Color Is Not

a Feeling

Most conversations about color in brand imagery are conversations about feeling. The words people reach for are impressionistic: warm, cool, muted, saturated, rich, airy. They describe what color does to us, not what it actually is. They are useful for briefing a mood. They are useless for briefing a system.

This is fine in casual conversation. For commercial imagery at any meaningful scale, it is a slow disaster.

The brands with the most distinctive visual identities — the ones whose images you recognize before you see the logo — did not build those identities by asking for imagery that “feels warm” or “looks clean.” They built them by making specific, technical, repeatable decisions about color and then enforcing those decisions across every touchpoint, every season, every market. The feeling is the output. The system is the work.

Understanding the difference between those two things is the first step toward building imagery that actually accumulates into a brand.

Color in commercial imagery is a decision system. Every pixel has a value. Every value is either a choice or an oversight. When you look at an image and respond to its palette — when a skincare ad feels clinical and precise, when a fashion campaign feels dusty and expensive, when a food photograph makes something look genuinely worth eating — you are responding to hundreds of individual decisions made during capture, processing, retouching, and output.

Most of those decisions, in most brand workflows, were never explicitly made at all. They accumulated. A photographer applied their house preset. A retoucher adjusted to taste. An art director approved because it “looked right.” Over time, the color shifted in ways that nobody could quite articulate, because nobody had defined what it was supposed to be.

The gap between a brand that has decided its color and one that has let it accumulate is the gap between a visual identity and a mood board.

Mood boards inspire. Identities compound. The former is useful at the start of a project. The latter is what makes a brand look like itself twenty campaigns later, with a different photographer, a different retoucher, and a brief that has been interpreted by a team that wasn't there when the original decisions were made.

There is a persistent misconception that color in commercial photography is set on set. That the lighting rig, the backdrop, the styling, the camera profile — that these are where the color lives. They are not. They establish a range. Post-production determines what's in it.

The lift of shadows. The hue rotation in skin tones. The temperature of whites against product surfaces. The saturation differential between foreground and ground. The way a specific fabric texture reads — whether it appears matte or dimensional, whether its true color holds or shifts toward the ambient fill. None of this is determined by the camera. All of it is authored in post.

This is why the framing of retouching as “cleanup” misses the point entirely. A retoucher working at the level commercial imagery demands is not correcting photographs. They are making decisions about what those photographs actually are. The camera records light. Retouching interprets it.

The brands that understand this hire for it, brief for it, and build it into every production workflow. The ones that do not treat retouching as the last mile of a photography job — something to be done quickly, cheaply, and without much creative direction — and wonder why their visual identity never quite coheres across channels.

Getting color right in a single image is achievable by almost anyone with time and taste. Getting it right across a campaign — across seasons, photographers, formats, markets, and years — is a categorically different problem.

This is where most brand color systems break down. Not in the hero campaign shot where the creative director is in the room and everyone is paying attention. In the secondary asset. The regional adaptation. The e-commerce variant. The product page image that was retouched by a freelancer working from a brief that said “match the reference.”

Without a technical standard — not “we like warm skin tones” but specific targets, documented curves, named and versioned LUTs — color drifts. It drifts between photographers. It drifts between retouchers. It drifts when the person who established the look moves on.

The brands with the most consistent visual identities have, almost without exception, reduced their color decisions to something that can be communicated, executed, and checked objectively. The look may have started as a feeling. But it was codified into a system. And the system is what makes it repeatable.

Codification does not mean rigidity. The best color systems are frameworks with intentional latitude built in — clear on what must never change (product color accuracy, skin tone approach, white point) and flexible on what can vary by context (campaign mood, seasonal warmth, editorial versus commercial treatment). What they are not is unspecified. Ambiguity in a color brief is not creative freedom. It is guaranteed inconsistency.

When people list what AI-generated imagery struggles with, the usual answers are anatomical: hands, teeth, text, reflections. These are real limitations, and they are improving rapidly. But for brand applications that require color precision, there is a subtler and more persistent problem: AI produces color that is probabilistic, not intentional.

A model trained on millions of images has learned what colors tend to appear together. It knows that cognac leather tends to sit against warm neutrals. That skin in soft window light leans slightly cool in the shadows and warm in the midtones. That a skincare product on white has a particular quality of light. These are statistical regularities. They produce images that look plausible.

Plausible is not the same as precise. The difference between “cognac leather in warm light” and your brand's specific version of cognac leather in your brand's specific quality of light is the entire gap between generic AI output and brand-ready creative. That gap does not close through prompting. It closes through retouching.

This is why expert post-production is not optional in AI workflows — it is the thing that makes AI imagery actually work for brands with real color standards. The model gets you to the range. The retoucher gets you to the specific. The model is fast and generative. The retoucher is where the brand identity enters the image.

Studios working effectively with AI imagery understand this and build their workflows around it. Generation is the starting point, not the finish line. The finish line is an image that looks like it could only have come from one brand — because the color decisions embedded in it are specific enough to be proprietary.

Treating color as infrastructure rather than instinct looks like a few specific things in practice. None of them are exotic. All of them require intention.

It starts with documentation.

Not a mood board. Not a reference folder of images you like. A technical specification that can be handed to any retoucher, any production partner, any AI image pipeline, and produce a consistent result. What is the target white point? How should product color accuracy be weighted against atmospheric grade? What is the acceptable variance in skin tone hue, and under what circumstances can it shift? These answers should exist as documents, not institutional memory.

It continues with process.

Color approval that happens at the right stage, reviewed on calibrated displays by people who have the authority and the vocabulary to make precise calls. Not “this looks a bit yellow” — but why it looks yellow, where in the curve that's coming from, and whether the fix is in the raw conversion or the final grade. The review process should be as technically specific as the work.

It ends with something most brands never quite reach: a proprietary color signature.

A color signature is what happens when you make the same decisions consistently enough, and specifically enough, that they stop being decisions and become identity. It is the point at which your imagery is recognizable before the logo appears — because the color itself carries the brand. It cannot be borrowed. It cannot be approximated from the outside. It is built through repetition, precision, and the patience to enforce standards that most people will never consciously notice.

That invisibility is the point. When color works, it does not call attention to itself. It simply makes everything feel inevitable. When it doesn't, the imagery feels slightly off in a way that people cannot name but reliably register — a faint wrongness that erodes trust without ever surfacing as a specific complaint.

Color is not a feeling. It is a set of decisions. The brands that treat it like one produce imagery that resonates. The brands that treat it like the other produce imagery that accumulates without compounding. The distinction, over time, is the entire difference between a visual identity and a library of photographs.