← Ideas

Why Brand Production Breaks at Scale

brandproductionscale

Eighteen years inside design studios taught me the same lesson everywhere. The brand system is perfect. The pipeline to execute it is not.

I've spent eighteen years inside design studios. Global ones. The kind with the unreasonable standards and the deadlines to match. I've built brand systems for companies where a misplaced pixel triggers a meeting and a wrong Pantone triggers a crisis.

The systems

Extraordinary on paper

The systems themselves are often extraordinary. Hundreds of pages of typographic rules, color logic, spatial grids, component libraries. The accumulated intelligence of very good people thinking very hard about how a brand should work.

And then someone needs a deck by Thursday.

What happens next has nothing to do with design.

It's production. Resizing, reformatting, swapping copy, adjusting layouts for a channel nobody planned for, exporting in four aspect ratios, getting feedback from three stakeholders with conflicting opinions, and doing it again when legal sends a note at 5pm.

The brand system is perfect. The pipeline to execute it is not.

System.

The waste

Not a talent problem

This isn't a talent problem. The people doing this work are often the best in the building. That's exactly the waste. You hired someone who can think at a world-class level and you've got them exporting PNGs.

It's not a tools problem either.

The tools are fine. Figma is excellent. The Adobe suite does what it does. There's no shortage of software. There's a shortage of architecture between the system and the output.

Architecture

The missing layer

What I mean by architecture: the layer that takes a brand system (the rules, the components, the logic) and turns it into finished work without requiring a human to manually translate every decision for every deliverable, every time. That layer doesn't exist in most organizations. So they solve it the only way they know how. They hire more people.

The loop.

And for a while, that works. More hands, more output. The queue gets shorter for a quarter. Then demand catches up because it always does. Marketing needs twelve formats instead of four. The company enters a new market and needs everything in three languages. Someone decides the quarterly report should be animated now.

The queue is back. Longer than before. And now you have a larger team grinding through the same structural bottleneck, just at higher cost.

Every level

Nobody has it figured out

I watched this pattern repeat at every level. Small agencies. Global networks. In-house teams at companies you'd assume have it figured out. Nobody has it figured out. The brand system gets more sophisticated and the production process stays manual. The gap widens every year.

Every other industry solved this decades ago.

Manufacturing separated design from production in the industrial revolution. Software engineering separated architecture from deployment with CI/CD pipelines. Even publishing automated the gap between editorial and distribution. But brand production? Still artisanal. Still manual. Still a person in a chair, moving things around a screen, one deliverable at a time.

The question

At some point I stopped asking how to make production faster

I started asking a different question.

What if the production layer didn't need people at all?

Not the thinking. Not the creative direction.

Not the strategic decisions about what the brand should say and how it should feel. Those are human problems and they should stay human problems.

Translation.

The mechanical act

But the mechanical act of translating a finished system into finished assets: the resizing, the reformatting, the export, the version control, the endless repetition of decisions that were already made months ago and documented in a PDF nobody opens?

That's not design. That's manufacturing. And it's waiting to be solved like manufacturing.

If your team is spending more time producing than thinking, the problem isn't your team. It's the architecture between the system and the output.

And nobody's building that layer.

I don't have a pitch at the end of this. Just the observation.

Almost nobody.