Skip To Content
Rapid Insight: Higher Education case study

Predictive Modeling to Streamline Recruiting: University of Nebraska-Lincoln

Predictive Modeling for Recruitment

Out-of-state recruiting

The University of Nebraska at Lincoln’s undergraduate enrollment was approximately 80% in-state students and 20% out-of-state students. When they took a look their traditional hotbed areas for out-of-state enrollment, including a few nearby states, they noticed that there were some pretty sharp declines. As David Burge, Associate Dean and Recruitment Coordinator, describes their recruiting process, “Just like everyone, every year we’d purchase a number of student search  names, load them into our CRM system, and mail to them. We were limited in terms of budget on how many names we could buy relative to how much we could afford to mail.”

Eventually UNL decided that this was not the best approach. “We came to realize that we were not getting everyone identified for our search process. We knew we needed to do a better job of identifying the right prospects.”

Evaluating analysis options

If you were to ask him, David Burge would tell you that he does not have a strong statistical background or a lot of prior experience building predictive models. These were important considerations when David and his colleagues were deciding which solution would be the best fit for them. Multiple vendors gave him six figure quotes to take over the entire process – purchasing names, generating the emails, and delivering inquired students. However, UNL was already doing much of that work in-house and so was looking for a more cost-effective solution.

After viewing a demonstration of the Rapid Insight solution, UNL chose Rapid Insight as its solution because of its cost effectiveness and ease of use. “I could see a million different applications of the software for modeling and data analysis,” says David. “We completed all of the training we needed by watching recorded webinars, taking notes, and experimenting – without ever opening a user manual!”

Building a recruitment model

With their decision made, UNL dug into their warehouse and pulled the last three years of information with which to build a predictive model. The next step was to load the variables into Rapid Insight’s Predict. “Building the model was easy because of the level of automation,” David asserts. “When the variables were loaded, we used the Automated Mining feature, which allowed us to click a button to determine which variables were predictive. The product is able to determine exactly how all of the variables are related and automatically assemble the predictive model,” While the model picked up on the usual suspects, David appreciated the new and deeper levels of analysis that Rapid Insight delivered – those not found by scanning spreadsheets or building pivot tables.

Using Rapid Insight

One of the things David valued most about working with Rapid Insight was utilizing the customer support whenever he had questions or concerns during the modeling process.

”Anytime I had a question I would get on the phone with someone at Rapid Insight and we would talk about it. We would talk about strategies and best practices. I don’t have a statistical background, so having the analysts at Rapid Insight on the other end of the phone line proved to be almost as attractive as the product itself.”

Model results

After scoring their search names, David and his team decided to target their mailings based on each student’s propensity to enroll at UNL. As he puts it, “We wanted to invest more money in finding out about the students, and not necessarily more money in physically sending mail to those students. We actually decreased the number of students we mailed to, based on our analysis.”

When it came time to look at their results, David was pleased. “Our responses followed the model, and interestingly, so did our applications. We believe that the modeling we did was relevant within the function we wished it to be.”

Financially, David believes they made the right choice. They were able to mail significantly less and target their recruiting efforts much more efficiently. Their total cost including purchases, mailing, and the Rapid Insight software suite saved them 6 figures under the quotes they were given initially. Overall, they were pleased with the number and quality of inquiries they received as a result using their predictive model.

Future plans

Since building their initial inquiry model, UNL has built two additional models – to predict which inquiries were most likely to convert to applicants and to predict which admitted applicants were most likely to apply. They’ve used the results of these models to further focus their admission efforts on those students who are most likely to attend UNL. They’ve incorporated their findings into their phone campaigns and have found positive gains in the number of inquiries they’ve been able to reach.

They have plans to build another model to look at enrollment deposits and predict who is most likely to walk away after May 1 –  assisting both faculty and residence life in accurately predicting the number of students who will be on campus on the first day of classes.

About University of Nebraska – Lincoln

The University of Nebraska-Lincoln, chartered by the Legislature in 1869, is that part of the University of Nebraska system which serves as both the land-grant and the comprehensive public University for the State of Nebraska. Those responsible for its origins recognized the value of combining the breadth of a comprehensive University with the professional and outreach orientation of the land-grant University, thus establishing a campus which has evolved to become the flagship campus of the University of Nebraska. UNL works cooperatively with the other three campuses and Central Administration to provide for its student body and all Nebraskans the widest array of disciplines, areas of expertise, and specialized facilities of any institution within the state.