Flexible Tools, Teams, and Thoughts: Adaptability in Enrollment ModelingReading time: 4 minutes
Over the past decade, higher education enrollment professionals faced a series of challenges to traditional admissions practices: a declining birth rate, financial challenges for institutions and students, and a reevaluation of standardized testing as an admissions metric.
In 2020, COVID-19 drastically compounded the situation. Staff and students were forced to work, teach, and learn remotely. On-campus recruitment events and tours were canceled entirely.
According to the National Student Clearinghouse Research Center, 2020 postsecondary enrollment declined at twice the rate it did in 2019. COVID-19 hit community colleges and private schools hardest; meanwhile, graduate and law schools generally saw increased enrollment. Without a shared understanding and experience of the crisis, a formalized, unified response was difficult to determine. In other words, most institutions had to learn what worked and what didn’t through trial and error.
Additionally, many high school graduates who would have enrolled in college under ordinary circumstances decided to take a gap year instead. Refunds and countermeasures to the pandemic for physical and digital classes created extreme budgetary pressure.
Confronted with these unprecedented circumstances, many enrollment managers worked under the reasonable assumption that predictive modeling, which relies on historical data to make forecasts, no longer had utility. After all, if the data from before the pandemic no longer applied to the new landscape, how could it possibly be a reliable predictor?
The Continued Utility of Enrollment Modeling
More than anything, the pandemic underscored that unexpected events can and will happen. While it is obviously impossible to forecast the exact nature of unpredictable events, colleges can establish good modeling practices that will prepare them to respond to the unexpected, whatever the specifics may be. In fact, institutions with flexible teams, tools, and thoughts are already developing effective solutions in the new enrollment landscape.
There are certain advantages to working with a consultant, whose industry knowledge might be applicable to the challenges you face at the moment. But there are also restrictions in the types of data and models you can expect when working with a consultant, especially during times when no one knows for sure which path forward is the correct one.
Rather, there are significant advantages to working with in-house analysts over consultants. As it relates to enrollment modeling, this chiefly means the capacity to quickly adapt models to a changed reality rather than adhering to pre-determined, periodic reports.
For example, when faced with the reality that in-person campus recruitment events were no longer possible, Lipscomb University’s enrollment team found alternative methods for measuring engagement, such as email replies and virtual interactions with recruitment materials. Using this data, they developed a digital engagement score and corresponding predictive model. The model allowed their admissions team to prioritize outreach to prospects who would benefit most from direct contact.
Meanwhile, Saginaw Valley State University’s retention team incorporated live Learning Management System data from online classes into their student success model. With this information, SVSU’s counselors could identify and connect with students who needed academic assistance most.
In both cases, creative in-house analytics teams enabled the institutions to change course and find success in response to unforeseen circumstances.
In addition to creative teams, rapid shifts necessitate flexible tools. An in-house team with transparent analytics software that allows for easy gear-shifting will enable you to change inputs and outcomes in short order, whatever the specifics may be.
For example, enrollment modeling teams often run into roadblocks when deciding which variables are the most statistically significant contributors to an outcome. When using Excel, Python, or other tools commonly used to build predictive models, variables most often need to be tested on a case-by-case basis and selected manually for inclusion in the predictive model.
This slow process is less than ideal when your institution needs immediate answers. Equipping your team with tools that automatically test for and include the most statistically significant variables is a major asset in remaining responsive under pressure. And the ability to generate unlimited models at a fast pace means your team won’t be restricted in the questions they can ask or the answers they can seek.
It is important to remember that predictive models give you a forecast of probable outcomes based on your data. In some cases, the outcome will be something you intend to change.
For example, you might predict that you will not meet your enrollment goals. Understanding this early in your enrollment season can motivate additional recruiting efforts and justify modifications to acceptance parameters.
Models do not necessarily predict a future that you are required to accept. In many cases, they alert you to take corrective action and reach more desirable outcomes. The more flexible and responsive you are with your strategic approach, the more likely you are to find success when faced with unpredictable circumstances.
Interested in learning how Rapid Insight’s intuitive modeling tools and experienced analyst support team can help you achieve your institutional goals? Click the button below for a demo!