University CIO helps boost graduation rates with analytics

As interim CIO at Valdosta State University in Georgia, Brian A. Haugabrook must ensure that the school's IT systems meet the needs of faculty, staff and 12,000-plus students. But he says an effective IT operation must do more than keep computers and networks running. It must also help the university fulfill its mission by giving officials a way to reach out to students at risk of failing. To do that, Haugabrook deployed analytics tools that allow faculty and staff members to determine which students need extra help and then coordinate that support. Less than two years into the project, the tools are yielding results. Here, Haugabrook talks about how VSU's IT department is helping to improve graduation rates.

The university is using BI tools to improve student retention and success. How did this strategy come about? The initial discussion started with the Complete College America initiative. We started Complete College Georgia in 2011. The goal is to graduate more students over the next six years. In that plan, the first critical goal is analytics -- having more data-driven decisions, having a better understanding of who your students are, the needs of your students. Our student support area, our advising, those are the areas key to student success. So we didn't focus on administrative reports; we focused on everything key to faculty, advisers and support services, to give them the information they need so they could tell which students were at risk. That's how we planned it out from the beginning.

What technology and business processes did you implement to support this plan? The first thing we did is decide that if we wanted to be innovative, we had to build a team. We created a team, changed our organizational structure. We partnered with our institutional research department, because we knew that partnership would be key to success, putting the right people on the bus in the right seat. Then we looked at what technologies were available. We found that we could quickly adapt with Oracle, and using the Oracle BI tools we could build out reports in days instead of months.

How are you using these tools? We have a two-phase approach: getting data out and then pushing data into the system. So we use Oracle Application Express to provide that two-way communication. That's a key component, the whole life cycle of delivering data to faculty. They can use the tools to say, "Yes, this student is at risk and they need additional services."

It was this combination that allowed us to quickly adapt to business needs. The first phase with those two tools began in January 2012 and was implemented in August 2012. It was released to our Freshman Learning Community faculty, which was about 10% of our faculty. But within two months the word spread around campus about how useful it was, so in mid-September it was released to all faculty.

Have you seen improvements because of this insight? On this first phase, we saw an overall increase of 2% in our retention rate. That's huge. We measured how often faculty used it, and for faculty who used the tool frequently, their retention rate jumped 4% in one year. We also measured course completion and grades. In introductory math, we saw a 5% increase in passing grades. That was in the first year.

How are you using analytics? We did a pilot in our math program. We found that math is a gateway course. If a student can't get passed through, there's a high chance they're going to drop out of college. We looked back over five years of data and created a model to see, based on what we know about a student, if they were likely to pass. We came up with an accurate formula. If we know a student won't do well, we now have them register for a lower-level course where they can get extra support and if they do well they can move up and get credit for that upper-level course.

What's next for your team around analytics? Our next big goal is to incorporate the unstructured data. In the past, you have the students' grades, GPA, the kind of standard models. Now we want to measure engagement. We know students who are more engaged have a higher level of success. So how do we measure engagement?

We worked with Oracle again. They have a tool called Endeca Information Discovery. We did a concept to mine unstructured data -- data from surveys, data from social media sites and databases we have -- to see if this tool can give us value to that data. We were able to merge in all the information we have on students -- whether they purchased meals, go to our cafeterias to eat, whether they checked out a library book, go to our rec center. We can pull in all that data using that Oracle tool to see what makes an engaged student and how that influences success rates.

Any insights gleaned from that? One of the first things I wanted to know was whether eating breakfast really does make a difference. Within one minute we could dig through all that data. Our overall retention rate is in the 70s, but students who eat breakfast are in the mid-80s. That's not to say eating breakfast is a cause, but now we can dig deeper.

What was your biggest challenge during this project? It took a cultural change to focus on BI, because IT has traditionally been focused on keeping the campus running. So it was a big change, creating a different team, moving people around. And any cultural change takes a few years. Now we see our other areas of IT becoming more innovative. For example, our technical support services team, they're creating a technical response unit. They're redefining what technical support is, and they're creating a rapid response. If someone calls with a problem and it can't be handled over the phone in five minutes, they'll have someone at the office in 15, so people aren't held up by technology and will have the support they need from IT.

-- Interview by Computerworld contributing writer Mary K. Pratt ( marykpratt@verizon.net)

Read more about management in Computerworld's Management Topic Center.

This story, "University CIO helps boost graduation rates with analytics" was originally published by Computerworld .

Editors' Picks
Join the discussion
Be the first to comment on this article. Our Commenting Policies