Case study · symplr

Building & Scaling the Alloy Design System

Building a design system from zero, then proving it could survive the business doubling underneath it.

Role

Senior Director of UX, Program Owner

Timeline

2020 – present

Team

Design agency, 4–5 engineers, 2 designers, a part-time product owner, Big Medium (2022)

Tools

Figma, Storybook, Zeroheight, Confluence

The challenge

When I joined symplr in January 2020, the company’s 26 products, assembled through a decade of acquisitions, had no shared brand, no shared codebase, and not one line of shared component code. Every product ran on a different front-end stack, built by a different original company, in a different visual language. There was no design-system team and no budget for one. My first priority was to change that.

My role

I owned this initiative end to end, from making the case to the executive team to running the program day to day. I secured the vision and buy-in with no dedicated budget, hired the agency that defined Alloy’s look and feel, built the case for (and then hired) a four-person engineering team, and two years later brought in outside experts to rebuild our governance model when the program hit its first real stress test.

Phase 1: Build (2020–2021)

No budget, no staff, and 26 products that looked like 26 different companies.

• Audit & plan. I audited the components already in use across our go-forward products, aligned with Engineering on a technical vision, and took it to the executive team to secure the mandate and a phased budget.
• Design & build. I hired an agency to define the visual language and secured funding for a four-person engineering team. We built the first 33 components, launched a public style guide on Zeroheight, stood up Storybook so engineers could preview components live, and named the system Alloy.
• Release & adopt. The first release shipped in July 2021. I hired two full-time designers, brought on a part-time product owner, published an ideas-to-features intake process, and tested developer-led, UX-led and product-led adoption strategies to see which actually moved products onto Alloy.
• Parity & expansion. We extended into accessibility and a mobile theme, reached operational parity with the legacy component libraries, and defined formal adoption levels so every business unit knew what compliance looked like.

After two years: 83 components shipped, 11 of 26 products actively adopting, and every new product built in Alloy from the ground up.

Phase 2: Survive real pressure (2022)

Early in 2022 the business decided to move most customers off legacy on-prem products. That meant doubling the engineering organization to roughly 600 developers, with no increase to the six-person Alloy team. It was the real test of whether the system could hold.

I was worried that low-quality or incomplete components would cost us the trust we’d spent two years building with engineering. So I brought in Brad Frost and Josh Clark of Big Medium for a six-week engagement to rebuild our governance model. Together we defined what Alloy would and wouldn’t build, and why, and published those values so product teams could set realistic expectations.

I also created an adoption template mapping each adoption level to the components and tasks a product needed, so designers and product managers could plan their own adoption without routing every decision through me. And with our engineering operations team I designed a quarterly adoption dashboard that made progress visible enough to report to the executive leadership team.

Phase 3: Extend to new frontiers (2023 to present)

Outcome

• 500+ developers worldwide build with Alloy’s components.
• No additional design-system headcount was needed, even as engineering doubled.
• 38 designers used the Figma component library in a single year.
• 100 components live, built and maintained by a team of four engineers.
• One adopting team estimated Alloy saved roughly three months of velocity on a new product; another estimated 30–40% savings in conversion time.
• The team now owns its own process and improvement work, such as revamping color and font tokens, without my day-to-day direction.

What others said