support processes such as change management, reference data management, and proactive communication. She should champion an enterprise data glossary and ensure critical data elements are defined. He should work with functional teams and institutional researchers to continuously monitor key data elements for missing or nonconforming values and take corrective actions before reports are generated. She might co-sponsor an end-user group to gather feedback from staff who use systems and data every day.

That's a Wrap

The coalescing forces that led USC to create the CDO position are hardly unique in higher education. With some variation in reporting based on public/private status, most institutions collect, produce, and maintain similar data on constituents, have similar mandatory reporting obligations, continuously try to improve student learning and educational outcomes, gauge faculty productivity, and seek greater cost controls and administrative efficiencies. We are bound by roughly the same federal laws and regulations as well as privacy and security imperatives surrounding data. Focused and coordinated leadership of data and information can assist with these concerns.

"When given the appropriate responsibility, authority and resources, a Chief Data Officer can effectively help shape institutional culture surrounding data. The Chief Data Officer carefully guides and directs the institution through a web of complex political and interrelated technical data issues. The Chief Data Officer is an advocate, a liaison, a working partner that bridges the gap among functional areas that share and rely on data to make decisions."

—Donald Miles, Executive Director for Institutional Research and Assessment, USC

These days, almost all postsecondary institutions want to find their way to analytics. Of course data are readily available to produce reports and support analytics, but data governance is the grueling, foundational work that goes beyond making data available to support decisions — it yields high-quality data that will support accurate understanding and well-informed decisions.

The CDO should evangelize the value of data for analytics and decision support, laboring with colleagues to ensure the very best data are available at the right time to support an institution's tactical needs and strategic objectives.

Words to the Wise

  • A data dictionary, or enterprise data glossary, may be more in demand than you could possibly anticipate. Even a partial dictionary might be a "quick win" that gives an institution a visible, derived benefit from data governance.
  • Universities are complex, and trying to solve data governance in all areas at once is impractical. At USC we constrained our initial efforts to administrative domains, namely students, HR, business, and finance. Data produced through academic research are likely in equal need of governance, but there's not a one-fits-all solution. Start somewhere.
  • In my previous role as an IT project manager at USC, I accepted that I was never the smartest person in the room. Around a conference table populated by engineers, database admins, analysts, and data center and network magicians, everyone is an expert in a critical discipline; you're there to hold them all together to achieve a specific outcome, on time and within budget. The same healthy dose of humility I had as a project manager helps me focus on interpersonal connections and recognize shared interests — skills critical to a CDO.

Recommended Reading

Raskino, Mark, and Debra Logan. "CEO Advisory: Chief Data Officers Are Foresight, Not Fad." Gartner (December 11, 2012); available to member institutions.

Notes
  1. Tony Shaw, "The Evolution of the Chief Data Officer," Dataversity, March 11, 2013.
  2. Laura Whitaker and Kevin Danchisko, "Hallmarks of the Data Driven Enterprise," EAB, The Advisory Board Company, October 2014; available to member institutions.
  3. Jane Griffin, “The Role of the Chief Data Officer,” Data Management Review, February 2008 (Vol. 18, Issue 2): 28.