The University of North Texas System (UNTS) is big as higher education institutions (HEIs) go, but it is leveraging data to make itself nimble and responsive.
UNTS employs more than 10,000 people at its various campuses, with a combined enrollment of nearly 44,000 students, and an annual consolidated budget of $1.2 billion. It includes the University of North Texas in Denton (UNT), the University of North Texas Health Science Center (UNTHSC) in Fort Worth, and the University of North Texas at Dallas (UNTD). In all, it boasts five major teaching locations.
Like many HEIs, UNTS generates a lot of data, but until a few years ago, it had little to show for it. The data was siloed, data governance was non-existent, and analysis required manual coding. Worse, after three failed attempts to establish a data warehouse and analytics program, hope and trust were in short supply.
“There was a significant lack of trust in the information and data resources that were being provided prior to the program start,” says Jason Simon, associate vice president of data, analytics, and institutional research (DAIR) at UNT. “It was the wild, wild west of data and there was no clear governance around information, resources, and data elements.”
Moreover, UNT’s reporting was largely time-bound and static, and did not allow for forecasting or longitudinal analysis.
“We needed to get our arms around our data resources and our data tools and really present the business with a fundamentally different way of doing business,” Simon says.
The turning point came at a data summit in 2015. UNT brought together stakeholders from across the university to discuss data issues with the DAIR team: what was working, what wasn’t working, what their pain points were, and what it was like to be a data provider at UNT. After hearing from stakeholders, UNTS decided to reimagine the role analytics plays in HEIs, an ambitious course that included not only a new program of data warehousing, analytics, modeling, and governance tools but an emphasis on culture. Called “Insights,” the program has earned UNTS a CIO 100 Award in IT Excellence.
Digging into data problems
UNTS launched its effort by turning to its unique in-house expertise: its anthropology faculty, who conducted a qualitative research study on the health and culture of the data landscape at the institution. The results would inform the program charter.
“That was a very creative solution because it did two things: First of all, it engaged faculty, which is a very important segment of our population here, but more importantly it allowed those of us in the data and analytics unit to step back and be listeners,” Simon says.
The faculty group met with more than 40 executives from across the system and distilled the resulting information down to major themes and pain point areas. Simon’s team combined the results with information from the data summit and other group meetings and met with UNTS’ technical subject matter experts to focus on culture building. Simon and UNTS’ chief enterprise architect did some groundwork of their own. Their five-day fact-finding mission to each of the university system’s campuses found that data was siloed, hard to replicate and programmed manually. Trends were difficult to recognize, and decisions were made by “gut.” Moreover, IT and institutional research (IR) didn’t work together; data requests were answered using a limited extract of variables; and ad hoc reporting required a lot of email back-and-forth.
Tools could help address individual areas, Simon says, but the DAIR team wanted to address what they saw as the root of the problem first: data quality, data veracity, and data governance.
The importance of culture
Recognizing that success would require ongoing support and financial resources, DAIR adopted a program orientation rather than a project orientation. The team brought together data warehousing, analytics, data visualization, and predictive analytics, with a data governance strategy, communication strategy, and training strategy.
With plans in place, UNTS engaged infrastructure vendors with a focus on the institution’s holistic data needs rather than point solutions. Weekly steering committee meetings between IT and DAIR leadership helped them focus on both short-term and long-term goals. The organization made strategic hires in both data governance and modeling, and the data modeling experts worked closely with subject matter experts. Training sessions were used throughout the organization.
Once the Insights Program was in place, Simon says everything was different:
- Data are governed, federated, and housed in a central data repository.
- Data are automated and updated automatically
- Analytic visualization products highlight outliers and trends, fueling decision making.
- IT and IR collaborate closely and share expertise.
- Users have secure web portals to access grading patterns, retention, payroll trends, demographics, outcomes, and enrollment trends.
- Expanded data are drawn from a new architectural environment built to collocate information across diverse data domains.
- Users can access self-service drillable, sortable, and exportable data visualizations.
In addition to helping UNTS boost student outcomes and retention rates, Insights has helped it analyze bus routes against where students live, enabling it to eliminate an entire route and better serve students based on where they live. It’s using Insights to study financial aid distribution patterns and the course-taking habits of students. A second-generation of Insights will soon leverage machine learning to help UNTS determine where and when courses should be offered and provide predictive scoring on the propensity of an individual to enroll at an institution based on past interactions.
Simon says the biggest key to success was a focus on culture rather than just technology.
“You have to be a culture-driven analytics leader. What I mean by that is it’s not good enough to just know your tool. You have to have emotional intelligence; you have to have the ability to understand previous failures and to recognize what a stakeholder’s concerns might be. You have to have the ability to think about historic challenges that your enterprise has faced and not make the same mistakes. You have to have an awareness of what your community feels in terms of your data landscape,” says Simon.