Design is a Function
An essay by Austin Knight
Summary
Design isn't a service or a subset of product. It's a peer function, just like product and engineering, with its own incentives, responsibilities, and authority. When it’s treated as a subordinate, product quality deteriorates. But when it’s empowered as a function, design protects what matters most: the end-to-end experience. Functional organizations like Apple and Airbnb prove the model works. Square is proving it again.
Apple’s organizational structure is a masterclass in functional design. Its leaders aren’t grouped by product or business unit - they’re grouped by discipline. In one of my favorite HBR articles, Apple executives describe how the functional model gives design room to thrive.
There is no General Manager of the iPhone. Instead, design reports directly into design. Engineering into engineering. Marketing into marketing. Each function has autonomy, depth, and long-term investment in excellence. The result is a consistent, coherent experience across every part of the ecosystem, from hardware to software to service.
This structure doesn’t just protect quality. It fuels innovation. Functional organizations don’t split design across teams. They concentrate it. And that focus allows design to lead.
Airbnb has followed a similar path, setting up design and engineering to be deeply partnered. They operate from shared values, not from handoffs. That’s the model. Brian Chesky summarized this perfectly in his Config Q&A (which is one of the best I've seen):
Design should not be a service organization. It’s not your job to catch things or stop them before they go out. It’s your job to work from the very beginning.
— Brian Chesky, co-founder of Airbnb
Square’s Return to Functionalization
Square originally operated as a functional organization. Over time, it adopted more of a GM-style structure, grouping teams by business units with cross-functional leadership. That model helped it scale, but it came with a cost: design quality began to suffer. Functional accountability blurred. Priorities became more about individual roadmaps than holistic user experiences.
In 2023, Square pivoted. It was one of the boldest structural moves I’ve seen. The company laid off thousands of people, including GMs and other executives. It leaned out. It moved to a performance culture. And it returned to a functional model.
That change required extreme conviction from Jack Dorsey. But it’s working. Design now has more influence at Square than at any other point in my tenure. Product quality is improving because design isn’t just invited to the table—it’s helping set it.
We often think of pivots as product shifts. But functionalization is a pivot too. One that affects every decision a company makes from that point forward.
Design Balances the System
In great software and hardware organizations, product, engineering, and design are the three legs of the stool. Remove one, or collapse one into another, and the system falls over.
Design, when functioning independently, plays a distinct role. It sets the vision. Product works out the practical details. Engineering figures out how to build it. That's a gross oversimplification, but you get the point: this is not hierarchy, it's harmony.
Each function is incentivized differently. Product often optimizes for roadmap velocity and short-term metrics. Engineering is rewarded for shipping and closing tickets. That’s fine, but it’s not enough. Over time, this approach leads to fragmentation. Every shipped feature is a win, but the overall experience deteriorates.
Design is built to prioritize experience quality. It filters, simplifies, and says no. But it can only do that when it’s structurally empowered to push back. If design is subordinate to the very functions it’s meant to balance, it can’t lead. It can only serve.
Engineering Can Unlock Excellence
Good organizations don't just leave quality to design though. The best engineering cultures assert that while shipping features and closing tickets is critical, quality isn’t optional. At Square, one of the core cultural principles is engineering excellence. It means engineering is responsible not just for velocity and raw output, but for the quality of the experience they ship. That quality is measured both quantitatively and qualitatively. It’s incentivized internally (to the point that it's an explicit part of the engineering career ladder). Engineers can flag critical UX issues for funding. There are mechanisms to invest in quality, even when it’s expensive.
This matters because in most large companies, engineering is rewarded for speed and output. Velocity gets conflated with impact. But closing tickets is not the same as building something great.
The only way to uphold quality at scale is for engineering and design to hold each other accountable. That accountability only works when the two functions are peers. You can’t enforce quality through process. You enforce it through principled culture, and whether we want to admit it or not, corporate structures are explicit statements of culture and priority.
QA Is Not a Safety Net
In too many teams, engineers write code with limited exposure to the design spec, and design only finds out when QA begins. By that point, timelines are tight and mistakes are baked in. QA becomes a fire drill.
Designers catch what they can, but they’re reacting to code that’s already been written. It’s not that design wasn’t present from the beginning. It’s that the process allowed engineering to run ahead without adequate alignment, and product didn’t account for QA time in the roadmap (it's the first thing to get cut when the deadline is approaching fast).
This is not a process problem. It’s a culture problem. One that starts with engineering leadership and product planning, and ultimately traces back to the CEO. Design can’t fix it alone, but it can help enforce the standard. And that’s only possible in a functional structure where design has lateral influence.
Prototypes Are Power
Here’s the truth: design can bulldoze the other functions - especially today. With tools like Cursor, Figma, V0, Lovable, and Goose (an open-source AI framework from Square) designers can prototype entire products on their own. They don’t need engineering buy-in to bring a concept to life. And they don’t need product sign-off to shape the roadmap.
That’s a lot of power. But power without restraint can become destructive. I’ve seen design organizations push ideas through brute force - successfully. And I’ve seen the bridges that get burned along the way. It’s usually not worth it, and almost always not the optimal route. There are better paths to the same outcome.
Good design leaders know when to push and when to pull others in. They lead with prototypes, not just words. They use their ability to make, not just to argue. And they understand that influence is earned by results, not by rank.
The Design Leader Who Still Designs
Functional models reward craft. They elevate leaders who can do the work, not just talk about it. Leaders that don’t just describe the vision, but can build it. AI makes that even easier. The ability to prototype entire workflows, UIs, and flows without writing a line of code gives modern designers more leverage than ever.
Historically, the best designers weren’t always the best leaders. That’s still true in some cases; soft skills matter. But the gap is closing. And today, the most valuable design leader is someone who combines taste, vision, and the ability to execute.
This is the same reason engineering has long commanded respect. Engineers can argue for a direction and then write the code. Designers can now do the same. The medium is accessible, the tools are powerful, and the expectations are rising.
Bold Leaders Make Space for Bold Design
Design needs strong leadership to thrive. Not necessarily loud, but principled. Opinionated. Willing to say no. In my experience, the most effective leaders are often the most polarizing. Not because they seek conflict, but because they have a point of view.
Only ship something that you’re proud of. Don’t test something until you’re happy with it. If you don’t want to put your name on it, don’t ship it.
— Brian Chesky, co-founder of Airbnb
That level of ownership is what separates good from great. And that kind of clarity is only possible when design has room to think - not just react.
Taste Is Not as Subjective as You Think
In my early essays like Design Is Not Art and Good Design Is Humble, I made the case that design should be objective, rigorous, and grounded in real-world constraints. I still believe that.
But I’ve also come to appreciate nuance. Taste matters. A principled, opinionated designer can often spot a problem or solution long before the data shows it. Not everything should be measured. Some of the best design decisions are ones that resonate - not because they were A/B tested into oblivion, but because someone with taste knew it felt right.
A/B testing is abdicating responsibility to the user.
— Brian Chesky, co-founder of Airbnb
Functionalization Is a Pivot
I recently spoke with a scale-up doing around $30M in ARR, aiming to more than triple that in the next 18 months. They were candid about their challenges: their primary competitor was winning on UX. That competitor offered a cohesive, self-serve experience. Their own product, by contrast, felt disjointed and required heavy onboarding support from account managers.
They expressed a strong desire for design leadership. Not just visual polish, but deep UX thinking: stronger information architecture, a coherent design system, and a long-term vision. They said they wanted a “change agent.”
But when we discussed the structure of the role, design would report into product, while both product and engineering reported directly to the CEO. In other words, design would be the only one of the three core functions not reporting to the ultimate decision maker. That matters. Because if you're asking design to drive transformation, but only within the boundaries of an existing structure, you're limiting its influence before it begins.
Frankly, the likelihood of that design leader being successful is low. They'll either end up hiring someone that doesn't spot this issue and thus isn't up to the challenge, or they'll squander good talent inside a heavy corporate structure and wonder why it didn't work out.
This is a common blind spot. Companies recognize the need for stronger design. They say they want it to lead. But changing the reporting structure often means someone else has to let go - of control, of responsibility, or simply of what they’re used to. That’s the hard part. Not hiring the designer. Redefining the role.
Design is not a service. It’s not a subset of product. It’s a function. And when companies treat it that way, both structurally and culturally, they create the conditions for quality to thrive.
