Back to Blog

Colour psychology is the most infographic-poisoned corner of marketing. Every chart that promises "red means urgency, blue means trust, green means growth" is downstream of a handful of claims that almost nobody has traced to a source. We traced them. This is the third post in our evidence series, after the marketing myths audit and the behavioural science replication check, and colour fared the worst of the three.

Same rules as always: every claim below is checked against the primary source, and the ones that collapse get named, not quietly dropped.

The famous stats have no study behind them

"Colour increases brand recognition by up to 80%"

This is the most-quoted number in colour marketing, usually credited to a University of Loyola, Maryland study. There is no such study. Fact-checkers who traced the claim found its earliest locatable source is an early-2000s Xerox marketing leaflet, sales collateral from a printer company, not a peer-reviewed paper.

The academic it gets pinned on is real. Dr. Ellen Hoadley of Loyola (Maryland) published a genuine 1990 experiment in Communications of the ACM. It measured whether colour helps people extract information from tables and charts faster than monochrome. It contains no brand-recognition measure and no 80% figure. A real paper about reading pie charts got repackaged into a fake statistic about logos, and two decades of marketing decks have been citing the repackaging.

"People judge products in 90 seconds, and 62% to 90% of it is colour"

This one has a real-looking citation: Singh, 2006, in the journal Management Decision. The paper exists. But it is a short review essay that opens with the 90-second claim without any reference. The figure's actual origin is a colour consultancy's website, sourced only to "CCICOLOR - Institute for Color Research." Asked for the study, the consultancy behind the page said it appears to be unpublished work from before 1990 that they could not locate. No study, no method, no sample has ever surfaced.

For scale: the fact-checkers who traced this claim point to eye-tracking work showing average total fixation on a package runs under one second. The "90 seconds of colour-driven judgment" picture does not survive contact with how people actually look at products.

The lab effects that collapsed

Red screens for detail, blue screens for creativity

A 2009 paper in Science (Mehta and Zhu) reported that red computer backgrounds improved detail-oriented task performance while blue backgrounds boosted creative tasks. It became a staple of design advice for a decade. The original studies were small: the key anagram experiment had 69 participants, and all six studies together totalled about 666 people.

Then it was retested. A direct replication of its key anagram experiment, with the same materials and nearly four times the sample, found no effect.

A 2016 attempt (Xia and colleagues, Frontiers in Psychology) found mixed results that depended on task difficulty and did not match the tidy red-detail, blue-creative split. And a 2020 meta-analysis of 67 effect sizes (Gnambs, Psychonomic Bulletin and Review) found that after correcting for publication bias there was no evidential value for an effect of red on intellectual performance at all. Screen colour is not a productivity lever, and any deck that says otherwise is citing a finding that failed its retest.

The other red effects fared no better

The 2008 "romantic red" study, which reported that men rated a woman as more attractive against a red background, shrank badly under retesting: pre-registered replications using the original materials put the effect near zero, and a separate three-study test totalling 830 men failed to find it. A broader meta-analysis still shows a small effect for men, and the original author defends it in print, but nothing approaches the original sizes. The 2007 finding that briefly seeing red before a test lowers scores fared worse: multiple close replications failed, and a 2016 study using real university exams printed on red versus green booklets found no difference in scores. Neither effect should be load-bearing in anyone's brand deck.

What actually held up

Colours carry personality associations

The strongest surviving finding is less exciting and more useful than the myths. Labrecque and Milne, in the Journal of the Academy of Marketing Science (2012), found that colours map onto brand-personality dimensions: red reads as exciting, blue reads as competent, reliable, and intelligent. Shade matters too, saturation and lightness shift the signal, so "blue" is not one decision. And the mappings replicated only partially in later work, so treat them as tendencies, not rules.

The companion finding is the one that should actually drive logo decisions. Bottomley and Doyle (Marketing Theory, 2006) found there is no single best logo colour: a colour that fits the brand's positioning is judged more appropriate than any universally popular one. An energy drink and an accounting firm should not fight over the same hex code. Fit beats fashion.

Blue tops the surveys, with an honest asterisk

A 2015 YouGov survey across 10 countries on four continents found blue was the most-picked favourite colour in every one, finishing 8 to 18 points ahead of the runner-up. It won with men and with women. The runner-up was the part that changed by country: green in some markets, red or purple in others.

The asterisk: this holds for surveyed, mostly industrialized markets, not for humanity. A 2013 paper in the Journal of Experimental Psychology: General is titled, in full, "Color preferences are not universal": Himba adults in rural Namibia showed none of the so-called universal patterns. A 2014 study found Polish participants picked blue most often while Yali participants in Papua favoured red and yellow. Blue is a safe default in the markets most brands sell into. It is not a law of nature.

Pink for girls, blue for boys: the history held, the hardwiring did not

The paper usually cited for "girls are wired for pink" (Hurlbert and Ling, Current Biology, 2007) is a three-page correspondence piece, not a full research article, and what it actually found was preference peaking in the blue region for both sexes; adult women's preferences were shifted slightly toward the red end of the spectrum, which is a nudge, not a pink instinct. The same lab then tested 4-to-5-month-old infants and found no evidence of a sex difference. Developmental studies place the emergence of girls' pink preference around age two to three, right when children start absorbing gender cues, which points at culture, not biology.

The historical record agrees. Pink-for-girls is young and commercially manufactured: American babies wore white for generations, a 1918 trade journal recommended pink for boys, a 1927 Time magazine survey of department stores found no consensus at all, and an archive search of nearly twelve million digitized newspaper pages found the two conventions close to an even split before about 1920. By historian Jo Paoletti's account, today's rule only became dominant around the 1950s and hardened further in the 1980s. If your category colour-codes by gender, you are following a merchandising convention from living memory, not human nature. Conventions can be worth following, they move product, but a convention is a choice, and choices can be rethought.

How to actually choose brand colours

Choose for fit, not folklore. The supported question is not "what does red do to the brain" but "does this colour match what this brand claims to be." A disruptive brand in a category of blue incumbents has a real decision to make, and the research says fit and context carry the weight.

Buttons are about contrast, not hue. No peer-reviewed work crowns a highest-converting button colour. What a button needs is to be unmissable against the page around it. Test contrast and placement before repainting red to orange.

Own something distinctive, apply it consistently. The defensible version of the claim is unglamorous: what the research verifies is that colour used consistently and congruently shifts how people read a brand's personality and can influence purchase intent (Labrecque and Milne, 2012). No study we could find quantifies a recognition gain at all. Anyone who quotes one at you is reading from the Xerox leaflet.

Frequently Asked Questions

Does colour really increase brand recognition by 80%?

No study supports that number. Fact-checkers traced it to an early-2000s Xerox marketing leaflet, and the Loyola academic it is usually credited to published research about reading charts and tables, not brand recognition. What research does support is that colour shifts how people read a brand's personality and can influence purchase intent; no study credibly quantifies a recognition gain.

What is the best colour for a logo?

The research says there is no single best colour. Bottomley and Doyle (2006) found that a colour judged to fit the brand's positioning beats any universally popular colour, and Labrecque and Milne (2012) found hues carry personality tendencies, red toward excitement, blue toward competence. Choose the colour that matches what the brand claims to be, then apply it consistently.

Do red or blue backgrounds change how people think?

The famous 2009 Science study claiming red aids detail work and blue aids creativity did not survive retesting. A direct replication of its key anagram experiment, using the same materials and nearly four times the sample, found no effect, and a 2020 meta-analysis of 67 effect sizes found no evidential value for an effect of red on intellectual performance after correcting for publication bias. Treat screen-colour productivity advice as an artifact of small early studies.

Is blue really the world's favourite colour?

In surveyed markets, yes. A 2015 YouGov survey found blue finished first in all 10 countries polled, 8 to 18 points ahead of the runner-up, and it won with both men and women. But cross-cultural research with the Himba in Namibia and the Yali in Papua found very different preference patterns, so blue is a strong default for most markets rather than a universal law.

Should my call-to-action button be a specific colour?

No hue is proven to convert best. What matters is contrast with the surrounding page, size, placement, and the words on the button. Pick a colour that stands out sharply from your palette, keep it consistent for the primary action, and spend your testing budget on the offer and the copy before repainting the button.

Related Reading

• Five marketing myths peer-reviewed research says are wrong, the first post in this series.
• Behavioural science and the replication crisis, which effects survived retesting and which quietly died.
• Building a brand that stands out, where colour actually earns its keep: distinctiveness and consistency.