Addaly is in open beta. Things will change, and AI answers can be wrong — check anything that matters.

Design, Before Any Software

Hierarchy, type, colour and spacing — the part of design you can do with a pencil, before any software.

Lesson 39 of 749 min

What colour blindness actually removes

Not black and white

The common mental picture — that a colour-blind person sees in greyscale — describes achromatopsia, which is extremely rare. What almost everybody means by colour blindness is a shift or a loss in one of the three cone types, and it leaves most colour vision intact while collapsing specific distinctions.

The categories:

  • Deuteranomaly — reduced green sensitivity. By far the commonest.
  • Deuteranopia — green cones absent.
  • Protanomaly / protanopia — reduced or absent red sensitivity. Also darkens reds noticeably.
  • Tritanomaly / tritanopia — blue-yellow, rare, and usually acquired rather than inherited.
  • Achromatopsia — no colour discrimination at all. Very rare.

The first four are the red-green family, and they account for the overwhelming majority.

The prevalence figures, honestly

The number usually quoted is about 8% of men and 0.5% of women, or roughly one man in twelve. That figure comes from studies of populations of northern European descent, and it is not universal: measured rates are lower among men of African and Asian descent, with published figures commonly in the 3-5% range depending on the population and the screening method.

The inheritance explains the sex difference. The genes for the red and green photopigments sit on the X chromosome, so a man with one affected copy is affected, while a woman generally needs two.

What this means practically: for a product with a hundred thousand users, several thousand of them are affected, in every market. The exact figure varies; the design consequence does not.

What converges

The distinctions that collapse are the ones between colours that differ mainly in the red-green direction and match in lightness:

  • A red and a green of similar lightness — the classic status dot pair.
  • Red text on a green background, or the reverse.
  • Orange against yellow-green.
  • Purple against blue, in some types.
  • Pink against grey.

The distinction that survives everything is lightness. This is the payoff for building palettes by moving lightness rather than hue: a palette built that way is already largely robust, without anybody having thought about colour vision at all.

Checking it, free

Simulation is built into the tools you already have:

  • Chrome DevTools: the Rendering panel has Emulate vision deficiencies, offering protanopia, deuteranopia, tritanopia and achromatopsia.
  • Firefox DevTools: the Accessibility inspector includes the same simulations.
  • GIMP: View > Display Filters > Colour Deficient Vision, which applies to the canvas while you work.
  • Coblis, a free browser tool, for uploading a flat image.

Run the design through deuteranopia first, since it is the commonest. Then look for anything where information has disappeared rather than merely changed colour.

The honest limitation on simulators. They model dichromacy — the complete absence of a cone type — reasonably well. They model anomalous trichromacy, which is the much more common condition, only approximately, and they represent an average rather than any individual. A simulator that shows a distinction surviving is weaker evidence than one that shows it disappearing. Treat a failure as conclusive and a pass as encouraging.

Designing so it does not matter

The systematic answer is in the next lesson: never let colour be the only carrier of meaning. Beyond that, three specific habits:

What common colour vision deficiency removesDistinctions that collapseA red and a green of similar lightnessRed text on a green groundOrange against yellow-greenPurple against blue, in some typesPink against greyDistinctions that surviveLightness, in every common typeShape, and an iconThe word itselfPosition in a list or a columnPattern or hatchingThis is the payoff for building a palette by moving lightness. A ramp made that way is already largelyrobust, before anybody has thought about colour vision at all.
What common colour vision deficiencyremovesDistinctions that collapseA red and a green of similar lightnessRed text on a green groundOrange against yellow-greenPurple against blue, in some typesPink against greyDistinctions that surviveLightness, in every common typeShape, and an iconThe word itselfPosition in a list or a columnPattern or hatchingThis is the payoff for building a palette by movinglightness. A ramp made that way is already largelyrobust, before anybody has thought about colourvision at all.
  1. Separate in lightness as well as hue. If a red and a green must sit together as categories, make one clearly darker. This alone solves most cases.
  2. Avoid the red-green pair for opposed meanings where an alternative exists. Blue and orange is a widely used substitute that survives every common deficiency, and it is why so many data visualisations use it.
  3. Use a tested categorical palette for charts. The Okabe-Ito palette was designed for exactly this and is freely usable:
#000000  #E69F00  #56B4E9  #009E73
#F0E442  #0072B2  #D55E00  #CC79A7

Eight colours that remain distinguishable under the common deficiencies. ColorBrewer, also free, offers sequential and diverging schemes with colour-blind-safe filters built in.

One thing not to do

Do not build a separate colour-blind mode as the solution. It requires the user to know they need it, find it, and turn it on, and it leaves the default design broken for everybody who does not. Fix the default. A mode may be a useful extra; it is not a substitute.

The one thing to keep

Common colour vision deficiency collapses red-green distinctions while leaving lightness discrimination intact, so a palette built on lightness steps is already largely robust — and a simulator showing a distinction disappearing is conclusive while one showing it surviving is not.

Before you move on

A chart uses six hues at matched lightness for six data series, and a deuteranopia simulation shows two of them merging. The team's proposed fix is to add a colour-blind mode toggle. What is the stronger objection?

Pick the one you would defend. Nobody sees your answer.

No ads. No data sale. No public scores on people. Ever.

© 2026 Addaly