The straight story on enrollment’s newest best practice
Many colleges and universities today are realizing considerable benefits from using predictive modeling. However, other institutions are hesitant, choosing to remain on the sidelines.
At first, predictive modeling may feel new and uncomfortable, but so did many of today’s most widely accepted practices: search strategies; The Common Application; and even institutional websites. As with these earlier innovations, predictive modeling is also now becoming an indispensable best practice for protecting the financial sustainability of colleges and universities.
The purpose of this second edition of Innovations is to shed light on how predictive models are being incorporated into enrollment management operations and how these tools help institutions achieve their enrollment and financial goals.
What is a model?
Enrollment officers use the information generated by predictive models to inform their decisions in many ways. The most successful predictive models are guided by an institution’s mission and values, overlaid with a healthy dose of common sense. A good model is created through comprehensive discussion of institutional priorities and goals. The models themselves do not – and should not – control your admission process.
Predictive models are conceptually simple and user friendly, although they can be mathematically complex. All the data about an inquirer or applicant you possess can be translated into knowledge that helps predict the likelihood of a specific behavior, such as applying (or not applying) or enrolling (or not enrolling) at your institution. Taken a step further, your data can be used to determine the optimal amount of institutional financial aid allocated to each admitted student to give you the best chance of influencing each candidate’s decision in your favor without breaking the bank.
While no one can predict the behavior of individual students, a predictive model shapes an institution’s aggregate results. Put another way, it’s about success with the overall class, not with every individual student.
Misperceptions about predictive models.
Predictive modeling provides many clear advantages. However, let’s first make sure that some basic facts are understood:
- Predictive models are not all the same. Predictive modeling should not be an off-the-shelf process. Because each college or university is different, each predictive model should be as well. The most successful models are far more than just “math.” The design and impact of a predictive model should be driven by the people who create and use it – and that means people from all levels of the institution as well as the enrollment managers who help translate a strategic vision into an achievable strategy. That means your model must be specifically appropriate to your institution’s values, objectives, strategies, and operating conditions. As customization decreases, so do the institution’s long-term benefits.
- Predictive models do not substitute for your judgment. A predictive model is much like a car. The car doesn’t drive wherever it wants. You drive it. It is a powerful machine, but you – as the decision maker – are in the driver’s seat. You determine where you wish to go, and a predictive model is designed to help you get there. If you want your model to help enhance diversity and increase student access, that’s where you’ll drive it. If you want to increase revenue and the size of your enrolling class, the model will strive to achieve that reality. You pick your enrollment destination, and a good predictive model will work in conjunction with your recruitment processes to get there faster and more efficiently. In other words, don’t speed past the strategic dimensions of building your model.
- Predictive models are not too complicated. While the process of generating a predictive model should be left to a qualified statistician, its use is quite manageable. And while the mathematical methodology is intricate, the student characteristics that predict the likelihood of application or enrollment nearly always align with an institution’s internal understanding of what it does well and where it needs help. In the end, predictive models provide accessible, highly relevant information to the people who make critical admission and financial aid decisions.
With predictive modeling you can:
- Visualize your class before admitting a single student. One of the great strengths of predictive models is that they allow decision makers to visualize the results of their stated goals well before admission decisions hit the street. For example, you can use a predictive model to see what would happen to your enrollment and revenue if you dropped your discount rate by five points or raised it by three points. Or you can find out what it would cost to enroll more out-of-state students. It allows you to simulate possible results of varying tuition increases or explore the feasibility of increasing class size and revenue targets. Predictive models allow enrollment managers and institutional leaders to be proactive rather than reactive in enrolling incoming classes.
- Offer more precise and effective financial aid awarding. Most colleges and universities rely on a “matrix-based” approach to determine awards for various groups of students, usually by internally defined measures of student “quality” and need. This method can be used successfully when the matrix is updated every year through analysis of prior-year results. However, a significant amount of enrollment and revenue may be lost by awarding fixed amounts or percentages of institutional aid to large groups of students. Predictive models can be used to transform a matrix-based approach into a highly refined, individualized approach. They do so by specifying the exact amount of institutional aid for each individual student needed to achieve ambitious yet realistic enrollment and revenue goals.
- Focus your recruitment and marketing resources efficiently and effectively. Well before students apply, institutions often spend a significant amount of money to buy thousands of prospective student names, employ a “search” strategy for these students, and engage in expensive marketing campaigns over multiple years. A predictive model applied to prospective students at the top of the admission funnel enables an institution to identify its best prospects, its “fence-sitters,” and those who are unlikely to apply. Using a model that predicts application likelihood can save even a smaller institution tens of thousands of dollars in marketing costs while strengthening application volume.
Use the best tools, get the best results.
Predictive models have been essential tools in market research for two decades or more, influencing our daily lives in many ways as consumers, investors, readers, and viewers. It’s no surprise that predictive models are now being employed as a best practice by colleges and universities, too, as enrollment managers utilize the most sophisticated techniques available to ensure robust classes and revenue.
While few chief admission or financial aid officers sleep well at night in April or May, those who employ predictive models get a little extra rest, knowing that they have probabilities on their side. Having a predictive model liberates them to be as productive and proactive as possible.
Maguire Associates offers the most comprehensive, customized, and collaborative enrollment modeling services in the United States. Our experience and success is unmatched – with more than a 90% historic client-retention rate. We provide analysis, modeling, and unlimited consulting with EMPOWR for application recruitment and EMFASYS for strategic financial aid awarding and pricing.
We at Maguire Associates have focused on our mission of advancing higher education through innovation and insight for over 25 years. Predictive modeling is a powerful Maguire Associates innovation that delivers an array of actionable insights. Visit our service pages and contact us to discuss your institution, our experience, and how we can help you use the best tools to get the best results.
Learn more about: