The Design System That Outlasted Me

In 2020, I built a design system at Wantable. Not because someone asked me to—because we were drowning. Every component was a snowflake. Every page rebuild started from scratch. The marketing team couldn’t ship a landing page without filing a Jira ticket.

So I built a system. Not a component library—a system. Shared primitives, consistent spacing tokens, composable layouts, and a Storybook instance that became the source of truth for both design and engineering.

The key decision was what not to include. Every design system I’d seen fail was too ambitious. They tried to cover every use case, every variant, every possible future need. Mine covered the 20 things we actually used, and it covered them well.

Buttons, form fields, cards, modals, typography, spacing, color. That’s it. No fancy animation system. No theme engine. Just the bones that every page needs, built to be composed rather than configured.

The real test came after I left. Four years later, it’s still running. Not because I wrote perfect code—I definitely didn’t. But because the system was simple enough that other engineers could understand it, extend it, and make it their own.

The best infrastructure disappears. You stop thinking about it. You just build on top of it. That’s the goal.

I still think about that system when I’m tempted to over-engineer something. Would this survive me leaving? Would someone else be able to pick it up and run? If the answer is no, I’m building it wrong.