Co-branding without the mush

Somewhere right now, a partnership launch is being designed by email. Two marketing teams are trading logo files — one sends a JPEG on a white box, the other sends a PNG with mysterious grey padding — and the result will be two logos on a slide, separated by a plus sign, each slightly different visual weight, both slightly resentful. It will look exactly like what it is: two brands that met at the airport.

Co-branding fails for a simple reason: it's treated as a layout problem ("put both logos on it") when it's actually a relationship problem ("who is the host, what is each brand for in this context, and what happens when it ends?"). Get the relationship explicit and the layout mostly designs itself. Skip it and no amount of kerning fixes the mush.

Here's the system we use when partnerships land in the studio — the same one behind the show-and-network lockups for Signal & Noise, where one brand had to host many without flattening any.

The pre-nup questions

Before a single pixel moves, both teams answer these out loud, together, in a shared document. The answers determine everything downstream, and the conversation costs one hour — the cheapest hour in the entire engagement.

Who is the host? Every co-branded surface has a host brand and a guest brand — the context the audience is in, and the reason they're there. An event hosted by you, sponsored by them: you're the host. Their product page announcing your integration: they're the host. "We're equal partners" is not an answer to this question; it's a way to avoid it. Equal partnerships still have a host per surface.
Whose audience is it? The brand whose audience is on the surface sets the voice, the palette leadership and the typographic voice, because the reader came to them and the guest is visiting the reader's home.
Which brand carries which job? Usually one brand carries trust ("you know us") and one carries novelty ("here's what's new"). Naming these jobs out loud prevents the daily negotiation of every headline.
How long does this live, and what happens at the end? A six-week campaign, a two-year integration and a permanent joint product need radically different co-brand systems — and different exit plans. More on the sunset below.
Who can say no, and how fast? Co-branding dies by a thousand unreviewed uses. Name one approver per brand, with a 48-hour response rule, or accept that the mush is coming.

If a partner won't answer question one, you don't have a co-branding problem yet. You have a partnership problem wearing a design costume.

Hierarchy: the three honest structures

Every durable co-brand is one of three structures. Choose deliberately — the accidental fourth structure, "both logos whenever someone remembers," is where mush comes from.

Hosted. The host brand owns the environment entirely: its palette, its type, its voice. The guest appears once — a lockup mark in a defined slot, often with a connector ("presented by," "in partnership with") — and then the surface behaves like a normal host-brand surface. Hosted is right for events, content series, sponsorships, and any context with a clear homeowner. It's also the structure that produces the least conflict over time, since only one design system is driving.

Endorsed. A new or weaker brand leads; a stronger one guarantees. The structure of imprint publishing and "a X company" footers. The endorser occupies a small, consistent, prestigious position — the seal position — and its restraint is precisely what carries the trust. We've covered the underlying decision tree in our piece on brand architecture; endorsement is the same logic, one level up, between companies instead of products.

Paired. Two brands of comparable weight, genuinely co-starring: a joint product, a co-authored report, a collaboration drop. Paired is the structure people assume they have and the hardest to do well, because it requires a real lockup — a designed unit — rather than two logos in proximity. If you can't commit to designing the lockup properly (below), demote to hosted and save everyone a month.

Lockup maths: the geometry of two marks

A paired lockup succeeds or fails on three measurements, and all of them are optical, not mathematical.

Weight before size. The single most common co-branding error: two logos set to the same pixel height that read wildly different weights. A thin wordmark at 40px next to a chunky symbol at 40px reads as a shout next to a whisper. Match optical weight — the perceived visual mass at the intended size — and let the pixel heights differ by as much as they need to, which is routinely 20–30%. Nobody measures logos with a ruler; everyone measures them with a squint. Design to the squint.

The divider and the clear space. Two marks in a lockup need a relationship element: a hairline divider, a "×", a "for", a "+", or disciplined space alone. Rules we hold to: a hairline divider is set at the thinner brand's stroke weight, in a neutral from the host's palette; an ampersand or "×" is set in the host's text face, never a decorative third voice; and clear space around the combined lockup equals the larger of the two brands' individual clear-space rules, not some negotiated average. The logo system, not logo principle applies double here: define the lockup as a unit with its own responsive behaviour — stacked under a narrow width, divider rules that survive a screenshot — or it will be re-invented badly, weekly, by everyone who touches it.

One baseline, one x-relationship. In horizontal lockups, both marks align to a shared optical baseline. In wordmark-to-wordmark pairings, match the x-height relationship of the glyphs, not the cap height — lowercase eye-height is where the eye actually compares, and matching caps while x-heights disagree is why so many paired lockups look like a tall person holding a short person's hand.

Shared palette: assign territory, never blend

The palette is where co-brands go to become mush most reliably. The instinct is to blend — tint everything toward both brands — producing that distinctive co-branded mauve with the energy of a wet paper towel. The better move is the opposite: assign territory.

The host owns the field. Backgrounds, large surfaces, the environment. Whose stage it is, whose colour it is.
The guest owns the signal. Accents, highlights, key interactive moments — applied with the guest's brand colour at full strength, undiluted. Full strength is the point: the guest reads as a confident visitor, not a ghost haunting the host's palette.
Never mix the two brand colours in gradients or tints. Brand A's blue blending into Brand B's green creates a colour that belongs to neither, contradicts both style guides, and photographs terribly. Blend nothing; juxtapose deliberately.
Type has one voice. The host's type system sets everything. The guest's typeface appears, at most, inside the guest's own wordmark. Two typographic voices on one surface is an argument conducted in fonts.

The test we run before ship: remove both logos from the layout. Can you still tell who is hosting and who is visiting from the colour and type behaviour alone? If yes, the palette is working. If the surface could belong to either brand, or neither, it's mush.

Voice and message: one narrator, one storyline

Collaborative copy dies the same death as collaborative palettes: by blend. Two teams approving one headline produce a sentence with the tempo of a committee meeting. The rule that saves it: the narrator is the audience's brand — whoever owns question two of the pre-nup — and the guest's contribution is substantive, not tonal. The guest supplies the facts, the numbers, the named expert for the quote; the host shapes the voice. Written down in advance, this rule prevents the fortnight-long duel in which every headline passes through two legal departments and emerges unbruised and unreadable.

For joint products, where neither brand is clearly the audience's home, the honest answer is usually that the product needs its own voice layer — lightly parented by both, owned by the product team. A joint venture trying to speak in two brand voices at once sounds less like either and more like a translation of a translation.

The trailer test and the credit roll

Every co-branded asset has a last frame — the end of the video, the footer of the page, the final slide — and both brands want it. Decide the trailer in the pre-nup: hosted surfaces end on the host, paired collaborations alternate or share a designed end-card, and nobody reshoots the ending after launch. It seems trivial until you've watched a partnership sour over whose logo closes the sizzle reel. The last frame is the possession arrow of co-branding; call it early.

The sunset clause: design the exit on the way in

Partnerships end — on schedule, early, or by acquisition — and the end is where unprepared co-branding gets expensive. A joint product page that took months to agree will take longer to unpick, with both teams suddenly remembering how much they loved the layout. So the co-brand system includes a sunset plan from day one:

An inventory of everything co-branded — pages, PDFs, templates, swag, that trade-show booth. If it isn't inventoried at launch, it will be discovered by a customer at the worst time.
Reversion rules. Which assets revert to which brand when the partnership ends, and within how many days. Evergreen content gets a pre-agreed "former partnership" footer treatment so nobody has to relitigate old blog posts.
The lockup's death certificate. The paired lockup has a written end-of-life: it stops being used, gets archived in both brands' living guidelines with the partnership dates, and both teams commit to not letting it haunt old slide decks forever.

Designing the exit isn't pessimism; it's what lets both teams commit fully at the start. The partnerships with the clearest endings are consistently the ones with the best launches.

Key takeaways

Co-branding is a relationship problem, not a layout problem. Answer the host/guest/audience/jobs/term questions before any design starts.
Choose one of three structures — hosted, endorsed, paired — deliberately. "Both logos everywhere" is not a structure; it's the mush.
Match optical weight, not pixel height. Give the lockup its own responsive behaviour and the larger of the two clear-space rules.
Assign palette territory: host owns the field, guest owns the signal, full strength, never blended. Type has one voice.
One narrator, one storyline — the audience's brand writes, the guest supplies substance.
Call the last frame in the pre-nup, and write the sunset clause on the way in. Clear endings make confident beginnings.

FAQ

Our partner's brand guidelines conflict with ours — whose win? The host's, per surface — which is why the pre-nup starts with "who is the host." On your surfaces, their logo is rendered as a guest within your system (their mark, your environment, correct optical treatment). On theirs, reciprocal. Where both brands' guidelines claim the same surface — a joint product homepage is the classic — that's a paired structure, and it needs a real, designed lockup and voice layer rather than a duel of PDFs.

How do we handle three or more partners on one surface? You almost never do. Beyond two brands, hierarchy must collapse into tiers: one host, one presenting partner, everyone else in a designed "with thanks" row at a set, subordinate scale. A row of six identical logos is not a partnership announcement; it's a dependency diagram. If a stakeholder insists all six deserve equal billing, the honest reply is that equal billing at small size is equal invisibility.

Can we create a merged "collab identity" — a new name and mark for the partnership? Yes, and for long-lived joint products it's often the right call — a child brand with both parents visible in its architecture. But it costs real money and ownership clarity to do well (who maintains it? for how long?), so we recommend it only when the partnership term and budget justify a genuine third entity. For anything under a year, a disciplined paired lockup beats a half-funded child brand every time.

What's the fastest useful co-branding artefact we can produce? A one-page co-brand sheet, agreed in week one: the three structures with which one applies, the lockup with its maths, the palette territory map, the narrator rule, and the approvers with response times. It's the single highest-leverage page in partnership marketing, and it prevents about ninety percent of the email-based design disasters this article opened with.