The Anatomy of a Rebrand

Nearly every rebrand brief contains a version of the same sentence: our brand no longer reflects who we are. It is usually true. It is also almost never specific enough to act on, because "who we are" could have changed in five different places — the customer, the product, the price, the promise, or just the kerning on a logo drawn in 2011.

Getting that answer right before anyone opens a design tool decides everything downstream, because the three things people call a rebrand are three different projects with different costs, different risks, and different failure modes.

Refresh, rebrand, reposition

A refresh leaves the strategy alone and modernises the execution. Same customer, same promise, same name — better craft. Type that holds up at 14px on a phone. A palette that passes contrast without a fight. A mark that still reads at favicon size. Nobody has to relearn anything, the risk is low, and so is the ceiling: a refresh will not rescue a business selling the wrong thing to the wrong people.

A rebrand changes how the business presents itself — identity, voice, sometimes the name. The strategy underneath may be perfectly stable, but the surface is deliberately different enough that people notice and have to update their mental picture. This is where most of the budget goes and most of the mistakes get made.

A repositioning changes what the business claims to be: the category it competes in, the customer it serves, or the price it can defend. That is a commercial decision that a brand system then expresses. Doing the identity work first and hoping the strategy will catch up is the most expensive sequencing error in this discipline. You end up with a beautiful, internally consistent system arguing for a position the leadership team never actually agreed to — and it will be quietly ignored within a year.

Ask which of the three you are doing. The honest answer moves the budget by an order of magnitude and the timeline by months.

Equity lives in a very small number of assets

Brand equity is not spread evenly across a brand system. It concentrates, hard, into a handful of things people can recall without being prompted: usually a name, one colour, one shape, and occasionally a phrase or a sound. Everything else — the secondary palette, the illustration style, the icon set, the photography direction — is craft. Craft can be replaced without anyone flinching.

Ask a customer to draw your brand from memory. Whatever they get right is the part you are not allowed to throw away.

Teams reliably protect the wrong things. They will defend a typeface no customer could name, then change the one colour doing all the recognition work — because it has been around long enough to feel boring internally. Internal boredom is not a signal. It is the sound of an asset working: you see it every day, your customers see it four times a year.

A workable method: inventory every asset and score each on two axes. How distinctive is it — could it belong to a competitor? And how known is it — do people outside the building actually attach it to you? Assets scoring high on both are load-bearing, and you change them only with a deliberate plan to transfer recognition to whatever replaces them. Assets scoring low on both are free, and probably should change.

What actually changes

Once the category is settled, the work divides more cleanly than briefs suggest:

• The argument. What you want to be chosen for, and who you are willing to lose in order to be chosen by the people who matter. If nothing is being given up, no position has been taken.
• The identity. Mark, type, colour, layout logic, imagery. The part everyone means when they say rebrand — and roughly a third of the work.
• The voice. Not a list of adjectives — a set of decisions about what you say first, what you refuse to claim, and how you sound when something has gone wrong. Voice is tested in error messages and refund emails, not in headlines.
• The system. Tokens, components, templates, and the rules for extending them. A brand that exists only as a PDF will drift within two quarters.
• The governance. Who decides, after launch, whether a new thing fits. Skip this and the brand is defined by whoever happens to be building the next landing page.

Sequencing a rollout so nobody gets stranded

There are two honest rollout philosophies and a lot of unhappy middle ground.

Everything at once is coherent and newsworthy, and brutal on operations. Every surface has to be ready on the same morning, including the ones nobody remembers: invoice templates, the email signature block, the 404 page, app store screenshots, the recruiter's slide deck, the sign above the goods entrance. The failure mode is a launch that looks immaculate on the homepage and unmistakably half-finished everywhere a customer actually spends time.

Phased takes surfaces in order of visibility and lets low-traffic material transition on its natural replacement cycle. Cheaper, calmer, and it accepts a long stretch where the brand looks like it is mid-argument with itself. That is survivable if the phases are short and the sequence is legible; it is not survivable as a permanent state.

Whichever you choose, one rule governs everything: the customer must never be uncertain they are in the right place. In practice that means:

1. Old domains redirect, they do not 404 — and they keep redirecting long after the project team has moved on. Links live for years.
2. During transition, the old mark and the new one appear together in a fixed lockup on the surfaces where identity is doing verification work: login, payment, transactional email, packaging. Everywhere else, switch cleanly.
3. Nothing else changes at the same time. Do not ship a new identity alongside new pricing or a rebuilt checkout. If two things change and confidence drops, you have learned nothing about which one caused it.
4. Customers hear it from you before they discover it. Someone who reads "we're changing our name next month" and then meets the new name is being reassured. Someone who meets the new name first is being surprised — and surprise, at the moment of a transaction, reads as risk.

The parts nobody budgets for

The long tail is where rebrands overrun. It is rarely the identity; it is the hundreds of places the old one is embedded. Contracts and legal templates. Automated emails written six years ago by someone who has left. Third-party listings and review profiles. Partner co-branding. Printed material sitting in a warehouse.

Search deserves specific attention if URLs are changing. A page-by-page redirect map is unglamorous, and it is the difference between carrying your visibility across and starting again. Expect a dip regardless; plan for it rather than panicking in week three.

Then there is internal adoption, consistently underfunded and consistently decisive. The highest-volume users of any brand are the people who work there, and they will use whatever is easiest at 4pm on a Thursday. Guidelines nobody opens lose to a component library and a set of ready-made templates every time. Make the correct thing the easiest thing and most of the enforcement problem disappears.

How you know it worked

"Does it look good" is not a test — everyone involved has been staring at it for months and has lost the ability to answer. Better questions, asked a year later:

• Can someone outside the company identify a cropped asset with the name removed?
• Has the number of one-off exceptions gone up or down? A system that generates constant exceptions was designed for a business that does not exist.
• Can the team produce a new page, deck, or campaign without asking anyone for permission — and does it look right?
• If you repositioned, are sales conversations happening on the ground you chose, or on the old one?

A rebrand is not a new coat of paint. It is a decision about what you want to be recognised for, plus a system that makes that decision repeatable by people who were not in the room when it was made. Get the decision right and the design work becomes obvious. Get it wrong and no amount of craft will save it.