March 15, 2024
How to build a design system the product team actually uses
By Vitor Botelho, Product Designer
Most design systems die from neglect, not from missing components. The problem is not technical. It is adoption.
The problem nobody names
Everyone wants a design system. Few teams keep one alive for more than six months.
The classic symptom: the component library grows, gets too heavy to use day to day, and the designers start working around it. Twelve months later you have two products — the official one, and the one the team actually ships.
That is not an execution failure. It is a failure of premise.
The wrong premise
The wrong premise is: "we will build every component, and adoption will follow."
It does not work that way. Adoption is not a consequence of completeness. It is a consequence of perceived usefulness. If the design system does not solve the real problems the product team runs into today, it gets ignored. Politely, but completely.
What works
Build from the outside in. Start with the concrete problems the team is hitting now, not with the components that "should exist" in theory.
In practice that means:
1. Audit before you create
Before you create a single new component, document what already exists. You will find four versions of the same button and three variants of the same card. That inventory is your real backlog.
2. High-return components first
Which components are used on more than five screens? Start there. One well-made button has more impact than twenty edge-case components.
3. Documentation as the product
Documentation is not an accessory. It is the product. A component without clear documentation is a component that does not exist for most of the team.
4. Short feedback loop
Weekly review with the people using the system. What is blocking them? What is wrong? What is missing? Without that loop, you are building into the void.
The metric that matters
Not the number of components. Not the coverage of the library.
It is this: when a designer has to build a new screen, what percentage of the work does the design system already solve?
If the answer is below sixty percent, the system is not mature enough to return on the investment yet.
That number grows over time, if you iterate with discipline.