Why Everything We Knew About Palettes Was Wrong
If you've ever spent hours clicking around a color picker trying to build the perfect palette, you already know the truth: color is incredibly complicated. It sits at the messy intersection of psychology, science, culture, and math.
For a long time, digital designers operated under a seemingly logical assumption. If we wanted a lighter blue, we just added white. If we wanted a darker blue, we added black. But as we started building massive, scalable design systems, this linear approach completely fell apart.
Here is the story of why our digital color systems were broken, and how science is finally helping us fix them.
The Trap of Linear Color Generation
When creating a color palette, the traditional method was entirely manual: a designer would pick a core brand color, and then eyeball the lighter and darker shades until it looked right. But as design systems grew, this approach wasn't scalable. How do you programmatically tell a computer to generate ten beautiful, usable shades of blue?
The obvious answer seemed to be math. You take a core color and step it evenly toward white or black in equal mathematical increments.
But there’s a massive problem with this: humans don't see in straight lines. As Linda Dong and the design team at Lyft discovered when building their color system, stepping through a color from light to dark in equal mathematical steps does not look equal to the human eye. The darkest colors in a linear scale will look almost identical to each other, while the lighter colors will look drastically different.