Open any design system that has been alive for more than two years and look at the button. Not the documentation page, the actual component file. Count the props.
Somewhere in there is a variant called "secondary-alt". Next to it, "secondary-alt-2". Someone added "ghost-danger-small" for one modal in a settings page that nobody has opened since. Every one of these made sense on the day it was added. A designer had a deadline, an engineer had a ticket, and saying yes took five minutes while saying no would have taken a meeting.
This is how design systems actually die. Not with a big rewrite or a rebrand, but with a slow pile of exceptions that each look harmless on their own. The system still exists. It just no longer says anything, because a component that can look like everything has stopped being a decision.
The most useful thing a design systems team can do is also the least popular one: say no, and write down why. A new variant should have to earn its place. Who needs it, on which screens, and what happens if they use the closest existing one instead. Most of the time the honest answer is "it would look slightly different and nobody would notice".
Deleting a variant feels like taking something away from people. In reality it gives them back the thing they came to a design system for in the first place, which is not having to think about buttons.
Get an email when Aino publishes
More from
Aino Korhonen →