IdentityendsinthebrowserIdentity ends in the browser

A founder writes, three weeks after the brand identity has been delivered. The design was beautiful. The deck was signed off. The developer she hired separately has finished the build, and the site is live. Something is wrong.

The typography on the homepage is set in Inter, not the typeface chosen for the brand. The easing curves on the hero animation are cubic instead of bespoke. The hover state on the buttons doesn’t match what was in Figma. The mobile breakpoint compresses the layout in a way that makes the wordmark look apologetic. The packaging system, photographed in studio, looks like one brand. The site, in the browser, looks like another.

The mockups are still beautiful. The brand, in the place ninety-nine percent of customers will meet it, is not the brand the founder thought she had bought.

What happened sits in the structure of the work, not in anyone’s hands.

An identity is often delivered as Figma files, a brand book, static assets, and a handoff document — design treated as the thing that was made, and the site that follows as a separate, downstream build. The brand book describes what the site should look like; the build is where that description has to become a living thing.

The brand book is not the website. It is a description of the website. A description and a thing are different objects. The translation between them is not free — it is, in practice, where a system can come apart.

Where do brands actually meet people now.

A skincare brand sells primarily through its own e-commerce site, secondarily through a marketplace, and often through a single Shopify-built funnel that doubles as the entire brand experience. A real estate firm publishes its inventory and its philosophy through a website that is, for most prospective buyers, the only encounter with the brand before the in-person tour. A fashion house operates a brand site that is, for the seventy percent of customers who will never visit a physical store, the brand. An architecture practice’s site is the portfolio. A B2B platform’s site is the product.

The packaging matters when the customer holds the bottle. The signage matters when the customer enters the door. The print matters in the rare moment a printed object is in someone’s hand. For the rest of the time — and it is most of the time — the brand is the screen.

A brand identity that does not extend to the screen has not been completed. It has been described.

What gets lost between Figma and production is not always large. But it is always real.

Typography in Figma renders differently from typography in a browser. Letterspacing, kerning, the way a typeface holds at small sizes, the rendering on a Retina display versus a 100 percent display — these are properties the designer feels in Figma and the user feels on the screen. Without the designer in the loop at the implementation step, those feelings drift apart.

Motion in Figma is a static prototype. In production, it is a curve. The easing of a hover, the duration of a transition, the moment a navigation collapses on scroll, the way an image fades in when it enters the viewport — these are decisions made not in the design file but in the CSS. When the brand book is silent on them, they get decided at the build step by default — and what gets decided by default was never written down as the brand.

Layout in Figma is a snapshot. In production, it is a system that has to respond to every viewport between 320 pixels and 4000 pixels wide. The decisions about what to compress, what to hide, what to stack, what to reorder, what to keep proportional — these are brand decisions, not engineering decisions. They cannot be made in a static frame. They have to be made in code, on a real device, while the brand is being shipped.

Performance is brand. A site that loads slowly arrives apologising; a site heavy with unnecessary fonts spends bandwidth the visitor never agreed to. Performance is decided at the build, and it is part of the brand, not a separate engineering concern.

These are not edge cases. They are the brand, in the place the brand actually exists.

The handoff introduces a translator. Translation introduces loss. This is true of any creative work, and it is doubly true of brand work, where the system is meant to be coherent across surfaces.

Split the work across a handoff, and a thousand small decisions get made in a language the design file never spoke. The brand book describes the homepage at 1440 pixels. It does not describe it at 768 pixels and 320 pixels, or at the moment the user has rotated her phone. Those gaps get filled at the build step — and the filling becomes the brand, whether or not it was ever decided.

No one in this is at fault. It is simply what a split makes: in a discipline whose deliverable is one coherent system across surfaces, the half that wasn’t written down gets invented at the build, because that is where it first comes due.

The studio that ships does not have a translator. The same hand that draws the wordmark sets the CSS. The same eye that approves the typography on a packaging photo approves it in the running stylesheet. The motion grammar lives in a single file, not in a brand book and a separate codebase. The decisions are made once, by the same person, and shipped to the same surface.

This is not a brand-identity project handed to a separate website project. It is a single discipline.

It is also, in practice, a faster engagement. A studio that ships does not negotiate with itself across a handoff. It does not write specifications for itself. It does not annotate Figma files for itself. The decisions are made and implemented in the same week, sometimes the same day. The brand book is the codebase. The codebase is the brand book.

A fair objection: most brand work is still handed to a separate build, and not every designer writes production code. Mine does — that’s the method here, not a verdict on anyone else’s.

This is true. It is also why this argument is worth making.

When design and build are split, an identity can stay Figma-deep and screen-shallow — coherent in the deck, approximate in the browser. Keeping both in one hand is how that gap closes: the brand the founder approved is the brand her customer meets, because the hand that drew it is the hand that shipped it. And the gap matters more each year, because brands live, more and more, on screens.

Shipping isn’t a mark of better taste. It is a structural choice — that the brand the founder approved is the brand her customers meet, in the place they actually meet it. The choice has costs. It limits how much work I can take on. It asks the designer to be the developer too, with the discipline of both. It compresses the engagement into one pair of hands.

What it produces, in return, is a brand that is intact at the only place it has to be.

The phrase brand identity often pictures a product that ends with a deck — a wordmark, a palette, a type system, a moodboard, a brand book, sometimes a packaging mockup. The deck is the deliverable. What the customer actually encounters in the browser sits, in that picture, downstream.

But that picture misplaces where the brand lives.

A brand is a position held over time. A system is a position held under pressure. The pressure that brands face most often, and most consequentially, is the pressure of running on a screen, in front of a real customer, on a real device, at a real download speed. The brand identity has to survive that pressure. If it does not — if it is a beautiful description that fragments in production — it has not done its work.

A brand identity that exists only in Figma is a description of a brand. A brand identity that exists in the browser is the brand.

If a founder writes asking for a brand identity, the fuller engagement is one that ends where the customer will encounter the brand. Not at the deck. Not at the brand book. Not at the moment the build takes over. At the moment the site is live, the typography is rendering correctly, the motion is running, the performance budget is held, and the system the founder approved in design is the system her customer will actually meet.

Anything earlier is a partial answer. Anything earlier is a description.

What you can return to is what you have built. What gets shipped is what was built. Everything else is decoration.

The brand is the running system. Identity ends in the browser.