Duquesne University transforms its enrollment decision-making with Microsoft Power BI

Joel Bauman, Senior Vice President for Enrollment at Duquesne University, quickly discovered that actionable data was historically very siloed, top-down, and sometimes unnecessarily guarded at the university, resulting in limited access across the organization. Additionally, the information wasn’t available in real-time and had a limited distribution.

“My VP of Marketing liked to say that a lot of the work we do here is artisanal,” says Bauman, laughing. “I had four people in my department who spent all day on Fridays creating the weekly reports. These would get distributed as PDFs that generated dozens of questions but no answers, and there was very little ability to analyze. Just straight, static, year-to-year comparisons without the ability to dig deeper or generate any useful observations, let alone decisions.”

With the support of the Provost, Bauman partnered with key colleagues across campus and principally, Dr. Jeffrey Miller, Vice Provost at Duquesne. “We needed a system that would support decision-making based on real-time data,” explains Miller. “These decisions ranged from admissions, retention, financial aid disbursement, hiring, developing programs, sunsetting programs, adjusting schedules, and workload considerations, among many others that needed current data.”

Deploying Power BI to drive efficiency

Duquesne began working with TRIVIUM BI to implement an improved system in undergraduate admissions and financial aid in March 2021. Today, halfway through the implementation, the university is in a “trust” phase. TRIVIUM BI founder and CEO Resche Hines and Angela Henderson, Chief Data Architect at TRIVIUM BI, work with the institution to build confidence among stakeholders as the new system rolls out across the different undergraduate schools. The graduate school will soon follow.

The goal of adopting visual data was to make the organization more efficient regarding the time required to create reports. Duquesne has 20 years of data built into Microsoft Excel tables. TRIVIUM BI transformed much of that information into a cloud-based solution using Microsoft Power BI, turning those Excel tables into visualized dashboards. This allowed the leadership at Duquesne to hold onto the ways they’ve always done things and see the new system’s efficiencies.

When Power BI was initially introduced, the team members who spent their Fridays producing the spreadsheets and PDFs asked, “What will we do? Are we no longer relevant?”

“My answer,” says Bauman, “was that there’s now a resource and partner to help us with this, and your job now will be to analyze, support, and help make decisions. It’s a much higher-level responsibility and service to the university.”

Duquesne also plans to open a medical school in 2024, which must be accounted for in the TRIVIUM solution. The Power BI solution will work seamlessly across the entire institution and can be scaled for future expansion. “Obviously, this is a courageous endeavor, and it was important for the institution to understand how they use their resources and understand their data,” says Hines.

Another consideration for implementation was the university accreditation process, where regional accreditors require unit assessment systems to be in place. “Academic programs often have accreditation processes that require formative and summative data, and non-accredited programs need to be reviewed periodically,” says Jeff Miller, Vice Provost at Duquesne University. “These macro assessment processes need to be informed by the same data that informs day-to-day decision-making.”

The data flows in the image below, with red arrows representing temporally based analysis of data sources, are now driven by Power BI dashboards and some proprietary consultant data sources.

The overall goal of TRIVIUM BI’s work is to shore up the running of the institution for better student outcomes and to attract the best-qualified class of students, ones that would be successful at Duquesne.

“It’s always mission-driven,” explains Hines, “whether it’s to find the right students for the right reasons or to shore up the bottom line. We need better analytical data to ultimately support students so we can help increase retention and graduation rates.” 

Viewing data in a “Decision Support” workspace

Miller touts the Power BI-based system now implemented on campus and has dubbed the new Microsoft workspace “Decision Support.”

“This workspace is a collection of dashboards that serves two purposes,” he says. “Every time someone needs to make a decision, they identify evidence to support that decision. For example, when deciding to increase the discount rate for a particular program, they look to support that decision with its impact on net new revenue for that program. The data must have functional utility on a day-to-day basis. That is, data display and use become regularized around functional decision-making.”

“Dr. Miller describes the new workspace as ‘revolutionary,’” says Bauman. “He told me, ‘I use it every day. I can’t believe we’ve lived without this for so long!’ He’s now able to visualize student registrations and other interactive reporting that he can click through when he’s talking with individual departments. Now with Power BI, we can have productive conversations supported by real-time data, rather than hunting down a static report in a PDF or spreadsheet.”

“For the visualizations to work, we had to rethink what data we were looking at,” Miller explains. “The rule was, if a data element didn’t inform decision-making, it didn’t get included in a dashboard. Thus, decision-making became the primary driver of what data we used.”

Spotting problems through visual data

Understanding where declines in enrollment are coming from geographically, what programs needed shoring up, and margins and net revenues (the amount students pay after scholarship funding) contributed to the university making better decisions about the institution’s financial health. 

“If there’s decline in one area where net revenues are higher, a smaller headcount is actually more impactful than one where it’s not, and we didn’t know that before,” says Bauman. 

The ability to compare current data with the past helps determine if benchmarks are being met and to adjust if they’re not. Interventions might involve alerting the financing team about lagging numbers so that they can ensure scholarship letters have been processed and sent out, prompting students to accept their enrollment invitations.

In the old reporting system, admissions data was static. With Power BI, the enrollment department can ask what happened after a student was invited to enroll? Did they accept? The team realized that one of the school’s more competitive programs, the Physician’s Assistant program, accepts only 40 students out of more than 1,000 applications each year.  With that knowledge, the enrollment department can now offer students on the waitlist spots in other programs of interest which have greater capacity, such as a pre-med track students might not have known about, and thus increases enrollment in other programs and overall.

With the new reporting, they discovered that the Liberal Arts program had great growth in political science and psychology but declines in computer science. That created an opportunity to ask more questions about that dynamic and compare it to competitor schools.

“It’s something to be able to see your data and trust your data, rather than having a hunch in your gut,” explains Hines. “It was interesting to discover which local community colleges had the highest transfer rate into the university, or which high schools had the highest recruitment rates. You may know these things intangibly, but the school is becoming much more efficient at understanding these trends.”

Moving forward with data democracy, not data siloes

The university has a very advanced institutional data security system, so questions of access to and tracking private information generated interesting conversations around who should and shouldn’t have access. In Bauman’s opinion, information should be democratized so that everyone can have an opinion based on real-time observations.

Once the information is public, he explains, the idea of data being “wrong” disappears. They can question sources and then collaborate and coordinate to come up with better reports. When information generated by the newly available data started being shared across the organization, teams coordinated data sources and time frames and discovered underlying discrepancies often amounted to a timing issue and not a calculation one.

“It’s baby steps,” he says. “Enrollment services are the sales and financing arm of the university. We’re on the front lines, projecting trends to understand the competitive landscape. That wasn’t what we were doing before we had these Microsoft Power BI data visualization tools. It’s been completely rewarding to see how the democratizing of data has changed the organization.”

“What’s exciting in all of this is how much more interesting the world becomes as more data is available. The ability to satisfy curiosity and find patterns that generate new questions, and seeing others have ‘aha’ moments through interacting with this information, too, has been very gratifying,” says Bauman. “It’s a legacy project for the university that will live on long after we’re gone.”

Hines agrees. “The reason data is so important to higher education is that every data point is a student, a life,” he says. “When we talk about increasing retention by two points, that’s 100, 200, or 400 students. We have the ability to change the lives of 400 people directly. When you think about their brothers and sisters, parents, grandparents, and community, it becomes transformative for our entire country and even the world.”