Notes · 29 April 2020
When clients send their own colours
The client's preferred colours are almost always wrong. A short post on what to do about it.
The client’s preferred colours are almost always wrong.
This sounds rude. It is, in fact, the kindest thing a designer can say to a client, which is why it is rarely said. The client has spent years looking at their own brand, on their own screen, in their own office, in their own well-lit photographs of their own products. Their preferred colours are the ones that have come to feel like the company. They are not, in most cases, the colours that the company should ship.
A few specific patterns I see.
The brand blue that is too bright. About one in three small businesses I work with arrive with a primary blue at roughly hue 220, saturation 85%, lightness 50%. It is the same blue Twitter shipped in 2007. It is on every SaaS landing page. It is, on a real screen at modern brightness levels, fluorescent. The client likes it because it has come to feel “trustworthy”. It is, in fact, generic.
The brand red that is actually orange. Clients who think of their brand as red are often, if you check the hex value they have been using, working with something around hue 10, saturation 80%, lightness 55%. That is not red. That is a warm orange. They have been calling it red because the original founder called it red on day one, and the swatch has drifted in everyone’s eye over the years.
The brand colour pulled from a logo that was rendered in CMYK. A great deal of the colour drift in small businesses comes from the fact that the original logo was designed in CMYK for print, the studio sampled the screen-rendered version of the printed PDF, and the resulting hex value has been used as the brand colour for years. The colour in the actual brand kit is not the colour anyone agreed on.
What I do, when the client sends their colours.
Take them as input, not as constraint. The colours the client sends are a starting point and a signal of intent, not a specification. The studio’s job is to deliver the colour palette the company should be using, which is sometimes the colour palette the client has sent, but more often a corrected version of it.
Show the client the comparison. Side by side. The client’s current blue, and a corrected blue with the saturation eased and the hue shifted slightly. Most clients, looking at the comparison, prefer the corrected version. They have just never been shown the alternative.
Use OKLCH for the work, even if the deliverable is in hex. OKLCH gives you a colour space that maps to perception more honestly than HSL does. The blue you pick at OKLCH 60 / 0.15 / 240 will look like the blue you intended. The blue you pick at HSL 220 / 85% / 50% will not.
Document the reason for each colour. The deliverable should not be a list of hex codes. It should be a short document that says, for each colour, what role it plays, what it is meant to feel like, and what it should not be confused with. The hex code is the smallest part of the brand colour decision. Most studios deliver only the hex code.
The clients who push back on the corrected colours are usually the clients whose websites I see on the next pitch round, with the original blue, six months later. The clients who accept the correction tend to keep the new colours for years.
For more on the technical side, the companion pieces on three colour pickers and the wrong one and colour at small sizes cover the underlying argument for OKLCH and the size-sensitivity problem. Lea Verou’s writing on colour is the cleanest current technical treatment, and Erik Kennedy’s piece on OKLCH is the practical introduction.