The gap between the colour theory people quote and the colour decisions people actually make
Ask a room of marketers why a brand is blue and someone will say trust. Ask why a discount sticker is yellow and someone will say urgency. These answers are repeated so often that they have hardened into folklore, and they survive largely because nobody checks them. The research they are loosely derived from was mostly conducted on small samples, in single cultures, using isolated colour patches with no product attached. Yet the folklore does real work: it gives a design team a shared vocabulary and gives a client a reason to say yes. Understanding brand colour means understanding both the technical reality — which is unglamorous and mostly about reproduction — and the social reality, which is that a colour decision is a decision made by a committee under commercial pressure.
This piece walks through how colour choices are actually arrived at in branding and product design: who is in the room, what constrains them, what gets tested and what does not, and where the process reliably goes wrong. It is deliberately more interested in process than in symbolism.
What a brand colour has to survive
A logo colour is not a colour. It is a specification that has to be re-created across a startling range of media, most of which the designer never sees. A single mid-range consumer brand might need its primary colour to appear on coated paper, uncoated paper, corrugated cardboard, injection-moulded plastic, powder-coated metal, woven label tape, embroidery thread, vinyl signage, a phone screen, a television advert, and an anodised aluminium fascia. Each of those has a different gamut — the range of colours the process can physically produce — and each has its own tolerances.
The colour that gets chosen is therefore rarely the most beautiful candidate. It is usually the most robust one: the candidate that degrades gracefully. A saturated orange that looks superb on screen may be unreachable in standard four-colour process printing, meaning every brochure will look slightly muddy unless the brand pays for a spot ink on every job. A very dark navy may read as black on cheap newsprint and on embroidered polo shirts. A pale mint may vanish entirely against white packaging under warm shop lighting.
The practical constraint list
- Process reachability. Can the colour be hit in CMYK, the four-ink process used for most commercial printing, or does it need a premixed spot ink? Spot inks cost more per job and are not available in every process, so a brand that depends on one accepts a permanent tax.
- Substrate shift. The same ink on uncoated stock looks lighter, duller and often warmer than on coated stock, because the paper absorbs ink and scatters light. Brands that care specify separate values per substrate rather than pretending one number works everywhere.
- Screen versus surface. A screen emits light and can be far more saturated than any pigment. Choosing a colour on a calibrated monitor and then discovering it cannot be manufactured is one of the most common and most expensive mistakes in the field.
- Small-size legibility. A colour used for a favicon, an app icon on a cluttered home screen, or a stitched label has to hold its identity at a few millimetres. Subtle colours lose their distinctiveness at that scale.
- Ownership. Someone will check whether a direct competitor already owns something close in the same category, because the point of a brand colour is discrimination, not beauty.
Who is actually in the room
The designer proposes. Very few designers decide. A typical decision path runs from a design studio to a brand manager, then to a marketing director, then in larger organisations to a legal review and a procurement or packaging engineering function, and finally to whoever signs. Each of those participants applies a different filter, and the filters are not weighted equally.
Design applies craft and differentiation. Marketing applies category convention and campaign flexibility. Legal applies trademark risk. Packaging engineering applies manufacturability and cost per unit. The signer applies personal taste, which is usually declared as strategy. It is worth being honest that the last filter is often the strongest, because the signer is the only participant who can end the discussion.
How this shapes outcomes
- Committee regression. Bold candidates lose votes because they are the ones anyone can object to. The survivor is often the least objectionable candidate rather than the strongest, which is why category palettes converge over time.
- The rescue narrative. A good designer anticipates this and presents each candidate with a rationale that gives the committee language to defend it. Colour choices with a story attached survive review meetings; naked swatches do not.
- Late technical veto. Manufacturing constraints often arrive after emotional commitment has formed, producing an awkward compromise colour that satisfies nobody. Bringing packaging engineering in early costs a week and saves a rebrand.
What testing can and cannot tell you
Brands with budget do test colour, and the tests are more useful than the folklore but less conclusive than the resulting slide decks suggest. Preference testing — showing people variants and asking which they like — measures preference, not behaviour, and preference for an isolated logo has weak links to purchase. Recognition testing is more useful: show the colour without the name and see whether people can attribute it. That measures the asset that colour actually builds, which is distinctiveness.
Shelf testing, where variants are placed in a realistic simulated retail environment with competitor products present and the shopper is timed, is the most informative and the least common because it is expensive. It captures the thing that matters at the point of sale, which is whether the pack can be found in under two seconds by someone who is not paying attention.
Common testing errors
- Testing on uncontrolled screens. Online panel testing shows each respondent a different colour, because their displays are uncalibrated and set to different brightness levels. The results are noisier than reported.
- Testing without context. A colour swatch on a white slide behaves nothing like the same colour on a pack surrounded by nine competitors under fluorescent lighting.
- Asking the wrong question. Which do you prefer produces polite answers. Which of these did you see first, or which of these is brand X, produces usable ones.
- Ignoring colour vision deficiency. Roughly one in twelve men has some form of reduced red-green discrimination. If a product range is coded only by red and green variants, a meaningful slice of shoppers cannot use the code.
Colour as a product decision, not a graphics decision
In physical product design the colour question is different again, because colour is inseparable from material. A colour applied to soft-touch polymer, brushed aluminium and matte ceramic will read as three different colours even when the pigment is identical, because gloss and texture change how much light is reflected specularly rather than diffusely. Designers who work in physical goods think in terms of CMF — colour, material and finish — precisely because separating them is meaningless.
This is also where cost bites hardest. A colourway that requires a second moulding run, an extra coating pass, or a pigment that fades under ultraviolet exposure changes the unit economics. Product colour ranges are usually narrower than the design team wanted, and the survivors are the ones that hold up over a three-year life on a shelf near a window.
Range strategy
- The anchor and the seasonals. Most ranges keep one or two permanent neutral colourways that carry the volume, plus rotating accent colours that create novelty and news without risking the core business.
- Forecast risk. A seasonal colour is a bet placed twelve to eighteen months in advance. Getting it wrong means markdowns, so the accent colours tend to hug whatever the trend services have signalled.
- Fade behaviour. Pigment lightfastness matters more in product than in print, because a phone case lives outdoors. Cheap organic pigments in the red and violet families are the usual offenders.
Writing a specification that actually holds
Once the colour is agreed, the deliverable is a specification, and weak specifications are why brands drift. A brand book that lists only a screen value and a spot ink number is incomplete. A robust one gives target values in a device-independent space along with a stated tolerance, so a supplier in another country can verify a match instead of eyeballing it.
- Device-independent target. Specify the colour in CIELAB coordinates measured under a stated illuminant and observer, because these describe the colour itself rather than one machine’s recipe for it.
- A stated tolerance. Give the acceptable deviation as a Delta E value with the formula named. Without a tolerance, every argument with a supplier becomes a matter of opinion.
- Per-substrate values. Different numbers for coated, uncoated, film, textile and moulded parts, each measured rather than guessed.
- A physical standard. A signed-off physical sample, stored properly and reissued periodically, because printed references fade and the drift is invisible until someone compares an old book with a new one.
Where the folklore does and does not help
None of this means colour meaning is irrelevant. It means the meaning is contextual and mostly learned from the category itself rather than from anything intrinsic to the wavelength. Green reads as environmental in cleaning products because cleaning products taught people that, not the other way round. Category conventions are real and worth knowing, because a brand that violates them is either invisible or memorable, and the difference is usually execution and budget.
The useful posture is to treat colour meaning as a hypothesis about the audience rather than a fact about the colour, to spend the testing budget on distinctiveness and findability rather than preference, and to spend the engineering effort on making the chosen colour survivable across every surface the brand will ever touch. The brands that look consistent are not the ones with the best-argued colour rationale. They are the ones with a measured specification, a stated tolerance, and someone whose job it is to check.
