Three lessons from twenty years of design systems

I have built design systems for a group with hundreds of stakeholders and for a start-up where three people decide everything. The tools differ, the problems barely do.

1. Systems fail because of people, not components

The best component library is useless if nobody understands why it exists. A design system is first an agreement and then a library. Skip the agreement and you are building decoration.

2. Order has to make you faster, not tamer

The most common objection to systems: "It limits us." The objection is valid — for bad systems. A good system takes the boring decisions off your plate so there is more time for the interesting ones. If your team has been slower since the system landed, the system is built wrong.

3. Start small, grow in public

The systems that survived started with a handful of rules that visibly paid off. The ones that failed launched as flagship projects with their own name, logo and roadmap — finished before anyone needed them.