The Case for the Unscalable: Why human weirdness is your only moat left

For twenty years the marketing profession has been organised around a promise: that the more we industrialised the work, the more of it we could produce, and the more we produced, the more customers we would reach. It was a reasonable bet in a world where distribution was scarce and attention was cheap.

That world no longer exists. Distribution is now effectively free. Attention is scarcer than at any point in the history of the industry. And the middle tier of work — the competent, forgettable, on-brief output that used to be enough — has been quietly absorbed by systems that produce it at a fraction of the cost.

The models are extraordinary at the average. They are, by construction, a compressed representation of everything that has already been said. What they cannot produce is the specific: the sentence that could only have been written by one person, the shot that was only possible because a particular crew was in a particular room on a particular Tuesday.

This is not a mystical claim. It is a mathematical one. Distinctiveness is definitionally what falls outside the training distribution. The moment your work is legible to the model, it is legible to every competitor with an API key.

The specific is expensive. It requires taste, which cannot be delegated. It requires time, which cannot be compressed. It requires a leader willing to defend a piece of work that will never survive an attribution model.

The good news is that the specific compounds. A single distinctive campaign resets what a category is allowed to look like. A single distinctive point of view becomes the reason a category-defining talent decides to join you. Neither shows up in the quarterly dashboard. Both show up in the ten-year story.

The most important budget line in your marketing plan next year is the one you cannot justify. Not because you should be reckless, but because the entire rest of your plan is now available to your competitors on subscription terms.

The CFO conversation is easier than it looks. Every mature category has one or two brands that command a premium the rest of the market cannot access. In each case, that premium was paid for by a decade of the expensive thing.

One piece of work per quarter that could not have been produced by a large language model with a well-written brief. One conversation per month with a customer who does not fit your ICP. One decision per week that closes a door your competitors will happily leave open.

That is the whole programme. It is not new. It is what distinctive brands have always done. What is new is that the alternative — being competent, on-brief, and forgettable — is no longer safe.