Design Tokens Are Not Just for Software

Ask a software team how they keep a product visually consistent and they will say design tokens: named values like primary color, heading font, spacing unit, stored in one place, referenced everywhere. Change the token, everything updates. It is one of those ideas that sounds bureaucratic until you see it work, and then you cannot believe anyone ships software without it.

Ask a communications team the same question and the answer is usually a PDF nobody opens and a shared folder of old files people copy from. Then everyone wonders why the newsletter, the slides, and the banner have drifted into three different blues.

I have lived on both sides of that comparison, and I want to make the case that the software people solved a problem the rest of us are still pretending is unsolvable.

How brand drift actually happens

Nobody decides to drift. Here is the actual sequence, which I have watched play out in every organization I have served over twenty years of doing this work.

Someone needs a flyer by Thursday. They find last spring's flyer, which looks close enough, and build on top of it. But last spring's flyer was itself built from a file two years older, and somewhere along that chain a blue got eyeballed instead of entered, a font got substituted because a machine did not have the right one installed, and a logo got stretched three percent because the layout was tight. Each copy was a faithful copy of a slightly unfaithful copy.

Multiply that by every volunteer, every staff member, every deadline, every year. The drift is not a failure of care. It is the compound interest of small approximations, and no amount of reminding people to "check the brand guide" fixes it, because the brand guide is a description and people copy artifacts, not descriptions.

The software insight is that the fix is structural, not motivational. Stop asking people to remember the right values. Make the right values the thing they touch.

Steal the idea

I manage visual identity for multiple brands at once, across web, email, social, print, presentations, and merchandise. One of them I built from zero for an organization that launched globally at the end of last year, which meant every template, every color decision, every typeface choice had to be made fresh and then hold steady while dozens of deliverables poured out of it in the launch season. The only reason it holds together is that each brand is defined once, as a set of named decisions, in one authoritative place.

Not a style guide describing the decisions. The decisions themselves. There is a difference, and it is the whole point.

A style guide says "our primary blue is a deep navy." A token says: this exact hex value, named, with a job description attached. This one is for backgrounds. This one is for headlines on light backgrounds. This one is the accent, and it appears sparingly or it stops being an accent. When a color has a name and a job, people stop asking "which blue" and stop answering that question with their memory.

Same with type. The token version of a font decision is not "we use a clean modern sans." It is the family name, the weights we actually use, and the fallback order actually written down, because the day someone builds a slide deck on a machine without your font installed is not a hypothetical day. It is Tuesday. If the fallback is chosen in advance, the deck degrades gracefully. If not, the software picks the fallback for you, and software has terrible taste.

Then the less obvious tokens: spacing and logo-clearance rules, minimum sizes, and the voice file. That last one surprises people, but words drift exactly the way colors do. Is it "sign up" or "register"? Do we say "campus" or "location"? A short file of named language decisions governs words the same way the palette governs color, and it costs an hour to write.

Every template points at that source. When something changes, I change it once and rebuild outward. When a new deliverable appears, a conference banner, a product listing, an app screen, it starts from the tokens instead of from a memory of the last file.

The launch season for that global brand was the stress test. In the span of a few months the identity had to show up on a website, in email campaigns, across social channels, on event graphics, in slide decks, and on printed pieces, produced by more than one set of hands under real deadlines. In the old way of working, that is exactly the recipe for drift: pressure plus multiple producers plus speed. Instead, every deliverable was assembled from the same short list of named decisions, and the result was that a brand that was only weeks old looked more consistent than brands I have seen limp along for a decade. Age does not produce consistency. Architecture does.

Print is where it pays off most

Screens are forgiving. Screens are backlit, they vary wildly from device to device anyway, and nobody holds two phones side by side to compare your blues. Print is not forgiving. Print puts your materials next to each other on a physical table, under one light source, where every inconsistency testifies against you.

I have spent two decades producing conference materials and print pieces, and the failure mode is always the same. A near-miss blue that looks fine on a phone looks obviously wrong on a matte brochure sitting next to last year's banner. And by the time you see it, it is printed. There is no pushing a fix to production. The reprint costs real money, or, more often, the organization just lives with it, and the booth quietly communicates that details are negotiable here.

Tokens with real values, including print-space conversions decided in advance, are the difference between a booth that looks like one organization and a booth that looks like a garage sale. The conversion point matters more than people expect: the value that renders your color on screen and the value that prints your color on paper are different numbers, and if you let each vendor derive the conversion themselves, you will get each vendor's opinion. Decide it once, write it into the token, and hand every printer the same answer.

There is a spiritual cousin of this in live production, where I spent many years. You do not decide the camera shading during the event. You match the cameras before doors open, lock it, and then the whole broadcast looks like one show instead of four opinions. Tokens are pre-show camera matching for your brand.

What changes when it works

The payoff is not just consistency. It is speed, and it is delegation.

Speed, because starting from tokens removes the fifty small decisions that make a "quick" flyer take three hours. The colors are chosen. The type is chosen. The margins are chosen. What remains is the actual content, which is the only part that was ever supposed to take time.

There is a third payoff that took me longer to notice: better arguments. When the brand lives in named decisions, disagreements get specific. Instead of "that doesn't feel like us," which is unanswerable, the conversation becomes "that's the accent color doing a background's job," which is fixable in a minute and teaches something in the process. Tokens turn taste into terms, and terms are things a team can actually discuss.

Delegation, because you can now hand production to a volunteer or a new team member without handing them your taste. This matters enormously in the world I work in, where much of the production capacity is volunteers with wildly different design instincts. With a real token setup and templates that point at it, the floor rises. The least experienced person on the team produces something on-brand by default, because the defaults are the brand. I have watched this transform what a small team can put out in a week, and it is the difference between a brand that depends on one person's eye and a brand that survives that person taking a vacation.

The one-afternoon version

You do not need software, a design ops hire, or a committee. You need one honest afternoon.

Name your colors and write down their values. The real ones, pulled from your actual logo file, not from memory. Give each a job.

Write down your fonts and their fallbacks. Which family, which weights, and what happens on a machine that has none of them.

Put both into every template you actually use. Not the templates you aspire to use. The document people really start from when a deadline hits. If that is a slide deck and an email header, start there.

Then the last step, and it matters most: delete the folder of old files people copy from. Yes, actually delete it, or archive it somewhere inconvenient. As long as that folder exists, it will outcompete your beautiful new system, because copying an old file is one click and finding the official template is two.

Drift does not come from bad intentions. It comes from good people copying the wrong file. Take the wrong file away, hand them the right one, and the brand starts holding itself together, which frees you to do the work that actually needs your eye.