Like most companies, Affirm had issues with building fast and siloed in different teams which resulted in a lot of custom, non-reusable code throughout the years.

In 2019, Affirm built a design system. The initial ask was was to create a Figma library to support the existing V1 code. However, both Figma and code for the V1 design system were built with poor properties, inconsistent, mismatched, Frankensteined, neglected, not supporting mobile operating systems, and outdated.

Unfortunately, there were major blockers to the path to a successful design system.

First, the V1 components we were tasked with supporting were, frankly, not usable at all. Next, the project lacked resources, since leadership wanted to preserve engineering time. Finally, any fixes to the V1 code or to create new components would result in breaking changes which were not allowed in order to preserve engineering time.

Common in several teams trying to build design systems, these issues created an uphill battle to get engineering support to build the Design system and support for adoption.

In an attempt to get buy-in from leadership, we built 2 Figma versions. A V1 component-build that matched the V1 build exactly and an industry-compliant V2 Northstar.

Each V2 Northstar component was audited internally, compared externally, validated internally with designer feedback, props built in a JSON ready matrix, systematic, pixel-perfect and documented.

As pushback for resources continued, we decided to provide data. We built a case that Northstar V2 really is the only way forward. We created user tests comparing V1 to V2 components and distributed them internally to designers in the team.

The results were clear that everyone felt V1 was a regression, and everyone felt overwhelmingly positive toward V2. We presented this to leadership and highlighted how broken V1 properties were and we finally got approval for more resources to help prioritize V2 components.

We were then able to continue designing the system in a better than yesterday approach.