Universities confront many of the same marketing challenges as corporations. While businesses need to attract potential customers, universities need to entice prospective students. A prospect needs to be converted to a customer, while a potential student needs to enroll at the university. Once you are an active customer, the business needs to maintain and grow the relationship. University executives should, likewise, nurture the rapport with their students. Attrition is always a hot topic for corporations. Retaining customers leads to increased investment returns. Increased student retention rates contribute to the success of a university, as it impacts competitive standing and prestige. And yes, financial returns are also positively impacted.
Do not jump to any conclusions. The goals of the corporation and those of a university are clearly not the same. However, some of the critical issues facing both are being addressed in a similar fashion.
Mining consumer behavior has become a fundamental mission of corporate thinking. Consumer data is tracked, and targeted strategies are developed. Guess what? Colleges are also monitoring applicants. Enrollment management systems may contain application data, demographics, student performance, results of various marketing programs, and data related to previous student educational performance.
Statistical modeling and Artificial Intelligence are being applied to predict student admissions for various academic programs. Whether it is college, graduate programs, or business school, machine learning tools have proven to be powerful and accurate. In an educational setting, even physician selection for surgical residencies have been improved through these tools.1
What are some of the attributes that a college may employ in their admission process? Integrating academic, demographic, behavioral, financial, and psychosocial variables, colleges can more accurately predict student outcomes and provide targeted interventions. This not only improves institutional efficiency but also ensures that students receive the support they need to succeed in their academic journey.
In an attempt to develop a model for student admission at Duke University, Njokko, employing similar attributes as alluded to above, concludes “Duke University can improve its ability to predict the chance of admission for each applicant more accurately, which could help reduce the chances of admitting students who have a history of poor academic performance.”2
Traditionally, the university admissions process has been labor-intensive, involving comprehensive review of applications, personal statements, letters of recommendation, test scores, and interviews. However, the addition of Artificial Intelligence in this procedure is transforming the way applicants are being evaluated, offering significant advantages.
Employing AI into university admissions processes results in significant productivity improvements. With the growing number of applications each year, review by admissions counselors can be labor intensive, and disposed to errors. AI algorithms can successfully manage massive amounts of data. By automating tasks such as sorting applications, evaluating essays, or assessing standardized test results, AI can simplify the admissions process.
Wait a minute. How are these essays reviewed? In a fascinating application of Artificial Intelligence for college admission, researchers “trained an AI model named RoBERTa to identify seven desirable personal qualities in applicants, using a rating system built by thirty-six college admissions officers. The model was then used to analyze over 306,000 student essays. The results showed a high correlation between AI analysis and human readers, suggesting that AI could process large volumes of applications more efficiently and consistently.”3
Often, corporations use a two-pronged approach for prospect acquisition. Firstly, will the prospect convert to a customer, and secondly, will that customer perform in an acceptable manner? Moore applies the same technique in an educational setting.“4 The first stage examines the admission decision process, while the second stage focuses on the prognosis for degree completion, or perhaps more familiar to us-customer or student retention.
Student attrition, the occurrence of students leaving a university before completing their degree, is also a significant concern for educational institutions. High attrition rates can have a profound impact on a university’s reputation, financial stability, and student morale. As a result, universities are increasingly turning to machine learning to predict, understand, and address the factors contributing to student dropouts. By leveraging vast amounts of student data and advanced predictive modeling techniques, machine learning offers a powerful tool for identifying at-risk students and implementing interventions to improve retention.
Customer retention is vital for businesses as it is significantly cheaper to maintain existing customers than to acquire new ones, leading to increased profitability and higher customer lifetime value. Focusing on keeping customers loyal generates greater returns. Admission is still a significant aspect of higher education, but it is even more important to provide students a route to successful achievement-completion of their degree programs. Recruiting new students is an expensive endeavor. Many universities have huge budgets in order to locate new students, schedule tours, set up meetings, and develop marketing and advertising material. However, if many of the students ‘attrite’ in the first year or so, it may be difficult to show a return on the budgeted expenditure.
In the business world, typical early signals for customer attrition may include decreased product usage, fewer communications with the firm, negative feedback in surveys, and increased questions concerning cancellation guidelines. What sort of indicators dominate student attrition? A variety of factors such as demographic information (age, gender, socioeconomic status), academic performance (grades), interest levels (attendance, participation in extracurricular activities), and even online behavioral patterns may foretell student retention probabilities.
Matz concludes “Analyzing the records of 50,095 students from four US universities and community colleges, [we observed} that the combined macro and meso-level data can predict dropout with high levels of predictive performance.” Further, and interestingly enough, the author comments that the results of their algorithms were performing well at other universities, not associated with the original sample space used to develop the study.5
Others have observed that “using historical click-stream data in conjunction with present click-stream data,” provided an effective tool to “predict dropouts weekly using a simple Support Vector Machine algorithm.”6
Yet others have displayed excellent results by “the use of data acquired from disparate sources in addition to more sophisticated algorithms such as deep feed-forward neural networks.”7
Artificial Intelligence (AI), and its analytic subsidiaries, are increasingly playing a pivotal role in addressing and transforming business issues. Not to be undone, higher education is leveraging this technology, as well. From enhancing administrative processes to providing personalized learning experiences, AI is helping colleges improve efficiency, accessibility, and the overall student experience. As AI technology continues to evolve, its potential to further enhance the college experience for both student and administration is significant.
3 https://verge-ai.com/blog/how-ai-is-being-used-for-university-admissions/
4 https://www.sciencedirect.com/science/article/abs/pii/ S0305048398000085?via%3Dihub
5 https://www.nature.com/articles/s41598-023-32484-w
6 Kloft, M., Stiehler, F., Zheng, Z., and Pinkwart, N. (2014). “Predicting MOOC dropout over weeks using machine learning methods,” in Proceedings of the EMNLP 2014 Workshop on Analysis of Large Scale Social Interaction in MOOCs (Doha), 60–65. doi: 10.3115/v1/W14-4111
7 Imran, A. S., Dalipi, F., and Kastrati, Z. (2019). “Predicting student dropout in a MOOC: an evaluation of a deep neural network model,” in Proceedings of the 2019 5th International Conference on Computing and Artificial Intelligence (Bali), 190–195. doi: 10.1145/3330482.3330514
 |
 |
 |
The Machine Learning Times © 2025 • 1221 State Street • Suite 12, 91940 •
Santa Barbara, CA 93190
Produced by: Rising Media & Prediction Impact
