Ask what a brand is and three answers come back. Brand is what we want to be. Brand is what we say. Brand is what we do. Marketing has been living almost entirely in the second, and the enterprise has settled for a thin version of the first: the mission statement, all ambition and no instruction. The third is where brand is made, and it is the one organisations least often think of as brand.
Most organisations respond with the inherited tools. Tone of voice documents and coffee-table brand books help with expression but add little that informs judgement. The brand onion and its relatives, the pyramid, the key, the wheel, the Venn diagram, the house, the temple… they get a workshop to agree on a vocabulary and do little else. The finance director, the engineers, the head of HR, the head of e-commerce, the people running customer service and the sales director have never seen them. Had they looked, they would have found no principles for making a choice, and none for settling a conflict between two good ones. The diagrams describe badly. They govern nothing.
The mission statement, the enterprise’s version of the first answer, has the look of authority without exercising any. A verb of ambition (empower, inspire, transform), attached to a beneficiary as broad as humanity and an outcome no one could oppose, generates no constraint.
The positioning statement, however essential it is as a building block, fares little better: For a given audience with a given need, the brand is the thing that delivers a given benefit, for a given reason. It is short, and that is its discipline and appeal. But leaving aside how readily its blanks fill with what April Dunford has characterised as “a hilariously awkward franken-statement of meaningless mumbo jumbo”, it gives little or no help when the company has to decide what gives way to what.
The root source of brand lies in the company’s value and growth philosophy, and it shows in its decisions: where (to borrow from Roger Martin) it chooses to play and how it believes it can win, what it makes, what it charges, what experiences and partnerships it builds, what it refuses, and yes, what and how it communicates.
Of course not every strategic decision the enterprise makes is a brand decision: capital structure and tax domicile carry little brand consequence. But every decision a customer, an employee, a reviewer or now a machine can read or feel is a brand decision, whether or not anyone labels it so, down to how customer service responds at 11pm.
A small company can get this right by default. Everyone sits near the founder. Decisions pass through a handful of heads that share the same instincts, and the product, the marketing and the culture feel like one thing because the same few people make them.
Scale removes that. New teams arrive with their own priorities, new markets reinterpret the brand for local conditions, agencies multiply, product drifts away from marketing… and each department makes reasonable decisions that add up to an unreasonable whole.
The fact is that showing up on purpose, with consistency, is about what we do, not just about what we say. Aristotle put action first: “Character comes in as subsidiary to the actions.” Hollywood still works by the rule. The screenwriting teacher Robert McKee separates characterisation from character. Characterisation, he argues, is the surface: looks, dress, manner of speech. But character is revealed in the choices a person makes under pressure, and the greater the pressure, the deeper the revelation. In Star Wars: Episode IV (1977) Han Solo presents as a gunslinger for hire, holster slung low. But we learn who he really is when he turns the Millenium Falcon around to cover Luke’s run on the Death Star. Adjectives and abstraction can’t capture character. It lives in choice and behaviour.
The Say-Do gap, between stated values or characterisation and behaviour under pressure, usually comes down to translation. The values are never turned into anything usable. “Customer first” and “integrity” are sentiments. They don’t tell a store manager what to do when a refund policy collides with a distressed customer. They don’t help an engineer decide whether to ship on time or ship right.
AI is about to make the gap much larger. The volume of output a company produces is rising sharply across every channel, and more of it is generated by machines. Increasingly, agents act for the brand, answering customers, writing content and making small decisions nobody reviews. Every one of those voices needs something to draw on when no human is in the room. The core logic a company sets down today will soon be the instruction set its agents work from, and an organisation without one will find its machines improvising its identity for it, at scale and in public. A brand diagram fed to a model is like showing a dog a card trick. Which is why its utility has dropped below zero. It gives the appearance of guidance where there is none.
The machines now sit on the buyer’s side as well. Customers increasingly use them to research for them, and the machines read what a brand delivers: its performance, its pricing relative to competitors, its returns terms, its service record, what reviewers report about its Say-Do gap.
BCG argue that “brand promise and brand performance must be fully aligned… CMOs and marketers are typically responsible for defining and stewarding the brand promise, but they have historically not had full responsibility for brand performance. While narrative, storytelling, and positioning are important, they cannot bridge the gap between what customers expect and what the organization consistently delivers… The work of defining the brand promise becomes more rigorous, grounded in what the company can demonstrably deliver and where it is genuinely differentiated from competitors, rather than what it claims...”
BCG correctly identify that the existing tools can’t reach the business. But (unsurprisingly) they stop there, without suggesting what can bridge the gap.
The best organisations will recognise that brands are not merely external veneers, the residue of promises and adspend alone, but the consequence of the company’s own decisions — manifested in consistent, coherent, legible patterns of behaviour in the marketplace that customers, and now machines, can read.
They will close the Say-Do gap by codifying their commitments so that a system (human and machine) can act on them. A commitment worth the name states the moment it governs, the trade-off it settles and the line it will not cross. “We put customers first” governs nothing. “When a customer reports a fault inside the warranty period, we replace first and investigate afterwards, even when the claim looks doubtful” governs a great deal. A store manager can use it on a Saturday afternoon. So can a service agent at three in the morning.
The work is to find the moments that matter across every surface area of the brand or organisation – product, service, pricing, communication, etc. – and to decide in advance which way the company leans in each. Culture then works as a decision-making system, and the same commitments brief both the people and the machines.
None of this can be delegated in its entirety to the brand team. It asks the leadership team to draw out what the company will commit to, rather than merely espouse, and what it will refuse; to decide which commitment wins when two collide; and to define each one precisely enough for a new hire or a machine to act on. I call it the Character Code. The policies, the manuals and the instructions given to every agent then follow, written by the people who own them.
The work is hard. Every Always and every Never closes down what an organisation most loves to keep open: its options. An Always forecloses. A Never forecloses. Each one shuts a door the business might later wish it had left open, and refuses to reopen it when circumstances or incentives shift. But that’s the point.
Duolingo shows this in action. Its own handbook concedes that sending learners more notifications would lift daily active users, one of the numbers it watches most closely. It caps them anyway, “regardless of what the short-term metrics suggest”, because people who feel bombarded eventually switch off or leave. The rule was written to survive a good week on the dashboard. And the messages it governs are sent by machines, millions a day, with no one approving each one.
Could a brand onion have produced that rule? Peel it ring by ring, through the values and the personality and emotional benefits and functional benefits to the essence at the centre, and nothing in it will tell a machine to send one message fewer. It never told a human either. The onion rarely informed or settled a decision inside the marketing department, let alone beyond it, and its utility was close to zero long before the machines arrived. People simply worked around it, drawing on instinct and on whoever had been there longest. Hand the machines the onion and they will borrow its adjectives and make up the rest. That’s how a useless document becomes a harmful one.
A description can be admired. Only a commitment can be followed.