How Quokka Cut Time-to-Prototype by 60% with a Radical Component Library Prune

Design · 5 min read

At 18 people and growing fast, Quokka's product team was drowning in options: dozens of variants of buttons, input fields, and complex containers that designers and engineers resisted using consistently. The design lead proposed a controversial experiment — prune the component library to a “core eight” set of components and enforce explicit extension rules for new additions. The hypothesis was that fewer choices early on would speed decisions and produce more consistent implementations.

The pruning process started with analytics-driven audits: identify the least-used variants, components that required frequent overrides, and those that repeatedly caused accessibility or parity bugs. Over six weeks the team archived 42% of the library, standardized naming, and published a small cookbook for composition patterns. Importantly, archived components weren’t deleted — they were hidden with migration notes to reduce friction.

Results were measurable: time-to-prototype dropped by 60% for new flows, engineering work for UI parity decreased 28%, and new designer onboarding went from eight days of hand-holding to three. The team also adopted a lightweight governance cadence: monthly “add or adapt” reviews to avoid creeping bloat while keeping the system responsive to real product needs.