In article 1 of this 7-part series on data science and AI in higher education, I introduced a simple categorization for thinking about five ways in which data is used and shaping the present and future of higher education: college search and connection data, student retention and success data, student learning/mastery/progress data, organizational health data, and data that enables graduates to connect with other people and organizations. In this second article in the series, I am focusing upon that first category, college search and connection data. Some of my regular readers may be less interested in this category, while this is right in the sweet spot for others. For those who cringe at language about recruiting, admission, and marketing; I contend that what I write here has quite a bit of relevance for the nature of the learning communities in higher education as well, so I invite you to consider this article with that perspective in mind. Regardless, be assured that the next articles will focus upon the learner experience in higher education.
How do colleges find and recruit students?
There is not a single answer to this question. When it comes to how colleges find and recruit students, the answer depends upon the type of student (prospective traditional undergraduates, adults seeking a first degree, adults returning to complete a degree, adults seeking to return for a second degree or part-time graduate student, people seeking full-time graduate study, etc.). It also depends upon the type of institution. Nonetheless, the modern world of data science plays no small role. People search for almost everything online, and that is certainly true when it comes to exploring opportunities for a college degree. However, the dominant ways in which people use the Internet to search for and reach out to possible schools has more than a few limitations. Consider this one.
The Facebook Peer Advice Story
As part of my research about how people search for colleges, I informally interviewed lots of adults seeking to return for graduate degrees. I also identified some of the online platforms, online resources, and communities targeting people who are looking for degrees. Beyond that, I played the role of a prospective student and tried my luck at using Google as I typed in various questions, taking special note to which featured and paid ads showed up for me and followed me around the web (search re-targeting).
However, one of the more interesting experiences came in some of the education Facebook groups in which I participate. I noticed as people asked others in the group for advice on good online graduate education programs. Then I read through the hundreds of suggestions and comments. It was amazing to me how many people claimed that the online degree program of their choosing was the least expensive, most flexible, highest quality, etc. Granted, some of these insights are going to be subjective, but price is not. Yet, countless people thought they were getting the lowest price. Or, perhaps on the subjective side, they thought that they were getting the best value. Looking at the advice as objectively as possible, I can say with confidence that it was riddled with errors and misleading information. This was more like a debate about favorite sports teams than a careful investigation of schools and analysis of the features/attributes that are most important to a given prospective student.
Such folk systems work for people today. They have no small influence on how traditional students choose colleges as well. A reference from a trusted source, even if it is not well-informed or a best fit, is a major factor in people’s decisions. In fact, in a world of information overload, trusted sources, even wrong ones, are increasingly valued by people.
How do students search, learn about, and select colleges?
They Google. They browse YouTube. They talk to peers, mentors, and trusted people in their lives. They read articles and follow social media about various schools. They ask people in person and digitally. They are unquestionably influenced by media, marketing, and advertising. If they were not, there is no way that so many schools would be spending tens of millions (or in some cases more) annually for digital ad campaigns. Yet, I can say with confidence that these ad campaigns do not improve the quality of match or connection between student and school. A school is not the right fit simply because it was willing to pay more than another school to show up at the top of a Google search. This current system is hiding learning communities that would be a great fit for some people, and they are driving people to the learning communities with the most clever marketers or that prioritize marketing in their budget over other things.
The Dominant Metaphor
Right now the dominant metaphor related to college search and connection, especially in the adult and graduate world of part-time and online learning is a sales metaphor. We advertise to students. Even if schools resist the language, they are selling. Students are shopping. As such, much of what goes with that metaphor informs how people select schools. They shop. They are influenced by some brands more than others. Sales strategies and tactics are used and result in students choosing. There is buyer’s remorse in some cases. Even the attitude and mindset of students once they are enrolled continues to be influenced by this sales metaphor. In addition, schools are often thinking about recruitment as sales and competing with similar “products” and “services.”
The Changing Metaphor
There is a strong competitor to the sales metaphor for college search and connection, and it is growing. It is the dating metaphor.
Note: While some might not appreciate the shopping or dating metaphors because they would prefer something from high culture, or at least not from popular culture, these are the metaphors that I notice animating the language and systems that people are using.
With regard to the dating metaphor, think about the idea of a dating website. People don’t shop for a date. They enter data and the system makes recommendations about others who might be a good match. They might also have the opportunity to search for a “good match” but it is qualitatively different from dating. It is more about seeking out a good connection or relationship and less about buying something. In addition, some of the algorithms behind these dating websites are getting increasingly sophisticated, drawing from years of insights about what types of matches did or did not work well.
The dating metaphor points us to the growing role of algorithms and potentially full artificial intelligence to match and connect people with other people, people with organizations, as well as people with various products and services. Consider MyOptions (formerly Admitted.ly), a site that collects up to hundreds of bits of data about a prospective college student, and compares that with the data set they have about a large number of colleges. Based upon everything from weather preference to level of academic challenge, the system makes recommendations for students to consider. While it started as a system only for prospective traditional undergraduates, in my interview with the founder in 2017, she indicated plans to extend it to other populations as well. For better or worse, this is going to be how a growing number of people select education opportunities in the future.
Reliance upon trusted networks and sources will remain a strong (perhaps the strongest) source that informs decisions for quite some time, however. Yet, these algorithmic developments will likely begin to further shape the nature of recommendations within these trusted networks, leaning more and more toward a dating versus a shopping metaphor.
This will change the nature of recruitment and marketing, especially for the adult, part-time and online student. In fact, there is a good chance that it will disrupt the multi-billion dollar business of higher education advertising in the digital space. Without question, there will be platforms that claim to be a neutral source of matching students and colleges (these already exist), but they are actually getting paid by colleges to rank some over others. Yet, the platforms that truly are neutral and that combine both trusted networks and algorithms will win the day. There will also be (or rather there already are) systems that simply rank colleges on the basis of something like the College Scorecard. Those lack complexity and matching capabilities, however, and will likely be secondary to the more personalized and algorithmic platforms.
This will challenge some schools that spend 20% or more of tuition revenue to recruit students through traditional advertising. As more people turn to these “dating” platforms for finding educational options, the advertising wars of higher education may well begin to wane, and I consider that an incredibly positive shift.
Note that there is no small number of entrepreneurs seeking ways to capitalize upon this shift from the sales to dating metaphor. For example, in a 2017 interview with John Katzmann (founder of the Princeton Review, 2U, and Noodle Companies), he articulated a vision for a an Amazon.com of the education space. Granted, Amazon at first consideration sounds like more of the shopping metaphor, but even in my short conversation with Katzmann, I suspect that what eventually comes about will be much more driven by that dating metaphor. And John represents one of more than a half-dozen significant entrepreneurs with whom I’ve spoken who are working on this problem (the higher education advertising wars) and opportunity (making better connections between students and colleges).
There are many important aspects and nuances to what I’ve written here about the metaphor shift, but for the sake of brevity, I offer what I deem one of the most important, especially in the early stages of this algorithmic revolution in college search and connections. Algorithms are created by people and they include the beliefs and biases of people. As such, an algorithm can mislead and do harm as well. It is for that reason that I argue for making these as open and transparent as is reasonable. In addition, when possible, why not give the learner some control over how the algorithms? In other words, instead of having a fixed system that matches students and colleges on the basis of what the platform owner/creator deems most important, why not leave some room for the end user to manipulate the algorithm, playing with prioritizing and re-prioritizing some aspects of the matching to see how that influences the recommendations? Better yet, why not create a system that shows people recommendations/rankings based upon four or five different configurations of the algorithm? This empowers the learners and invites them into being a co-creator of the system as opposed to someone who is just matched and sorted. Agency matters for people and society, and as this is in the early stages, this is the time to think about how we are going to create systems that are more open, humane, and nurturing of agency.