NerdWallet did not have a missing-components problem. It had an adoption problem. The system existed, but ten-plus product teams built around it because finding the right component took longer than rebuilding it. Currency reframed the system as an internal product, named, searchable, documented, and tooled, so the official path was finally faster than the workaround.

Teams were being rational, not non-compliant. The system had no clear identity, the documentation did not match how designers and engineers searched, and recreating a pattern locally was often faster than reusing it. Those small local decisions compounded into design debt: parallel patterns, inconsistent implementation, and a shared system that was available but rarely used.

Naming was adoption infrastructure, not decoration. “The component library” described a folder; Currency gave teams a named product to reference in meetings, link to in docs, and hold accountable over time. The word fit the company too, implying value, exchange, and shared acceptance, which is exactly what a design system needs to work. Paired with nerdwallet.style, it read as maintained infrastructure with an owner, not a static archive.

Teams came to Currency to find the right component variation and know when to use it. The new documentation moved that decision to the top of every page, so designers could compare options without reading a wall of guidance while engineers found implementation details below. Three reusable templates, Foundation, Component, and Pattern, kept the system growing without every new addition becoming its own project.

Documentation solved the information problem; tooling solved the behavior problem. A Chrome extension brought the NerdWallet grid onto live pages so teams could check alignment without opening a separate spec, and the Figma library made the approved path faster than rebuilding from scratch. Adoption rarely comes from discipline. It comes when the right choice costs less effort than the wrong one.

A small systems team could not support every product vertical through one-off decisions, so the contribution model had to remove ambiguity at the source. Teams needed a clear answer to one question: where does this component belong? Broadly reusable components moved into Currency, product-specific ones stayed local, and unclear cases went to office hours. That gave teams a visible path in without routing every decision through the core team.

The launch changed the system’s relationship with the organization. Weekly documentation visits tripled within the first month, 80% of employees rated the refreshed guide useful, and the new templates cut variant-discovery time by more than half. Those numbers measure behavior, not intent: teams returned to a resource they had abandoned and finished the most common task faster. Currency became the standard because it lowered the cost of the right decision, not because anyone was told to use it.

The refresh solved adoption, but it exposed the next constraint: contribution throughput. The decision tree clarified where a component belonged; it did not solve who had the capacity to review, refine, and ship every contribution that came in. The next step would be team-level system advocates, accountable owners who keep contribution quality close to the work. The front door was designed well. The review pipeline behind it needed the same rigor, and it did not get it before I left.