SAP for Higher Education and Research Blogs
Discover practical tips and insights to optimize operations and enhance learning experiences with SAP. Share your own experiences in higher education and research.
Showing results for 
Search instead for 
Did you mean: 

BY: Dr. James David Hardison, Principal, Industry Value Engineering, Higher Education, SAP

After working in higher education for decades, I can tell you this: Anyone who’s going to devote his or her life to academics has a sense that they’re shaping the future – for individual students and for society as a whole.


Because of that larger mission, people in academia are concerned about dropout rates for two reasons:


  1. There’s an economic impact – and not just for the university. It’s true that a student who drops out early in his or her college
    career represents a loss in tuition and fees for years to come. But beyond that, college graduates
    earn more than non-graduates and are able to
    contribute more to the economy through taxes and spending.
  2. A student who doesn’t graduate means that the institution has failed to fully execute its social responsibility. People in higher education take that personally.

Technology has impacted how colleges and universities are run. While advisors may be more technologically connected to their students than before, one-on-one, in-personal interactions between professors and pupils are sometimes collateral damage.


The days of a professor inviting 25 students to his or her home for dinner are largely in the past. But face-to-face connections can sometimes be game changers, helping
determine whether a student stays in school until graduation or drops out.


So how can we utilize technology to empower busy advisors to identify and reach out to students who are struggling? Using powerful in-memory computing to
create granular-level personas of at-risk students might just hold the answer.


The unengaged student won’t take initiative

College and universities have made some attempts – with mixed levels of success – to use technology to help the troubled student.  Typically, those efforts have involved early-warning systems to alert pupils that it’s time to seek extra help based on their grades.


This is akin to an “open door,” or “shopping center,” approach. The message to the student is that the university is providing everything you need – including office hours with professors, tutoring, and study materials. You just need to take advantage of the resources.


The problem is that students on the verge of dropping out may be having difficulties in areas that go far beyond grades. These people are often unengaged in collegiate life and may be somewhat isolated. A smartphone warning is unlikely to prompt an unengaged student to seek help on his or her own.


That’s where technology can play an exciting role. If institutions can segment their student populations based on characteristics that may indicate a lack of engagement, there’s an opportunity to help those students turn things around.

Start by making the problem smaller


When I worked at the University of Kentucky, I had the pleasure of tackling some of these issues with a team led by Dr. Vince Kellen, Senior Vice Provost and CIO.  Our group realized that if we could increase the graduation rate – what we called the “student success rate” – by just 1%, it would have an immediate positive impact of over $1 million per year on the budget.

It was big problem to solve. We needed to figure out how to put the problem into a manageable scope, so we decided to do what marketing folks had been doing for years: segment our audience.


To do that, we had to contend with some big data. We wanted to pull information from multiple and disparate sources to create a profile of students who might need additional help. We were already collecting a lot of data in the regular business of running the university and in student lifecycle management: attendance, student
information, library usage, learning management, and ID swipes when students entered campus buildings. All of those things could be markers of student

Our goal was simply to uncover groups of 40 students – 1% of a class of 4,000 – with similar characteristics. We realized that handling this information and getting it into useable form was going to require running complex and time-consuming database queries. As with any undertaking like this, we knew that we’d start asking
better questions as we moved along. What we needed was a way to fail fast – we wanted to plow through the early questions quickly so we could get to the information
that we really wanted.


We turned to the massive power of in-memory computing. Because the system was able to handle so much information at one time, we were able to get instant answers to queries that previously might have taken weeks. Because all of the information was in memory at one time, we didn’t have to go through all the data manipulation, extraction, and translation. We could change variables and get answers instantly.


Using that information, and running multiple queries, we were able to get a much better picture of our students. And since we were working with granular data rather than the typical summary data, it was a lot easier to identify certain characteristics so we could approach students who really needed guidance.


The results were surprising. We not only exceeded our goal for graduation rates, we saw an ROI of 509% in five years.


Change in open-door policy


Technology can make it a lot easier for advisors to reach out to students who need help, rather than the other way around. It’s important to remember that technology doesn’t have to replace a human interaction; instead it can help guide, direct, supplement, and inform as to where those interactions are most needed. That way, institutions can move from an open-door policy to a policy of “Please, come in our door.”

Top kudoed authors