Six months after a design system launches, it starts rotting. Twelve months in, half the new components being built are one-off variants that do not conform to the system. Eighteen months in, the system is legacy and the team is quietly talking about a rebuild.
This is not a failure of the original system. It is the natural state of any living artefact that is not tended to. Design system debt is real, it accumulates daily, and it compounds.
The antidote is not more documentation. The antidote is a steward — a named person whose job includes reviewing every component that ships and deciding whether it belongs in the system or whether the system needs to evolve to absorb it.
At every studio I have worked at, the moment a design system got a steward was the moment it became sustainable. Before that, it was a Figma file nobody owned. After that, it was a living product with a roadmap.
If you cannot staff a full-time steward, a rotating twenty-percent-of-one-designer role is the minimum viable version. Less than that and the rot sets in.