Four Things That Get Sold When a Brand Stops Working

The brief usually arrives already answered. "We need a rebrand." Sometimes it comes with a mood board attached and a launch date in the subject line, which means the decision was taken in a meeting nobody minuted and the diagnosis was skipped entirely.

Underneath that single word sit four different interventions. A refresh, a reposition, a rebuild and a rename. They take different lengths of time, cost amounts that are not in the same range, and fix genuinely different problems. Usually only one of them matches the symptom that started the conversation.

That list runs roughly from cheapest to most expensive and from least disruptive to most. Teams tend to buy from the wrong end of it. This piece is about how to tell which one you are actually looking at, using evidence you can gather in a fortnight rather than a conviction formed in a corridor.

Four things are on the menu

A refresh changes how the brand looks. A reposition changes what it claims. A rebuild changes the system that produces everything the brand makes. A rename changes who you are in a search box, on an invoice and in a contract. Four scopes, four bills, and only one of them usually matches the complaint.

The blur between them is understandable rather than sinister. All four eventually produce new artwork, so all four look alike in a presentation. The differences only surface eighteen months later, in whether the original problem came back.

Refresh: the surface, and only the surface

A refresh keeps the name, keeps the claim and modernises the expression. Typography, a tightened palette, photography rules that reflect how images are actually used, a mark redrawn rather than replaced. It is a visual identity project with a boundary around it, and the boundary is the point.

What a refresh genuinely fixes

It fixes looking older than the work you do now. It fixes an identity designed for print being asked to survive on a phone. It fixes eleven versions of the same logo circulating internally because nobody ever published a correct one. Those are real problems with a real cost, and a refresh solves them completely. It is the most under-sold of the four for exactly that reason: the fix is unglamorous and the announcement is quiet.

What a refresh cannot touch

A refresh cannot change what people believe about you. If prospects think you are expensive, generic or junior, new typography delivers that same belief in a better font. If the sales team keeps losing on one objection, a refresh spends the budget and hands the objection back intact. The belief lives in the claim, and the claim is a different project.

The smallest honest version of a refresh is worth naming, because plenty of brands do not own fifty pages of website to restyle. For a founder, a small practice or a single-location business, the entire public surface might be one page, and something like Nichevio assembles precisely that: a polished, mobile-first profile site built from structured widgets. Match the scope of the fix to the scope of the surface, not to the size of the ambition.

Reposition: changing what you claim

A reposition changes the claim. Same name, same or lightly refreshed look, a different sentence about who you serve and what you are better at. Most of it is not a design project at all. It is research, then a decision somebody has to be willing to sign, then a rewrite of every sentence that contradicts it. The output has to survive a customer with three alternatives open in other tabs, which is the standard we set in choosing a territory you can actually defend.

Where the money actually goes

Not into artwork. The cost of a reposition is agreement first and consistency afterwards. Every page, deck, proposal, job advert, onboarding email and support macro still arguing the old position has to be found and changed. The ones nobody changes become the evidence people cite when they say the reposition never landed.

This is also the intervention with the most predictable failure timetable. The launch goes well, the first six months hold, and then the old vocabulary creeps back through everybody who was not in the room. We have set out that pattern in why rebrands fail in the second year, and the remedy is boring: fewer people with authority over the words, and rules written down where the work happens.

Rebuild: the system under the logo

A rebuild replaces the machinery rather than the picture. Tokens, a type scale, spacing rules, components, iconography, motion, templates for the things you actually produce, and written answers to the decisions nobody thought to write down. The mark may barely change. Everything about making work with it does.

What "system" has to mean to be worth the money

• Decisions, not swatches. A palette is a list of colours. A system says which colour carries an action, which carries a warning, and what happens when the two sit next to each other.
• Templates for the work you really make. The proposal, the case page, the recruitment post, the invoice. Not three hero variations nobody will ever build again.
• Rules for the awkward cases. Long names, narrow screens, dark backgrounds, one-word headlines, the mark sitting on a photograph somebody else chose.
• A findable home for the files. Working files, licensed fonts and export rules, stored where the next person can reach them without asking a favour.
• Defaults that settle arguments. Contrast, focus states and target sizes decided once, in the tokens, so nobody relitigates them screen by screen.

That last point is where brand work and build work stop being separate budgets. If a rebuild touches how pages are constructed, the accessibility decisions worth making before any audit belong in the same pass, because contrast and focus are far cheaper to settle inside a token set than to retrofit across a hundred screens.

A rebuild is the intervention that pays back longest and excites clients least, because the deliverable is not a picture. It is the thing that stops the brand drifting back within a year of the launch party.

Rename: the one with a bill attached

A rename is the only one of the four that changes what a contract says. It is also the only one where a registrar, a bank, an app store and a search engine all hold opinions you cannot design your way around.

It is warranted for a short list of reasons: the name is legally contested, it is unusable in a market you are entering, it describes something you no longer sell, or it belongs to a merger that has since dissolved. What makes a name worth the trouble of changing works through that test properly. Everything outside the list is taste, and taste is an expensive reason to renegotiate your own paperwork.

The redirect map is the artefact that decides whether the move holds. That is why we treat a rename as a migration project carrying a brand component, rather than a brand project with a technical footnote at the end of the deck.

There is also a cheaper alternative that gets skipped. When the problem is a new offer that does not fit the existing name, the question may be architectural rather than nominal: does the new thing get its own name, or sit endorsed under the one you already have? That is the argument in branded house or house of brands, and our parent company has written the small-company version in brand architecture for a small company. Answer it before anybody commissions artwork.

The evidence that points at each one

Diagnosis is not mysterious. Each of the four leaves a distinct fingerprint, and the fingerprints are things people in your own building already say out loud, usually in the week before somebody books an agency call.

Two rules for reading that table. First, one symptom can have two causes, so a row is a hypothesis rather than a verdict. Second, the right-hand column is a floor and not a ceiling: finding a rebuild does not license a rename on the same invoice.

Gathering the evidence takes about two weeks. Interviews with people who chose you and, more usefully, with people who did not. Win and loss notes read rather than summarised. An inventory of what exists, who made it and which parts are still in use. That inventory is most of what a brand audit is for, and it is worth commissioning even when the answer turns out to be that very little needs to change.

The honest concession: this evidence is never conclusive. Ten interviews will not settle a contested position. They narrow six candidate explanations to two, which is enough to stop you buying the wrong intervention and not enough to promise the right one will work. Anybody offering more certainty than that is selling something else.

Why the biggest one gets sold by default

Three reasons, none of them a conspiracy. A rebrand is legible internally in a way a refresh is not: it comes with a launch, a film and a moment of applause, and applause is difficult to budget for but easy to want. Scope is also simpler to sell than diagnosis, because diagnosis is billed in days and scope is billed in months. And a new mark feels like progress in a way a rewritten positioning sentence never does, even when the sentence is the thing that changes the pipeline.

The most expensive rebrand is the one that fixed a problem you did not have, because the problem you did have is still there on the Monday after the launch.

The counter-move costs nothing. Write one sentence describing the business outcome that should be different once the work is finished, and have the person paying for it read that sentence aloud. Vague outcomes produce vague scopes. Something like "mid-market prospects stop asking whether we do implementation" points straight at a reposition and quietly rules out three other quotes.

The order we would run it in

Where there is genuine doubt, this sequence limits the damage of guessing wrong.

1. Write the symptom as a sentence somebody actually said. "Our biggest prospects think we are a web shop" is a symptom. "The brand feels tired" is a mood, and moods do not have fixes.
2. Gather two weeks of evidence and then stop. Interviews, win and loss notes, an inventory of what exists. Enough to choose between explanations, not enough to build a research programme around.
3. Settle the position before anybody draws anything. One sentence on who it is for and what it is better at, agreed by every person who could veto it in month nine.
4. Fix the system before the surface. Tokens, templates and rules first. A surface built on a system holds its shape; a surface built on enthusiasm drifts by the second quarter.
5. Treat the name as the last question, not the first. It is the most expensive item and the easiest to defer. When the position is right and the system holds, a merely unfashionable name often stops being the complaint.
6. Publish the rules where the work happens. A system nobody can find is a system nobody uses, and the drift then starts in month three rather than year two.

When the right answer is to do nothing yet

Sometimes the brand is not the problem at all. If the product is half-finished, the pricing is unreadable, or nobody has published anything for two years, a new identity simply delivers an unsettled message more attractively. The work will be blamed for the result, and it will not deserve the blame.

The test we apply before committing anybody to a scope: name the decision this work makes easier. If the honest answer is "we would look more current", that is a refresh and it should be priced as one. If it is "our best-fit prospects would understand what we do without a call", that is a reposition, and no amount of artwork substitutes for it. Our parent company argues the same case across the wider services question in what gets sold when a brand stops working.

The rule that survives contact with most briefs is this one. Buy the smallest intervention that could plausibly fix the symptom, then check whether it did. It is unfashionable advice, it produces smaller invoices, and it leaves the budget for the larger intervention intact on the day you can prove you need it. If you want a second opinion on which of the four you are looking at, tell us the sentence somebody said out loud and what it was about.