A company decides to redo its identity for one of two reasons. Either something in the market has stopped working and the identity is suspected, or something internal has changed — a new marketing director, a generational handover, a funding round — and the identity is the most visible thing available to change. The second reason produces most of the work commissioned in this sector, and almost none of the results.

It is worth being specific about what identity work can and cannot move, because the distinction is not obvious from outside.

Brand work changes recognition and preference. It makes a company identifiable at speed, gives a buyer language for why this supplier rather than another, holds a range together so that growth compounds instead of fragmenting, and lets a company be described accurately by people who do not work there. Those are real commercial outcomes with real financial consequences, and nothing else in the business produces them.

It does not fix distribution. If the product is not on the shelf, or the listing was lost, or three markets are covered by an agent who also represents a competitor, a new identity changes the appearance of a problem that is structural. The pack looks better in a place the buyer is not looking.

It does not fix a price position the buyer's own cost model has already fixed. In categories where the retailer builds the retail price from a landed cost and a target margin, a supplier's presentation has limited influence on where the price lands. It has influence on whether the supplier is in the conversation at all, which is a different question and should be argued as such.

It does not fix a sales team that cannot answer why-us. A well-built system gives them the answer; it does not install the habit of giving it. If the commercial organisation has spent fifteen years answering specifications, the identity arrives into a company with no muscle for using it, and it decays inside two seasons.

And it does not fix a product the market has quietly decided against. This is the hardest one to raise, because by the time a rebrand is being discussed, several people have staked their judgement on the product. But a redesign applied to a product with weak repeat purchase does not produce demand. It produces a more attractive version of the same decline, and it consumes the budget that could have funded finding out why the repeat rate is what it is.

There is a diagnostic that costs nothing and settles most cases. Ask what changed in the last twenty-four months. If the answer is that a competitor entered, a listing was lost, a distributor changed hands, or the category's price band moved, the problem is commercial and identity work is at best a component of the response. If the answer is that nothing changed — the company simply stopped being chosen, or was never chosen for a stated reason, or cannot be described consistently by two people in the same building — then the problem is one that identity work is actually built for.

A second test: who is uncomfortable. If the sales team is losing deals it expects to win and cannot explain why, that is a preference problem. If the marketing team dislikes the logo and the sales numbers are stable, that is a taste problem, and taste problems are expensive to solve and produce no return.

None of this means the identity is fine. Most are not. It means the sequence matters more than the quality of any single piece of work. A company with an unresolved distribution problem, a good product and an ugly identity should fix the distribution first and will be able to fund better work afterwards. A company with a resolved commercial position and no identity is leaving money on the table every quarter and has the clearest possible case for acting.

The reason this rarely gets said out loud is structural. The people who diagnose the problem are usually the people who would be paid to solve it, and the diagnosis tends to arrive at their own service. A company can protect itself from that by asking one question before any brief is written and insisting on a specific answer: what exactly is expected to change, in which number, and by when.

If nobody can answer, the work is not a bad idea. It is simply not yet a decision — and a redesign commissioned in place of a decision will be redone within four years, by someone who inherits the same unanswered question.