Using First Principles to Push Past the Hype in Edtech

A lot of what passes for "innovation" in edtech does not go far enough, because it is not grounded in first principles of the learning sciences

This piece was co-authored with the Lifelong Kindergartner Yusuf Ahmed. Hopefully, the first of many more to come. Make sure to follow him on twitter.

---

It has become widely accepted that a grounding in first principles drives breakthrough innovation. Sadly, this approach has rarely been the reality in edtech.  

Shiny new technologies are too often combined with a shallow understanding of how people learn. Claims about innovation often amount to little more than rearranging chairs within a traditional classroom. Surface features may change, but the fundamental design remains the same.

As software eats the world, first principles thinking requires a three-pronged approach: developing deep domain expertise, grappling with the possibilities of new technologies, and understanding user needs within the context of complex systems. You can think of these as three legs on a stool. 

Breakthrough innovations engage all three legs of the stool. While others have written about understanding users and software, we will focus on why a grounding in first principles from the learning sciences is required to create truly transformative edtech. 

While there are many important facets of the learning sciences, the most important first principle for transformative innovation must be a theory of how people learn. 

Being explicit about your theory of learning is the first step to breaking from traditional paradigms and building from first principles. It creates clarity about what you’re designing for. An explicit theory is testable and falsifiable. It forces you to ground your work in experimentation. Without that clarity, we risk moving chairs without meaningfully changing the design of the classroom.

The power of analogy: Moving beyond a “Netflix for Education”

A superficial theory of learning is most apparent in the popular refrain to “build the Netflix of education.” This desire relies on a misguided notion of learning as passive content consumption. This perspective guides how many organizations invest in edtech: focusing on highly produced video courses, with active engagement as an afterthought.

Tech in this model is used to serve passive content, not as a mechanism for active engagement or connection. Such a design works in only two instances: if you are focused on developing an entertainment company or if you are resigned to superficial learning (e.g. focused on rote memorization). 

While a content recommendation engine can be a powerful part of a learning experience, content alone doesn’t drive deeper learning. It needs to be paired with contexts for practice and opportunities for transformation (i.e. a learning pathway). Content alone is your answer when you believe that people learn by consuming information (whether you are conscious of that theory or not). 

Applying first principles points us in a different direction: instead of building the “Netflix of Education,” we should build “a Google Maps for learning,” a “Bumble for finding mentors and peers,” or “Minecraft for professional practice.” None of these analogies are perfect, they are only better. 

Let’s take the Google Maps analogy. It immediately puts agency into the hands of a learner and frames learning as an active process of discovery and pathfinding. Learning is something that the learner does, not something that is done unto them. Instead of passively watching content, a learner might use a Google Maps for learning to map how to get somewhere (e.g. how to do X or understand Y) from where they are now (using reflection or formative assessment tools). 

This draws on research from the learning sciences that sees learning as an active process that “leads to change, which occurs as a result of experiences and increases the potential for improved performance and future learning.” Importantly, learning happens much more effectively when the learner has agency - when it is something “students do themselves.” 

To be clear: starting with the learning sciences doesn’t point in a single direction. Companies like Quizlet focus on applying cognitive science principles to aid memorization. In stark contrast, companies like Synthesis leverage constructivist ideas about learning to create simulations where kids work through complex problems with peers.

Despite their differences, these companies both start with a clear theory of learning that (a) drives better product development and (b) creates opportunities for improving a product over time as your core ideas about learning are tested and refined. 

Going beyond the (cohort) hype

Recently, there’s a lot of hype around Cohort-based courses (CBCs). But here’s a question: universities have used cohorts for centuries: what’s different this time around? 

What makes the best CBCs special aren’t cohorts, but how they bring learning principles to life. This involves three features: (a) a clear theory of active learning (b) systems to operationalize that theory (c) and feedback loops with strong accountability that drive improvements over time. 

The current wave of CBCs is nascent -- think MOOCs circa 2012. The winners in the space will leverage learning principles to create robust communities, attract members who add value, and improve over time. 

Bringing it all together

The current pandemic is a once in a generation black swan event. While it’s reaped devastation on families around the world, it also offers a possible inflection point: do we continue rearranging chairs within the architecture of a traditional classroom? Or do we use this opportunity to build new infrastructure to support lifelong learning? 

Grounding our work in first principles offers us a way forward: one that better positions investors, operators, and system leaders alike to build tools and learning experiences that actually support deeper learning. It allows us to distinguish superficial innovation and tech from products that actually drive value for end-users.

Humanity has had these kinds of windows before in the past. Let’s not whiff at this one. A grounding in the learning sciences won’t guarantee our success (there’s luck, and the two other parts of the stool) but it certainly increases our chances of transforming education for the better.