Design Systems Enter the Age of Complexity
When Google introduced Material Design in 2014, it leaned on paper as a guiding metaphor. Digital elements were stacked like sheets on a desk, with seams and shadows telling users what could be touched. As VP of Design Matías Duarte put it: "Material has physical surfaces and edges. Seams and shadows provide meaning about what you can touch." That metaphor worked because users still needed visual cues drawn from the physical world. Nearly a decade later, Google has dropped it, pointing to an expanding computing landscape. People no longer need that kind of help, the company argues—our collective grasp of digital navigation is simply more advanced now.
That shift reflects a broader reality: design systems are being asked to handle far more than they once were. New form factors, affordances, accessibility standards, performance expectations, and pressure to stay visually fresh all pile on top of each other. And the products themselves are growing more ambitious. Instagram's move from a three-dimensional camera icon to a flat, gradient one in 2016 wasn't just aesthetic—it signaled that the app was now about discovery, commerce, and partnerships, not just photo sharing. Apps rarely do one thing anymore, and their users have accordingly more sophisticated needs.
"Today, design is more complicated, often encompassing larger teams working together on multiple, interconnected systems," says Figma Product Manager Jacob Miller. "This requires better solutions to organize and make sense of it all."
Borrowing Structure from Code
One answer teams are gravitating toward is structure borrowed from software development. Branching and merging—long a standard workflow in engineering—has become a popular pattern in design systems. Contributors can work on isolated branches, submit fixes, or propose new components without destabilizing the primary system. Managers can review changes and adopt them at their own pace, effectively letting teams build with their communities rather than merely for them.
Spotify's Encore system takes this further with local systems, where sub-teams are encouraged to fork the core system and build on top of it. Their ads team, for example, developed a range of video players through this approach, which in turn informed how the company handles video across all surfaces—expanding the reach of the design systems team itself. Some organizations open their systems entirely to the public, allowing anyone to contribute or recreate designs in full.
"Open-source systems can be great for gaining a feedback loop with a wider range of users," says Louise Macfadyen, a former design advocate on Google's Material Design team. "It also demystifies the products themselves, raising the organization's profile, while also sharing that knowledge which creates greater trust in the industry as a whole."
When Structure Becomes a Straitjacket
Too much structure, however, creates its own problems. A restrictive system chokes off freedom and experimentation. "At the end of the day, our design systems should lower the barrier for creative expression," says Miller. "Not make it harder."
Shopify Designer José Torre proposes a mindset shift: "Rather than approaching design systems like architecture, we should approach it like gardening." Architecture implies permanence—the blueprint is drawn, the building is finished. A gardener, by contrast, "plants seeds and tends to them. They know what they planted, but they don't know exactly what the garden is going to look like. They find out as it grows." Components may spread unexpectedly, or fail to thrive, and the manager intervenes as needed rather than locking in a plan.
Another alternative to rigid, atomic models is the rise of sub-components or nested components—modular units that allow customization to fit particular needs. Product design manager Matthew Ström of Stripe sees them as part of a shift toward a "functional model" of design systems, where components map more directly to actions like click, type, input, search. "Sub-components allow for scalability and a new level of efficiency beyond the object-oriented design systems we've been building over the past decade," he says.
Their advantage is flexibility in the face of an unknown future. "You don't need to anticipate the super-mega-menu you'll need three years from now," Ström explains. "You have all the sub-components, and when that opportunity comes up, you put them together according to the rules of the system. And it just works."
There's also a data dividend. More components mean more variables and richer metadata about how those components are actually being used. That information helps system managers measure what works and what doesn't—and could even feed future applications, such as an AI that generates new components from existing patterns.
Finding the Middle Ground
Deciding whether to tighten or loosen a design system isn't straightforward. Constraints can unlock creativity, and well-defined standards improve efficiency. If every aspect is perpetually up for debate, designers can't trust the system; if nothing can change, they can't work within it. The challenge is optimizing for both stability and evolution simultaneously.
Regular reviews and evaluations help ensure systems keep pace with the products they support, even when the future is hard to predict. Like any good product, a design system should solve real problems and work toward a clear goal. Building a culture where participants can contribute ideas, feedback, and components allows a design systems team to improve and extend its reach. Miller suggests teams treat their systems like products—checking in with users and responding to their needs. Ultimately, the answer lies not in the tools or tokens, but with the people using them.
