Skip To Content
Rapid Insight: Higher Education case study

Lipscomb University Improves Enrollment ROI using Digital Engagement Data

Download a Summary
Lipscomb Enrollment
  • Developed new metrics to measure virtual applicant engagement

  • Built new predictive models to forecast enrollment likelihood

  • Improved return on investment for marketing spend

For several years, Lipscomb University, a private Christian institution in Nashville, Tennessee, used Rapid Insight’s analytics tools to predict which applicants were most likely to enroll. When COVID-19 struck, Lipscomb realized that applicants would not exhibit the same behaviors that Lipscomb’s models had used as predictors in the past.

To address this challenge, Lipscomb created a digital “Enrollment Engagement Score” using data from their Customer Relationship Management System (CRM) and applicant interaction with Lipscomb’s outreach. Lipscomb incorporated this new metric into their existing models and was once again able to target outreach toward persuadable applicants despite the impact of COVID-19.

Customer Testimonial

Rapid Insight equipped us to better identify those applicants who were likely to apply, better spend our marketing dollars, and develop a strategy for who to send admissions materials to.

Kelley Graham, Director of Enrollment Technology, Lipscomb University

The Challenge

In 2017, Kelley Graham, Lipscomb University’s Senior Director of Enrollment Technology, used the Rapid Insight analytic platform to build a predictive model to forecast enrollment. Graham’s goal was to identify the criteria that pushed applicants to enroll at Lipscomb so that Admissions staff could better target their outreach efforts. 

The model assigned an enrollment probability value to each applicant based on activities and behaviors. It identified on-campus visits as the highest indicator of likelihood to enroll, followed by test score values. Essentially, if a student attended an on-campus recruitment event or tour, their likelihood to enroll dramatically increased—likewise, good performance on standardized tests correlated with a higher probability to enroll.

Altered Enrollment Landscape in Higher Ed

In 2020, the COVID-19 pandemic eliminated the possibility of on-campus events and largely halted standardized testing, thereby removing Lipscomb’s strongest indicators of enrollment likelihood. 

Customer Testimonial

COVID virtually made our model ineffective.

Kelley Graham, Director of Enrollment Technology, Lipscomb University

The model needed new inputs to grapple with the altered enrollment landscape.

The Solution

Building an Enrollment Engagement Score

It had long been a goal of Graham’s to build a score to measure online applicant engagement across Lipscomb’s graduate, undergraduate, and online schools. The shift to fully-online education offered Graham the opportunity to dedicate resources to the project.

With guidance from VPs and Admissions Directors, Graham created a ranking system that gauged applicant interest based on incoming activity. For example, if an applicant attends a virtual daily visit, they receive ten points. For attending an online orientation event, the applicant gets twenty points. The applicant receives an additional point for replying to an email, calling a counselor, or otherwise reaching out to contact admissions staff.

These metrics, taken together, comprise an Engagement Score, which admissions staff use to prioritize outreach and send targeted messaging. During 2020’s turbulent enrollment windows, it was a critical piece of Lipscomb’s strategy to maintain a strong applicant-to-enrollee conversion rate.

Predictive Modeling with the Engagement Score

With the Enrollment Engagement Score established, Graham now had a data point she could use to build a new “inquiry-to-applicant” model using Predict from Rapid Insight.

Predict allowed Graham to generate models with a single click. The tool mines data to identify the strongest predictive variables and generates models in a matter of seconds.

REQUEST A DEMO

Graham also updated her existing Applicant to Enroll model to incorporate the new engagement score. Graham further improved the model by working with Rapid Insight’s support team (which provides free and unlimited assistance for users with data prep and modeling projects). She created new data points (such as “distance from campus” and “days between application creation and submission”) in Rapid Insight’s Construct data wrangling software.

The model automatically updates daily, so the information is always relevant to admissions staff.

With predictive models updated for the fully-virtual application cycle, Lipscomb had an even greater ability to target outreach to likely enrollees.

Better-informed Recruitment Decisions

Using the predictive model results, Lipscomb’s admissions team can more readily target “persuadable” students. Admissions staff receive streamlined reports (which are automatically updated daily) which drill down to the student-record level.

Sending recruitment materials to applicants ranked 5 (highlight likely to enroll) would not be the best use of resources since they will most likely enroll without the extra encouragement. Similarly, sending materials to those ranked 1 or 2 is also not likely to have a dramatic impact.

Instead, Lipscomb’s recruitment team targets the applicants ranked 3 and 4 plus the lower-tier 5’s) – those on the fence and would benefit from outreach pushing them towards enrollment.

Before the development of the Enrollment Engagement Score and Graham’s updates to the predictive models, Lipscomb’s recruitment team expected that, as a result of COVID, they’d lack the type of directed outreach that led to success in the past.

Instead, thanks to Graham’s efforts and Rapid Insight’s flexible tools, they had a functioning predictive model and could conduct targeted, efficient outreach. “Before the new model, it felt like we were just throwing a dart out the window,” Graham said. “This way, you’re throwing a dart at a true target.”

Looking to the Future

The enrollment engagement score and model addressed Lipscomb’s immediate needs in 2020, but Graham plans to develop the model further and produce even better results.

She plans to compare engagement modeling scores to Lipscomb’s geomarkets to see if the data supports their perceptions about their strongest markets. The enrollment department will use that information to identify markets where the school is succeeding, even if it isn’t currently aware of that success. This understanding will help Lipscomb identify and capitalize on new market opportunities.

Additionally, Kelley believes that the engagement model will remain accurate and valuable once Lipscomb begins hosting live recruitment events again. “This year has been different, but the model is resilient enough to handle both on-campus and off-campus events,” Graham said. “Moreover, it’s made us a more data-aware and more data-driven institution.”