Complementary colours sit directly opposite each other on the colour wheel. Red and green, blue and orange, yellow and purple. Put either one next to its opposite and both look more intense than they do alone.
The theory takes about a minute to learn, which is why it turns up in every design article. What takes longer is noticing the three places the pair stops working, because none of them are visible on the wheel.
Mistake one: using equal amounts
Two colours of equal strength in equal quantities compete for the same attention, and the eye gets nowhere to settle. The result reads as a warning sign rather than a brand.
The fix is a ratio somewhere near ninety to ten. One colour holds the surface, the other appears rarely and does the pointing. Orange on deep navy works because there is very little orange. The same two tones in equal halves is a road sign.
Dropping the saturation on one side does the same job by another route. Muted green against a clear red is something a person can look at. Two fully saturated tones are not.
Mistake two: using them for text
The wheel shows contrast in hue. It says nothing about contrast in lightness, and lightness is what makes text readable.
Red and green are opposites in hue while sitting at almost identical lightness. Set one on the other and the boundary shimmers, and body copy on that background cannot be read at any size. The same pairing is the single most common point at which information disappears for readers with colour vision deficiency.
So a complementary pair belongs on the thing being emphasised, never on the relationship between text and its background. That job goes to a difference in lightness, which is measurable and which the wheel does not describe.
Mistake three: assuming there is one wheel
This is the one that catches people out with a printer on the phone.
The wheel most articles show is the painter's wheel, with red, yellow and blue as primaries. Screens do not work that way. A screen mixes light, its primaries are red, green and blue, and the opposites therefore land somewhere else.
| System | The opposite of blue |
|---|---|
| Painter's wheel (RYB) | Orange |
| Screen (RGB) | Yellow |
| Print (CMYK) | Yellow |
Both answers are correct inside their own system, which is why a pair chosen on a wheel can look wrong once it is printed, and wronger still on uncoated stock that absorbs ink and dulls everything a step.
The pairs, for reference
| Colour | Complement |
|---|---|
| Red | Green |
| Blue | Orange |
| Yellow | Purple |
| Red-orange | Blue-green |
| Yellow-orange | Blue-violet |
| Yellow-green | Red-violet |
Brown is missing from that table because it is orange, darkened and dulled. Its complement is therefore a muted blue rather than some fourth thing. Beige, olive and maroon work the same way: find the colour underneath, then take its opposite.
Related schemes read off the same wheel and are usually easier to live with. Analogous takes two or three neighbours and stays calm. Split-complementary takes a colour and the two either side of its opposite, keeping the contrast and losing the tension. Triadic takes three evenly spaced colours and demands that one of them clearly leads.
What survives into a brand palette
Colour theory narrows the field. It does not make the decision, because the wheel knows nothing about the audience, the competitors, or the places the colour has to appear.
The order that works in practice: choose the one colour that carries the brand, then use the wheel to find its opposite for the things that need to stand out, then check both in the places they will actually be seen. Choosing that first colour is its own problem, covered in the guide to choosing a brand colour.
Turning the result into rules and codes that survive a printer and a new designer is branding and visual identity work, and it ends up written down rather than remembered.
