Yellow Drift — Findings
The internet does not see
the same yellow.
A word enters the internet.
A culture comes back.
32 languages · 36,451 images · Measured against #FFFF00
Core observation
Yellow is not one image.
It is a field of competing associations.
The drift is the message.
Languages studied
32
Images analysed
36,451
Maximum drift observed
ΔE 41.3
01
Largest drift
Māori drifts furthest

The word kōwhai pulls strongly toward the warm ochre of the native kōwhai flower. The resulting average is richer, more organic, and further from pure spectral yellow than any other language.

Māori
ΔE 41.3
Bengali
ΔE 30.6
Persian
ΔE 25.2
Dutch
ΔE 9.1
Hebrew
ΔE 3.7
02
Closest to reference
Hebrew stays closest

צהוב (tsahov) produces imagery that remains remarkably close to #FFFF00 (ΔE 3.7). The results are dominated by flat, graphic yellows: signs, products, and digital interfaces. A culturally compressed, almost “pure” yellow.

03
Gold as attractor
Gold pulls many languages

In more than a third of the languages, yellow drifts toward the gold-amber spectrum. Arabic, Hindi, Thai and several others show this strong cultural overlap between yellow and gold — currency, ceremony, divinity.

#fff176
#f5df79
#efbd16
Gold
#c8960c
04
Warm bias
Some languages lean warm

Portuguese, Swahili, Tagalog and others show a clear shift toward skin tones and tropical sunlight. In these datasets, contextual warmth (people, nature, environment) outweighs graphic yellow objects.

05
National symbols
Flags leave a visible mark

In Ukrainian, Swedish, Spanish, German and Lithuanian results, national flag yellow is a strong signal. Yellow rarely appears alone, it is anchored by its national colours.

Central thesis
Search engines do not return the same colour.
They return culturally compressed visual associations.
The word does not move. The image field does.
Yellow Drift · 2026
Explore all data
See the full grid
Open Grid →
Behind the project
Read the method
Open Method →
For the press
Download materials
Open Press Kit →