The Psychology of Colour in Web Design

In web design, colour is the silent ambassador of your brand. It communicates values, evokes emotions, and—most importantly—Search for colour psychology in web design and you will find the same article three hundred times. Blue means trust. Red creates urgency. Green means growth and sustainability. Yellow makes people happy. Somewhere in there, a claim that up to ninety per cent of a first impression comes down to colour alone.

Almost none of this survives contact with the evidence, and building a brand on it is how businesses end up with a palette that looks like every competitor in their category.

This is the version we would give a client, which means separating the small amount that holds up from the large amount that does not.

The statistic everyone quotes

The claim you will see most often is that between sixty and ninety per cent of a snap judgement about a product or website is based on colour alone.

It traces back to a 2006 paper about consumer product marketing, not websites, and the figure has been detached from its original context and repeated so many times that it now reads as established fact. It is not. Nobody has isolated colour from typography, layout, imagery, photography quality and copy in a way that would let you attribute a specific percentage to any one of them, because in a real design they are inseparable.

The related claim, that people form an opinion of a website in about fifty milliseconds, is on firmer ground. That comes from research by Lindgaard and colleagues in 2006, and it has held up reasonably well. But note what it actually says: people form a rapid judgement of visual appeal. It does not say that judgement is driven by colour, and it does not say the judgement is durable or that it predicts whether they will buy anything.

We are pointing this out because the stats get used to sell palette work as a conversion strategy, and it is not one. Colour matters. It just does not matter in the mechanical, deterministic way the industry pretends.

Colour meaning is learned, not universal

The idea that colours carry fixed emotional meanings is the weakest part of the standard advice.

Blue is described as universally trustworthy. In practice, blue dominates finance, insurance and technology because those industries copied each other for forty years, and now blue reads as institutional because that is where we have all seen it. That is learned association, not psychology. It also means blue is the least differentiating choice available to you in those categories.

Colour associations vary by culture, by industry, and by individual history. White signals purity and cleanliness in Australian and European contexts, and mourning in parts of East Asia. Red is danger and stop signs here, and prosperity and celebration in China. Green means environmental in a marketing context and means go in a traffic context and means unwell in a health context.

Within a single market, context does most of the work. Red on a hospitality site reads as appetite. Red on a bank statement reads as debt. The colour did not change.

So when you read that a colour means something, the useful question is: means something to whom, in what context, compared to what else on the page.

What actually influences behaviour

Here is what holds up, and it is less romantic than the folklore.

Contrast, not hue

The most reliably demonstrated colour effect in interface design is not about which colour you choose. It is about how much a colour differs from what surrounds it.

The isolation effect, sometimes called the Von Restorff effect, is well established in the psychology literature: an item that differs from its surroundings is more likely to be noticed and remembered. This is the entire basis of a call to action working.

The practical rule is that your primary action colour should not appear anywhere else in significant quantity. If your site is largely charcoal and cream, a single warm accent used only on buttons will pull the eye. The moment you also use that accent for headings, icons, link text and section dividers, you have removed the contrast that made it work. You have not got a brand colour, you have got wallpaper.

This is the most common palette failure we see. Not the wrong colour. The right colour, used everywhere, until it stops meaning anything.

Hierarchy and consistency

Colour is most useful as a system for telling people what is what. This is a link. This is a warning. This is the primary action, and that one over there is the secondary action.

That works when it is consistent across every page. It stops working the moment your contact page uses a different button colour to your services page, which happens constantly on sites assembled from templates over time.

Legibility

Text you cannot comfortably read is the single largest conversion problem colour causes, and it is almost always self-inflicted by a designer working on a calibrated monitor indoors.

Light grey body text on white looks refined in the design file. Outdoors on a phone in Queensland sun, it is gone. Your customer is not in a studio.

Accessibility is the part that is not optional

This is where colour choice stops being taste and becomes a requirement.

WCAG 2.2 sets minimum contrast ratios: 4.5 to 1 for normal body text against its background, 3 to 1 for large text, and 3 to 1 for interface components and meaningful graphics. Level AA is the standard most organisations are held to, and it is the baseline we build every site to.

Around one in twelve men has some form of colour vision deficiency, most commonly red and green. That has a direct practical consequence: never use colour as the only way of conveying information. A form field that turns red to indicate an error, with no icon and no text message, communicates nothing to a meaningful slice of your visitors. Add the icon. Add the words.

Beyond compliance, this is simply commercial. Every person who cannot read your text is a person who cannot become a customer, and the population that struggles with low contrast is much larger than the population with a diagnosed vision impairment. It includes everyone over fifty, everyone outdoors, and everyone on a cheap screen.

Checking is free. WebAIM’s contrast checker takes seconds per pair, and browser dev tools will flag failures automatically.

Colour of the year, and why it is a poor input

Pantone named Cloud Dancer, a soft white, as its Colour of the Year for 2026, following Mocha Mousse in 2025 and Peach Fuzz in 2024. The stated reasoning is a cultural appetite for calm and clarity in a noisy world.

Trend forecasts are interesting as a read on the mood of a moment. They are a bad foundation for a brand palette, for one obvious reason: a brand is meant to last longer than a year, and anything visibly of a particular year will look dated on schedule.

There is a more useful way to read the current direction of travel. The move toward softer neutrals, generous white space and lower visual noise is real, and it has been building for several years across interface design. If your site is loud, high contrast in the aggressive sense, and crowded, it is reading as older than it is. That is worth acting on. Repainting your buttons in this year’s shade is not.

How we actually choose a palette

For what it is worth, this is the sequence we work through on a client build.

We start with what the business actually is, not what the category expects. A family concreting business operating since 2008 is not the same brand as a venture-backed construction tech startup, and if both end up in safe corporate blue, one of them has wasted the opportunity.

We look at what every competitor is doing, then deliberately avoid the centre of it. If every operator in the category is navy and grey, that is an argument against navy and grey, not for it.

We pick a dominant neutral for most of the surface area, one or two supporting tones, and exactly one accent reserved for actions. The accent gets used sparingly and nowhere else. That restraint is what makes it work.

We check every text and background pair against WCAG 2.2 AA before anything is signed off, not afterwards, because retrofitting contrast into an approved design means renegotiating the design.

Then we look at it on a cheap Android phone, outdoors. That is the real test and it fails more palettes than any other step.

The short version

Colour will not make somebody buy from you. It will decide whether they can read your page, whether they can find the button, and whether you look like everyone else in your industry or like yourself.

Those three things are worth getting right. The mystical bit, where a specific hue reaches into a customer’s subconscious and produces a sale, is folklore that got repeated until it sounded like research.

Be suspicious of anyone who quotes you a percentage.