Andrew Drinkwater
Over the last few weeks, we’ve talked about how enrollment targets are created (Part 1), how they become a reality from a business process perspective (Part 2), and where institutions get student with target setting (Part 3). Today we look to the future, and discuss how Predictive Modelling and AI are transforming enrollment planning.
Enrollment Forecasting is fundamentally a predictive modelling process: given a series of inputs, what are the outputs likely to be?
But enrollment target setting is a little different. It’s a human process, informed by data: what is our strategic goal, and how do we get there?
Both predictive modelling and AI could help an enrollment planner create better targets.
Predictive models can be used to determine which factors are most likely to influence success. Rather than having rigid targets around particular programs, it is theoretically possible to make targets based on certain kinds of experience or other characteristics (while being mindful of the ethical implications of some choices in this space).
By way of example, perhaps students who completed an International Baccalaureate program are 5x more likely to complete their studies as other students. Many institutions will target IB students specifically both because of their likelihood of success (and higher customer lifetime value in many cases) and because the process by which their application is evaluated is different.
With enough data on past precedent, predictive models could also be used to determine realistic targets. Perhaps you are able to combine labour market data, demographic data, your own internal data on student enrollments, and past target data. A predictive model could take these data sources and predict a realistic target in the future. Artificial intelligence could be used in conjunction with this method to: find the external data, propose a methodology, write the code, and potentially even run the model (please ensure you’re taking proper care of student data).
Artificial intelligence could be used in many ways to support the target setting process as well.
One avenue is to use artificial intelligence to critique the existing targets. By combing through past data and combining it with public data, AI could suggest which programs should have targets raised or lowered.
AI could assist with research to help determine which areas are in demand and which areas are declining. It could be used to suggest which programs should be discontinued or revised substantially.
As we’ve shown in some of our recent presentations, AI can also be used to help understand the targets as they exist. It can summarize and help you understand how targets are changing over time, what your future campus will look like, and identify pressure points before they come issues.
We can also use AI to make the targets process more human-friendly. For example, at Plaid, we’re experimenting with using AI to devise the targets themselves, so that you can set your targets based on an aspiration (ie: “grow domestic enrollment by 5% per year”, with Plaid Forecast determining which programs this gets applied to. In this case, we’d apply it to all programs, with the exception of those for which our AI knew there were hard constraints).
As AI capabilities continue to develop and models be refined, the possibilities for its use to improve enrollment target setting processes will expand. Keep up with us as we continue to explore and review the latest developments in AI and potential impacts for education technology and enrollment management: