Should Universities Be More Like Amazon?

I was recently asked to imagine how universities could better serve students with lessons taken from major online experiences like Amazon. The question led me to consider what it is about these experiences that actually helps to serve the customer, leaving aside for the moment the possible difference between customer and student.

When I think of Amazon’s main features, I think of the recommendations based on my previous viewing habits, shopping habits, and those of customers like me. I also think of the many reviews sourced from other customers which help guide my decisions. Amazon stores my information, texts me updates, and even lets me subscribe to deliveries of cat food to further ease my life.

When I think of how universities serve students, the most applicable analogy that comes to mind is the registration system. It has been a few years since I was in college, but I remember searching and registering for classes. The only reviews I could find were on RateMyProfessor.com, and the process of searching and registering was clumsy at best, opaque at worst. I certainly didn’t think of registration as a way to serve my needs, the way I may think of an Amazon app. Nor was it an enjoyable browsing experience like Netflix, where I can hope to easily find something I like.

Comparing these two experiences leads me to suggest a few ways that we could improve the student registration experience by modeling it off of similar features in Amazon. While simple improvements in search and interface are always welcome, I’ve focused below on the recommendation system that Amazon, Netflix, and similar sites use, as I think it may have the most potential for interesting outcomes in the student experience. I do think there is room for student ratings of classes, but I think that topic deserves a post of its own!

I can think of several ways that a recommendation system could help students make better decisions about their class schedules while also improving the classroom experience for everyone.

• For example, imagine if you could recommend classes to students with the express goal of improving diversity within a classroom. Could you raise the number of female students in engineering disciplines, simply by suggesting that a class was suited to them?
• What about your pre-med students taking biology who still need to fill their humanities distribution? The system could suggest a class on Darwin and philosophy, helping them both fulfil distribution and gain a new perspective on core elements of their professional knowledge.
• Perhaps the system could take into account the choices of former classmates and attempt to recommend classes to a student based on what their classmates have chosen, trying to use those social connections to encourage more exploration on the student’s part.
• A robust system could look at a student’s history of classes and suggest extracurricular activities. It can be difficult to navigate all the possible opportunities college provides, especially for first-generation students like myself. Having those opportunities presented while a student is already planning their schedule and plotting their next term could help engage them and further deepen their relationship with the university community.

But of course, these positive outcomes have another side. Recommending classes with the goal of diversifying the student body means using demographic data to inform decisions. I would be very surprised if such a system did not eventually exhibit some racial and/or gender bias. It would be very challenging to ensure that a system could effectively use personal data to guide decisions that were always beneficial. In addition, such a system may encourage administrators to cut back on student advising, but there are some elements of student choices that only a human can discern. I think of noticing when a student’s habits change—perhaps their emails arrive later or not at all; perhaps they are late to appointments when they weren’t before. Perhaps when they visit, they are downcast and discouraged. No system we can currently create could see such signs or address them.

Not all challenges are so insidious; rather, they are simply resource-intensive. Making the Darwin recommendation in my example above requires coding every class in a useful taxonomy. This means faculty would have to submit syllabi early, and someone would have to ensure that the syllabi from last term are still accurate for this term. Even for a small department, this can be a monumental obstacle. A similar challenge exists for the example of recommending extracurricular activities. The administration of student groups and volunteer opportunities would also need to be robust and timely in order for the system to be useful.

But overall, the main question I have is: What is the goal of this system? The goal of Amazon’s system is to sell more products. The goal of Netflix’s system is to get people to watch more movies/shows in order to keep them as subscribers. A university’s goal would, I hope, be different, and here is where I think the question of student vs. customer becomes especially pertinent.

Students do pay us money (tuition) for a service (teaching) and a product (credential). However, that is a very simplistic way of looking at this relationship. Universities exist to do much more than simply sell knowledge and certification. I hope we would all agree that universities exist to change students’ lives, to nurture young and old into wiser citizens, and to broaden minds and widen perspectives.

Before we engage in any such systems, we must consider what we are trying to achieve. The purpose of Amazon’s systems is to sell (me) more things, and it does so quite successfully. But the purpose of a university system is to serve the student better. If a system could be built to both nurture a student’s educational journey and support their decision-making process while also working to make that journey the most diverse and productive it can be, then great. But that’s a far harder task than just selling a few more sundries I hadn’t realized I needed.