Cohort Based Courses: Something old, Something new

Cohort-based courses are all the rage these days. I am taking one now with David Perell.  The course has been an amazing educational experience. Above and beyond anything I expected, prompting me to question my own role as an educator. I spent 22 years in school (yes, I counted) and 16 years teaching. Am I doing it wrong all these years?

Cohort-based learning is a collaborative learning style where a group of students progresses through an educational program collectively.  You are probably asking yourself how is that different from the traditional way of learning. It’s not. This has always been the academic style of learning, ever since the School of Athens and the Academy of Plato and Socrates.  If you were my student at Wharton or Kellogg, most likely, you studied in a “cohort-based course.”  

How is that different?

The difference is that the term is now used to describe online courses, which were traditionally self-paced.  And indeed, the last few years have seen a significant explosion, both in interest and in funding going to these Online Cohort Based Courses (CBC): Scott Galloway announced the $30M dollars raise to his Section 4 startup. Wes Kao (who ran altMBA with Seth Godin) and Gagan Biyani (founder of Udemy) have launched a startup aiming at democratizing the ability to launch such courses. And Reforge announced they raised $21M from Andreessen Horowitz to help them build a school devoted to Cohort Based Courses on the topic of scale.

Back to the one that I am taking now. We have multiple meetings a week as well as assignments. We can attend multiple optional groups, where we discuss different aspects of the course (which feel like TA sessions). We have a discussion board, where topics and off-topics are discussed. Until now, it should sound somewhat like a traditional course, just online. The first difference is that we have alumni from previous cohorts actively joining the sessions and discussions. The second is the fact that the course experience is much richer at 200 people as compared to 20 students.

But none of that demonstrates the true experience of being in such a course. As David Perell describes it: it’s like Coachella for learning. I have not been to Coachella (count the years I been in school and you will understand why), so I will have to rely on his judgment here. But the idea is clear: there is the main stage (the main weekly sessions led by the main instructor), and there are many other side events: TA sessions, alumni events, speakers. We even have a cross-fit for writing.  You are not expected to attend all of them. The reality: you don’t sleep so you can attend all of them.  You don't want to miss a single opportunity to learn.  It’s exhilarating.  This has been my best learning experience since “Brownian Motion and Stochastic Calculus” with Karatzas. 

What makes these courses work? Tiago Forte, one of the early developers of these writes that 

The learning experience that is emerging resembles a video game or a virtual world as much as it does a university classroom. Polls, interactive whiteboards, and emoji reactions enable many-to-many communication that can keep hundreds (or even thousands) of people engaged at once…. allowing the excitement, the joy, and the fun of learning to shine through more strongly than ever. 

Tiago outlines several dimensions along which Cohort Based Courses improve over existing online courses:  creating a community, which increases social accountability, and thus further increasing the impact of the educational experience. In his article, Tiago goes back to MOOCs (Massive Open Online Courses) and describes CBC as the first internet-native education experience. 

History of Courses

The main tradeoff we face in education is between scale (or reach) and quality of outcomes: making an impact on those learning and endowing them with skill or knowledge that alter their future.  The first innovation that altered the scale of education was the Gutenberg press. The fact that any idea could be printed transformed the ability to distribute knowledge at scale. But, as Andy Matushek writes, Books Don’t Work. Retention is very low and people barely learn from just reading books. 

The internet reduced the cost of writing and distributing knowledge.  Retention is still low. In fact, my colleague and I show that with the abundance of information sources, we know even less. 

MOOCS took that idea and improved it in two critical ways: videos and testing. Most people prefer watching videos to reading since multiple senses are involved. But it’s deeper: the fact that learning is personified through an instructor helps increase the authenticity of the content and thus learning. MOOCs providers like Coursera and aggregators like Khan Academy also added short tests and quizzes. While we all hate tests, there is quite a bit of research that tests are helping Make it Stick by forcing students to constantly engage with the content. 

But completion rate of MOOCS is abysmal. Retention of knowledge is usually very low.

The implication is that knowledge was never scarce. We removed the barrier to create and replicate it. We reduced distribution costs. But we did little to make sure it actually becomes immersive. 

While the majority of my education and teaching is done in person, I had early courses on Udemy and Coursera. While teaching these my colleagues and I ran a randomized experiment. We showed that using a very simple set of nudges, encouraged students to visit the discussion board or communicate with other students, the majority of students had a higher social engagement, higher quiz completion rates, and higher course grades. Simple nudges. Quite a significant impact. 

In my experience, Cohort Based Courses are the ultimate nudging experiment. You don’t want to miss a single thing, and through this constant Fear Of Missing Out, you keep engaging with the content, and the community. Pushing yourself and others. To learn. 

Are Cohort-Based Courses Scalable?

So does it mean that we solved the issue of scalability of knowledge dissemination?  When compared to in-person education CBC’s are clearly cheaper and more efficient: the classes have no physical limit on the number of students, and no expensive facilities are needed.  But my experience is that they are not as cost-efficient as they may seem.

The reason for this is driven by the notion of the scalability continuum. On the one hand, you have instantly scalable solutions (such as books and MOOCS). They have unlimited reach at zero marginal cost. The issue is that they are very transactional, and thus have limited lasting impact. On the other hand, you have the highly curated and bespoke MBA experience, which is very relationship-based, and thus inherently unscalable. The issue with Cohort Based Courses is that for them to work, they need to be very relationship-based: the relationship among the students, and relationship with the instructor. And this just doesn’t scale. 

Call it the educational Dunbar number. Dunbar's number is the number of people with whom one can maintain stable social relationships, limited by our cognitive limitation. I claim that for education to be truly immersive, there is a limit on the ratio of students to instructors. Even in an online setting.  

Cohort Based Courses are a step in the right direction. They are more scalable than traditional education. They are more immersive than traditional online courses. The best ones feel more like Burning Man (I assume. Never been) than a college class. But you still need a sage on the stage. And thus, they aren’t as scalable as everyone believes.