I’ve been a big believer, probably for the past six to eight years, that those of us in the enrollment management (EM) profession would have to, and frankly should start leveraging advanced analytics to confront the market, demographic, and the financial realities that would face EM leaders in this decade and beyond—and that was before Covid.
Yet whether it was when I was in the hot seat like all of you, or in my current role as a member of a team delivering advanced analytics to higher education, I have found that many don’t understand nor see a reason they should move from descriptive and diagnostic analytics, which look backward and provide insights into what happened and why, to advanced analytics, predictive and prescriptive analytics, which provide foresight and insight to predict what will likely happen and how the leader can influence that outcome[1]. Then, it hit me as I was planning a drive to a location that I had to get to by a certain time, that I may have come up with the best analogy!
Getting to Your Destination on Time
Think about this: you are planning a driving trip that will cover a lot of distance and take a lot of time. You must arrive at your location at a specific time. You can arrive early, but not too early, and you can’t arrive late. This is a lot like achieving next fall’s enrollment goals – it is a long trip and leadership has given you some challenging goals. You can exceed the goals, but not too much, but you cannot fall short, especially the revenue goal! In the current enrollment environment, this requires just about everything to go right!
During the trip, you’ll pass through urban, metro, suburban, and rural areas. You know from prior success in getting to your destination on time that there is one ideal route, but that there are other routes that will get you there on time as well. The challenge is:
- you don’t know much about these alternate routes and you have no insights on whether you should choose one of them because of problems along the primary route
- the weather forecast indicates that there is an over 50% likelihood you will encounter severe storms along the way.
- you will be moving through some of the urban and metro areas around rush hour and it is road construction season.
My question is, to take this trip and have the highest likelihood you will arrive at your destination on time, would you use a paper map, or would you use a smart map app like Google or Waze?
I am guessing that all of you said you’d use a smart map app unless you are a lot older than me and I’m old! Why? Because you know that, despite your knowledge of the route and extensive experience making this trip, the app will pull in a lot of data from other app users and sources about current road conditions, traffic density, accidents, etc., and it will use advanced machine learning to suggest alternate routes automatically and keep you on track to reach your destination on time.
Looking in the Rear-View Mirror
Diagnostic and descriptive analytics are like a map and your experience. You’ve driven to the location before and most of the time you arrived on time, until recently when unforeseen conditions have caused you to be late. As you progress on your drive, you know if you are ahead of schedule or behind, but at any specific moment you don’t know what lies ahead of you and if conditions have changed. If you run into heavier than usual traffic, you may not know what to do to stay on time. From the map and experience, you may know you need to take an alternate route, but you don’t know if that will help you stay on time or delay your arrival further.
This uncertainty, am I really on track to achieve my goals, is pervasive throughout the enrollment funnel.
For instance, if you are up in inquiries and applications, or it is April 15 and you’re up in deposits, experience may tell you to rest easy, you are in great shape to achieve your enrollment goals. But are you really in great shape? Are this year’s inquiry, applicant, and admitted cohorts identical to years prior? Is their behavior the same, e.g., did the same proportion visit campus, what was the time between inquiry and application for these students, and have they responded the same to your publications or e-communications? A difference in any or all these behaviors, plus literally hundreds of other behaviors or demographic realities, e.g., where they live, socioeconomic status, etc. could be indicators that you are not on track to achieve your enrollment goals.
The Road Ahead
Predictive and prescriptive models are like the map apps you use every day. They are advanced analytical models driven by machine learning. Using machine learning, in this case, means that the models can process a lot of data from many sources quickly and render a much more dependable indication of whether you are on track to achieve your enrollment goals, and if not, offer prescriptions, actions you can take that will improve the probability that you will achieve your goals. Just like the map app which is providing you with an estimated arrival time and recommending course corrections as required, predictive and prescriptive models are dynamic. Machine learning means these models are constantly updating as they get more information from student behavior or additional factual data, for instance, an admitted student pings on your website or who sends in a FAFSA and when, or hundreds of other indicators.
There has been recent criticism that advanced analytic models can have negative consequences in higher education. However, just like the smart map app does not make the driving decisions for you when it offers an alternate route, predictive and prescriptive models do not make decisions, leaders do. So, if the model offers suggested actions that are not consistent with institutional mission and values, then it is the leader’s responsibility to weigh the recommendation and act accordingly. Further, through the power of these advanced analytic models, enrollment leaders can run simulations of different decisions, which allows them to manage the oft-times conflicting goals that they are given by university leadership while making decisions consistent with institutional values and goals.
I still like having a map to begin planning my trip. I still have them in my glove box or the storage along the door where I put them years ago! I like to see the “big picture” view of the route. Couple the map and my experience, I have a sense of when I need to leave, where I will stop for breaks and get gas, and the sites along the way. But as I get closer to my departure day, I start plugging my destination into a smart map app to determine when I need to leave to get to my destination on time, and the smart map app stays on the whole way.
Mapping Your Enrollment Future
Are you ready to start using a smart map app, i.e., leverage predictive and prescriptive analytics, to inform your decision-making and guide you toward achieving your enrollment goals? If so, I look forward to talking about your market and your challenges, and how advanced analytics can help. CLucier@liaisonedu.com
Resources:
[1] This report Machine learning in higher education | McKinsey provides an excellent resource to understand the data analysis framework