Before you build any design system components, spend a week reading your own product. Not using it — reading it, the way you'd read a magazine whose design you were asked to critique. Open every page. Take a screenshot of every button. Write down every place you see a blue that is slightly different from every other blue.
The output of this week is a single document that lists every colour, every type size, every spacing value, every shadow, every border radius, every button shape currently in production, with a screenshot of where each one appears. I called mine "the zoo" because it felt like a catalogue of wild animals. My zoo had forty-seven distinct blues. Forty-seven. At a company with one designer. The blues had arrived one at a time, each one with a reason that made sense on the day.
You cannot design a system that fixes the zoo until you know how large the zoo is. Most design-system projects fail because they start with a Figma library before the founders know how much chaos they are actually replacing. Your zoo is your scope document. It is also your evidence, when you need to explain to a skeptical founder why the design system is worth a month of your time. Nothing makes the case for consistency like forty-seven blues on one piece of paper.
Sign in to leave a comment.