Seeing the Systems in EdTech
How ‘pace layers’ can help designers understand complex systems
As problem solvers in the volatile world of EdTech, how can we ensure that our efforts are meaningful, and avoid getting lost in superficial concerns? Understanding the many-layered systems in which we work is key.
The tantalizing power of systems thinking
UX practitioners in EdTech work at the intersection of staggeringly complex systems. The field of formal education is in constant flux, as new trends and research push up against old practices, educators attempt to navigate multiple layers of laws and requirements, and school curricula become political footballs. Where all of that overlaps with the dynamic and chaotic systems of technology, from which our products are developed, things can feel utterly overwhelming.
Amidst all this complexity, it is common to hear UXers in this space trumpet the value of systems thinking. Whereas mere mortals are stuck in an endless game of Whac-a-mole, in which for every problem squashed two more pop up, systems thinkers are able to design solutions that work elegantly in the big picture and in the details. They are like Neo, once he is able to recognize the Matrix and draw power from his new understanding.
But what actually is systems thinking? Can we define it, or do we just know it when we see it?
Seeking a better understanding led me down a rabbit hole that brought me here, to a particular framework for systems thinking known as pace layers. It’s a concept that is both actionable and profound.
The system and its layers
This story starts in the 1970s, when an architect named Frank Duffy conducted an analysis of money spent on a building over the course of its existence. He found that, typically, only a third of the cost goes into the original construction, while the rest is used for ongoing maintenance and remodeling. In other words, the bulk of the expense is not in the creation of the thing, but in its use over time. At the root of this finding was a key insight:
“Our basic argument is that there isn’t any such thing as a building. A building properly conceived is several layers of longevity of built components.”
Duffy came to see a building not as an indivisible whole, but as a system composed of layers, each with distinct lifespans. He called these “shearing layers” to reflect the fact that, by their interconnectedness and their different rates of change, a building was constantly in the process of tearing itself apart.
From this understanding emerges a clear lesson. Since each layer moves at its own pace, if we constrain the rate of change of one layer to satisfy the needs of another, we create complexity, headaches, and extra cost. On one level, this conceptual model leads to seemingly obvious rules, like “don’t pour the concrete foundation over the electrical wiring.” On a deeper level, it implies something essential about design:
“The unit of analysis for us isn’t the building, it’s the use of the building through time. Time is the essence of the real design problem.”
The architectural principle Duffy describes is conceptually similar to what we today call responsive design — the idea that what we design for screens should adapt gracefully to fit into a variety of screen sizes. Just as the contents of an app or website must flow elegantly into any two-dimensional space, the design of a system must be responsive to the passage of time.
In 1994, the writer and thinker Stewart Brand expanded on Duffy’s work with the publication of How Buildings Learn, the book from which the above quotes from Duffy were sourced. Brand sought to understand what led to a building’s enduring success or failure, and came to believe that the interactions between the shearing layers — how they relate to each other — were key:
“How buildings learn over time is that the fast parts gradually suggest things to the building which get integrated into the slow parts. That’s how buildings become wise over time, and eventually become loved.”
Brand found a commonality in buildings that had longevity: a harmonious balance between the layers. Rather than causing destruction, they influenced each other so that the building as a system evolved to meet the changing needs of its occupants.
Zooming all the way out
In 1999, Brand went on to publish a book called The Clock of the Long Now, in which he devoted a chapter to applying Duffy’s core concept broadly to all of civilization, and renaming the shearing layers as pace layers, as shown here.
I won’t attempt to describe each layer, although I highly recommend reading the chapter for more detail. I will say that this grand expansion of the concept brings to the surface a deep insight about the nature of systems, and how they can be resilient to changing conditions.
The upper layers, what Brand calls “Fashion” and “Commerce,” change quickly, constantly, and discontinuously. They are chaotic and experimental. In contrast, the deeper layers display constancy, consistency, and embody the years of wisdom that the system has absorbed. The slow layers constrain the fast layers, resisting manic change based on whims and fads.
Even as they face this resistance, the fast layers play a vital role in the life of a system. As ideas are quickly and constantly auditioned, some will prove valuable and will gradually move downward into the slow layers. Brand summarizes this dynamic in one great paragraph:
“Fast learns, slow remembers. Fast proposes, slow disposes. Fast is discontinuous, slow is continuous. Fast and small instructs slow and big by accrued innovation and by occasional revolution. Slow and big controls small and fast by constraint and constancy. Fast gets all our attention, slow has all the power.”
In a talk he would give years later, Brand noted that a large portion of the response to “How Buildings Learn” came, unexpectedly, from the tech community, written as though the book was about the design of software systems. Many attempts to apply the pace layers framework to software development have followed in the ensuing years.
The layers of resilient software
One of the most influential adaptations of pace layers to the practice of software development came from the consulting firm Gartner, in 2012, in the article Accelerating Innovation by Adopting a Pace-Layered Application Strategy. The article lays out a fundamental tension software companies face between the need to quickly and nimbly address problems and opportunities, and the need to establish systems that are stable, reliable and secure.
Gartner proposes to solve this problem by defining the architecture of an application system as composed of three main layers.
- Systems of Innovation are where experimentation happens as new ideas are conceived. The point here is to throw spaghetti against the wall and see what sticks. Stakes are low, change happens rapidly, and work gets thrown away frequently.
- Systems of Differentiation are a company’s special sauce. As new ideas prove their value, they graduate from the fastest layer and move here, where they will stick around and help differentiate a company from its competitors.
- Systems of Record correspond to common ideas in business, widely accepted functions that any company in a given industry must serve. This layer is mature, and slow to change. The point here is not to be different, but to be more stable, more efficient, and more performant.
To bring it back to our world of EdTech, let’s say we think we have a brilliant idea for how to connect student assessment to practice and instruction, but turning it into reality would require building new and complicated ways to share, store and interpret student data between products. Rather than making a large infrastructural investment upfront, Gartner’s recommendation would be to hack together a cheap interim solution that gives it an opportunity to prove its value. As we evaluate, we gradually move the things that are working deeper into our architecture. In this way, the system adapts over time to incorporate the few ideas that turn out to be good, and we avoid making large bets based on guesses.
Deep applications for UX
While the Gartner piece was intended for system architects, there have been many compelling applications of the pace layer framework within UX, to design, research, and our workflow within the development process. When asked about the similarity between the pace layers of civilization and Jesse James Garrett’s landmark Elements of User Experience diagram, which was published around the same time, Brand noted that the concept is fractal, meaning that within the multi-layered super-system of civilization exist smaller systems, each of equal complexity. The ever-evolving system of UX design contains its own set of pace layers.
One way to envision the pace layers of UX is through the tried-and-true game of 5 Whys, applied to the UX practice itself. As each answer takes us a layer deeper, note how we get further removed from tools and technology, and closer to our underlying and enduring purpose. In the context of EdTech, the exercise may look something like this:
Why are we using Figma? (Tool layer)
To prototype an experience for teachers.
Why do we need that prototype? (Artifact layer)
To see if teachers understand how to use our design.
Why do we want teachers to understand how to use our design? (Workflow layer)
To see if using this design will help them accomplish a task.
Why do we want them to accomplish this task? (Strategy layer)
Because we believe that by accomplishing this task, teachers can more effectively help students learn.
Why do we want students to learn? (Mission layer)
Because that is our mission.
Of all of the attempts to define the pace layers of UX, one of the most meaningful comes from Cyd Harrell, a civic and service design leader. In a podcast discussion with Jorge Arango (who created a version of pace layers for design from which Harrell’s version is adapted), Harrell warns of the limitations of our fixation on the faster, surface-level layers at the expense of deeper, longer-lasting institutional layers:
“A lot of the conversation around user experience right now is kind of around optimization… making more exciting, more delightful, digital products that convert or that sell or that sell ads… working in small cycles in an agile way that…suits the kind of metrics, methods of the time and that I think actually makes it difficult for UX practitioners to really think about the deeper people-oriented values that we have to get into.”
This warning feels increasingly salient in an era in which the latest Figma features and AI accelerators have come to dominate the discourse. Paradoxically, as our tools continue to get smarter, faster, and more automated, they seem to demand more and more of our attention. Optimization in the fast layers of UX is absolutely wonderful, as long as we recognize it for what it is: a means of working on the fast layers even faster, bypassing less meaningful work that, in the past, would eat up the lion’s share of our time. Its value is that it allows us to think about those things less, and to invest our recovered time meaningfully. When we fixate on shiny objects and navel gaze about our own tools and workflows, we become distracted from the reason we design in the first place.
Our obsession with “delight,” too, can be a distraction. To delight our users is a noble and worthy goal, but to elevate it as the pinnacle of achievement in design is to get swept up in the fast layers. Purpose lives in the slow layers.
Where do we go from here?
So, what does it mean to focus instead on the slow layers? It means understanding our users’ deep needs, figuring out where they intersect with our values and purpose, and working our way up from there. It means influencing our organizations’ policies and practices, learning how decisions are made and by whom, and helping to guide our priorities to align with our mission.
To work on the slow layers, we need to slow down ourselves. We need to be careful and deliberate, gathering a wide range of perspectives on what we’re working toward, and why. We need to heed researcher Erika Hall’s great advice:
“You just need to sit in a room with people representing different disciplines: researchers, designers, writers, business leaders, technologists. And then you need to ask: ‘What does everybody hope happens? What does everybody fear? What could go wrong from each party’s perspective? What are the contexts we need to consider? What do we have a reasonable certainty about, and where are we unsure?’
“If you work in design today, 80% of your job should be that. It should be talking to people. And then once you’ve done that, and a concept emerges, you can spend a little time making stuff.”
Being at the hectic intersection of education and technology can feel overwhelming, which makes it all the more imperative for us to take a breath. Seeing the system for its layers allows us to step back from the medium of our work and focus on the outcomes. It brings clarity about the very human reasons why we do what we do.
I’ve read a number of EdTech mission statements in researching this piece. Most do not mention technology at all, and those that do only do so in service of the larger goal: to help students learn. Let’s design at the deep layers to drive that purpose for all learners, ethically and effectively, in a way that will last.
The UX of EdTech Community’s purpose is focused on helping UX practitioners who work on products and services that support learning. We also welcome educators and those in other disciplines looking to develop a UX mindset applied in the industry of education and future of learning and work.
Subscribe to our newsletter for more EdTech-focused UX resources and job opportunities.
