Answer in brief
You need a rebrand when the identity itself is the obstacle: a trademark conflict, a merger, a name that blocks the categories you now sell, or a promise the brand no longer matches. You do not need one for falling sales, inconsistent materials or a dated look — those are product, governance and surface problems, and each has a cheaper fix. Test the trigger against evidence before anyone draws anything.
Verified facts
- Source review
- 7 September 2026
- Reader need
- rebranding readiness checklist
Start from the problem, not from the logo
Almost every rebrand request arrives as a conclusion — we need a new look — rather than as a problem. The useful first move is to translate it back: what is observably happening, to whom, and since when. Most of the answers that come back are not identity problems at all, and the ones that are become obvious once they have been stated plainly.
Gathering that evidence is mostly mechanical, which is where AI-assisted preparation is quick and cheap: collecting what already exists, listing every place the current identity appears, comparing what different teams publish. The judgement it cannot make for you is whether the recognition you already hold is worth keeping. That one needs a person who will own the consequence.
Sales are falling
Usually means: a product, price, distribution or service problem that a new identity cannot reach. Do first: check whether the customers who already know you are still buying. If they are, the gap is reach or offer.
We look dated
Usually means: typography, spacing and photography on a handful of surfaces rather than the mark itself. Do first: rebuild the three surfaces you publish most, mark unchanged, and see whether it still reads dated.
Nobody knows who we are
Usually means: low awareness, which replacing the identity resets rather than repairs. Do first: separate never heard of you from heard of you and got the wrong idea. Those two need opposite treatments.
Our materials are all different
Usually means: no written rules and nobody owning them, which is a governance gap. Do first: establish whether documented rules exist at all. If they do not, a guideline document is the fix, at a fraction of the cost.
We have outgrown the name
Usually means: the name describes a narrower business than the one you now run. Do first: test whether it genuinely blocks what you sell today, or only what sits on the roadmap.
We merged or were acquired
Usually means: a real structural trigger. Do first: settle brand architecture — one brand, a house of brands, or an endorsed structure — because that decision governs everything drawn afterwards.
A trademark or market conflict
Usually means: a forced trigger on a deadline you do not control. Do first: get the legal constraint in writing, with classes and territories, because it fixes the scope before taste gets a vote.
Test the trigger before you accept it
Ask who first said the brand was the problem, what changed around that moment, and what they would accept as evidence that it is not. A trigger that survives those three questions is usually real. One that came from a single meeting, a competitor's launch or a new executive's preference usually does not, and it is considerably cheaper to discover that now.
Write the trigger as something observable: a conflict notice with a date, a category you cannot enter, a name that cannot be registered where you are expanding — trademark registers are public and searchable by class and territory, and WIPO is the usual starting point for international marks — or a promise that contradicts what you now sell. If it cannot be written that way, it is a preference. Preferences are allowed to win, but they should be labelled rather than dressed as necessity.
Measure what you already hold before replacing it
You cannot price a change without knowing what exists. Before anything is drawn, record what people type when they look for you, what they actually call you in conversation, which cues they recognise with the name removed, and where your own materials already contradict each other. Take it from sources you control and can show to somebody else.
Keep the measurement honest about its limits. A handful of customer conversations is a signal, not a study, and internal opinion is not recognition data. Label which is which inside the document, because the whole value of this step is that the decision can be re-read later by somebody who was not in the room.
Compare the cheaper interventions first
There is a ladder below a rebrand, and most requests are answered somewhere on it. Write the rules that do not exist yet. Rebuild the two or three surfaces carrying the most attention. Retire a sub-brand nobody defends. Change the typeface and the photography while the mark stays. Every rung is a fraction of a full identity change and finishes in weeks rather than quarters.
Price the rungs against the whole programme before choosing between them. If a cheaper rung would plausibly resolve the stated problem, take it and set a date to look again. A rebrand that follows a genuine failed attempt at something smaller is a far better-argued rebrand than one that skipped the comparison entirely.
Protect the recognition you already have
Recognition is an asset with a rebuild cost, and it rarely sits in the whole identity at once. More often it is one element: a colour, a silhouette, a single word, the shape of a container. List every candidate explicitly as keep, change or retire, and make somebody argue for each retirement, rather than letting it happen by default inside a concept presentation.
That list is also what separates a refresh from a replacement. If most entries say keep, you are planning a refresh and should price it as one. If the trigger genuinely forces most of them to change — a conflict, a merger, a name that cannot be used — then say so plainly, because the dependency work that follows is a different size of project.
Map the dependencies, because that is where the cost lives
The design fee is usually the smaller number. The real scope is every place the current identity is embedded: packaging artwork, signage, vehicles, uniforms, the product itself, contracts and documents, invoices, application icons, email templates, marketplace listings, and anywhere a legal entity name stands next to the mark. Count them before agreeing a budget, not after approving a concept.
Some of that work answers to rules your identity does not override. Packs carry codes governed by published specifications — GS1 publishes the barcode standards retailers verify against — so that artwork is reissued to specification rather than simply restyled, and anything regulated on the pack is reapproved rather than redrawn. This is ordinary work, but it is work, and it belongs in the estimate.
Write go, pause and stop criteria before you see concepts
Decide in advance what would make this a yes, what would make it a wait, and what would make it a no. Criteria written after the first concept presentation are not criteria: by then a persuasive drawing is competing with a spreadsheet, and the drawing tends to win. Name who holds each decision, and what evidence each of them requires.
A pause deserves defining properly, because it is a legitimate outcome: what is being waited for, who checks, and by when. Most programmes that ran badly did not fail at execution. They were started without anybody agreeing in advance what would have counted as not starting.
Be explicit about what this cannot fix
A new identity changes what you look like and what you are called. It does not change what you sell, what you charge, how quickly you answer, or where you are available. If the evidence points at any of those, the rebrand can be delivered correctly and the original problem will still be there afterwards — which is the most common way this kind of programme disappoints the people who paid for it.
Record those limits inside the decision itself. Naming what this work is not expected to address protects both sides later: it stops the programme being judged against an outcome nobody scoped, and it keeps the real problem visible to whoever does own it.
Record the decision, including a no
The deliverable of this stage is a short written decision: the trigger as stated and as verified, what was measured, which smaller options were compared, what is being kept, the dependency count, the criteria, and the date. A page or two is enough. Its purpose is to make the decision reviewable instead of remembered.
If the answer is not to proceed, that is a result rather than a failure, and it deserves the same care — including what would have to change for it to become a yes, and when somebody should look again. An undocumented no returns every quarter and costs the same argument each time.
Practical checklist
- Stated trigger rewritten as an observable problem, with the date it was first raised
- Recognition measured among people who already know the brand, before anything changes
- Cheaper interventions priced and compared: written rules, one surface, a retired sub-brand
- Every recognition asset listed explicitly as keep, change or retire
- Dependency map complete: packaging, signage, product, documents, listings, legal names
- Go, pause and stop criteria written and agreed before any concept is reviewed
- Known limits recorded: what this programme is not expected to fix
- Readiness decision written with its evidence and date, including a decision not to proceed
Questions and answers
How do I know whether we need a rebrand or just a refresh?
Ask whether the identity is the obstacle. If the mark, the name or the promise blocks something you need to do, that is a rebrand. If the same assets applied consistently would resolve it, that is a refresh or a governance fix, and it costs a fraction as much.
Is falling revenue a reason to rebrand?
Not on its own. Revenue moves for reasons a new identity cannot reach: product, price, distribution, service. Check first whether the customers who already know you are still buying. If they are, the problem is reach or offer rather than identity.
What does a rebrand cost beyond the design work?
The design is usually the smaller number. The cost sits in the dependency map: packaging artwork reissued to specification, signage, uniforms, product surfaces, documents, application icons, marketplace listings, and anywhere a legal entity name appears. Map it before approving a budget.
Can we keep our logo and still fix the problem?
Often, yes. Inconsistency is a rules problem and a dated feel is frequently a typography and layout problem, and both can be fixed while the mark stays. Replacing a mark people already recognise is the most expensive way to solve a problem that was never about the mark.
What if the decision is not to rebrand?
Write it down anyway, with the evidence and the date. A recorded no stops the question being re-litigated every quarter, and it names what would have to change for the answer to become yes.