New York Times columnist Bret Stephens writes in his article How Plato Foresaw Facebook's Folly, “The deeper reason that technology so often disappoints and betrays us is that it promises to make easy things that, by their intrinsic nature, have to be hard.”
Companies have entered the higher education world, purporting to be an IBM Watson for enrollment management—artificial intelligence that can provide the best financial aid leveraging at the click of a button. But those teams that build these "artificial intelligence algorithms" for use in admissions often lack any significant knowledge or experience in the field themselves.
We at Maguire Associates don't shun machine learning models—in fact, we test and use several methods frequently in our work to increase accuracy. But machine learning and automated modeling still require the human touch and the critical thinking that evolves from decades of experience.
Here are some problems with these A.I. offerings:
- The outcomes are only as advanced as the data fed to the algorithm, so it can't account for qualitative information that can be gleaned from experience.
- Machine learning in a vacuum has no way to understand shifts in the industry (regional enrollment surges, like in California), changes in your market (a competitor closing or merging), or decisions that are new to your institution (going to the Common App).
- Teams without admissions experience won't predict the impact of timeline changes without historical data, like Prior Prior Year financial data.
- Machine learning without an experienced guide could make decisions that are counter to your goals and institution's mission, recommending that you decrease aid based on the number of clicks on your website, penalize campus visits, or potentially discriminate according to other metrics in awarding.
Bret Stephens further affirms in his article, "We tend to forget that technology is only as good as the people who use it."
Partner with us at Maguire Associates where we have both the technology and the experiential know-how to develop a model that will meet your goals in net total revenue, headcount, quality, and demographics. You'll understand why those companies touting "A.I." and "machine learning models" aren't a cure-all to the complex challenges of admissions.