Building the Design Function at Knapsack

How I grew design from an execution role into an organizational partner

What the function was

When I joined Knapsack as a Senior Product Designer in September 2021, design existed in a familiar but limiting way: we made things look good. The function was execution-focused by default, brought in after decisions had been made to apply polish and produce deliverables. There was no expectation that design would have a perspective on what to build, when, or why. That wasn't anyone's fault. It was just how the company had grown up.

The problem with that model isn't just that designers feel undervalued. It's that the work suffers. When design enters a conversation after the shape of something has already been decided, the best it can do is optimize the surface. The harder, more valuable questions, about what a feature actually needs to do, what mental model a user is operating with, where the real friction lives, those questions get answered without design in the room.

I started working on that problem before I had a director title. By December 2022, when I was promoted to Director of Product Design, the work of changing how design was perceived was already underway. The promotion gave me more leverage, but the approach didn't change.

How I thought about the job

My core belief was that design's influence had to be earned through usefulness, not argued for through process or hierarchy. I didn't come in with a playbook. I didn't propose new rituals or lobby for design to have a seat at the table. I started by understanding what each stakeholder actually cared about and finding ways to be genuinely useful to them on their own terms.

That meant resisting the instinct to formalize too early. A lot of design leaders in this position start by establishing critique cadences, design reviews, definition-of-done checklists. Those things have value, but installed before the organization has seen why design is worth the investment, they read as overhead. I chose to earn trust first and put structure around what was working once I had it.

It also meant staying a working leader. With a team of two to three designers, I could have stepped back from hands-on design work entirely. I chose not to. Being in the work kept me credible with the team, gave me real context for the conversations I was having with stakeholders, and meant I wasn't leading from a distance on a product I didn't fully understand.

What I actually did

The most important work happened in relationships, one at a time. With the CEO, I tried to be a thinking partner on product direction, not just someone who could execute a vision. With the CTO and head of engineering, I invested in understanding the technical landscape well enough to have honest conversations about tradeoffs rather than just advocating for design preferences. With the head of product, I built a working rhythm where design and product were solving problems together rather than passing work back and forth.

One relationship that paid off in ways I hadn't fully anticipated was with the head of customer service. Most design leaders focus their stakeholder energy upward and across to product and engineering. Customer-facing teams at a B2B company see patterns in user frustration that never make it into analytics, and getting design into those conversations earlier meant we were working on real problems rather than inferred ones.

On the team side, I was deliberate about who I hired. Knapsack's product is inherently technical, a design system platform used by design and engineering teams together. I prioritized designers who could hold their own in technical conversations, who were curious about systems thinking, and who didn't need a lot of hand-holding on product context. The team stayed small, but it operated above its weight class because of that hiring bar.

The shift that happened

The signal that something had changed came gradually and then felt obvious: I started getting pulled into initial conversations across almost every workstream. Not just product and engineering, but customer service, strategy discussions, early problem framing. Across nearly every function, the default shifted from "bring design in when it's ready for polish" to "bring design in while we're still figuring out what to build."

That kind of broad organizational pull doesn't come from a single champion or a well-timed pitch. It comes from enough people having enough good experiences that it becomes the expectation. By the end of my tenure, design had relationships and visibility at every level of the organization, from the CEO down through individual contributor teams, that it simply hadn't had before.

There are no clean metrics for this kind of work. What I can say is that the change was visible and broadly felt, not the result of a single initiative but of consistent presence and earned trust over roughly three years in the director role.

What I'd do differently

The relational approach worked, but it took longer than it needed to because I was too patient about formalizing what was working. Once I had enough trust built with key stakeholders, I could have moved faster to turn those informal patterns into explicit expectations. I waited for things to feel settled before putting structure around them, and some of that structure would have helped the team, not just me.

I'd also document the design function's value more explicitly and earlier. The work I was doing in stakeholder relationships was largely invisible inside the organization, which is fine when things are going well, but it means the function's contribution can be hard to articulate when it matters. I got better at this over time, but I wish I'd started sooner.

Finally, I underestimated how much team composition shapes organizational perception. A small team means every hire is load-bearing, not just for capacity but for credibility. I was careful about this, but I'd be even more deliberate in a future role about treating each hire as a statement about what the design function is and what it can do.

Next Selected WorkFigma Integration: From Embeds to a Real Design Model   ↗