Imperial Advertising (In A Feudal Media World)

The territory fragmented. The advertising never did.

Alan Wolk called it feudal media: the monoculture collapsed into a thousand disconnected bubbles, each with its own rulers, its own borders, its own rules of entry. The diagnosis was accepted almost immediately, because everyone in the industry could feel it was true. 

What nobody looked at is what happened to the advertising travelling across that map.

The territory broke into duchies. Each one developed its own language, its own rhythm, its own etiquette, its own relationship between the viewer and the thing being viewed. And the advertising that moves between them is still built exactly the way it was built for the monoculture: one message, one format, one grammar, dispatched identically into every territory, as if the empire were still standing. Not that there are no versions. We ship them by the hundred. But they vary by format, duration and crop, not by what they say, and not by the relationship they assume with the person watching.

The industry noticed the fragmentation — or said it did. Either way, it answered the wrong question.

Advertising solved addressing. It never solved language.

Confronted with a thousand bubbles, advertising responded with hypertargeting — and targeting is an addressing solution. It answers who sees this and where. It does not answer does this belong here. We spent two decades becoming surgically precise at delivering the same message to the right person inside the right bubble, and almost no time making the message native to the bubble it landed in.

This isn't hindsight. The IPA's databank showed it years before generative AI existed: high-performing campaigns are more likely to use big data for insight, while low-performing campaigns use it more often for targeting. The same dataset found that broader, earlier targeting consistently beat tightly targeted real-time communications tied to purchase intent.

We read that finding as a media planning note. It was a warning about language.

The variable that explains half the outcome

Meanwhile the evidence on what actually drives results kept pointing in one direction. Circana's analysis puts creative impact at nearly half of all incremental sales generated by advertising — the single largest driver of campaign performance. Not the targeting. Not the plan. The work itself.

And the work has been getting worse at its job. When Peter Field last ran the numbers across the IPA's twenty-four-year databank of effectiveness case studies, creatively awarded campaigns were less effective than at any point in that record — and no more effective than campaigns that won nothing at all. Field's diagnosis was short-termism: the steady replacement of brand building with disposable tactical activation.

Put the two findings side by side. The variable that explains roughly half the outcome is the one the industry has been quietly disinvesting from, across precisely the decades it became most sophisticated at knowing where to put it.

And the market is moving in exactly the opposite direction. Retail media is the fastest-growing channel in advertising, which is to say the industry is scaling the one thing the data rewards least: communication one step from conversion, aimed at people already looking. The system is giving up on sowing and specialising in harvesting. Not because harvesting works better, but because harvesting produces a receipt and sowing doesn't.

None of this is an argument for spending less on media. It is the opposite. If creative accounts for roughly half the incremental return, then every dollar of media currently carries a passenger: a message built for nowhere in particular. The media isn't the waste, it is the part that still works. What we send through it is the part that stopped. Change what travels, and the same budget buys more.

AI arrived to fix the wrong scarcity

And then, at the exact moment when fragmentation made creative adaptation the central problem, a technology appeared that could finally solve it at scale. A thousand native versions for a thousand contexts. Creative that changes its grammar depending on whether it lands in a feed, a living room, a podcast, a creator's video, a second screen at 11pm. Adapting creative as a form of hyper-personalisation to the kind of relationship with the people watching their chosen content at a given moment. Adaptation as a default condition and as a relationship builder rather than a budget line.

Instead, AI is being used to make one ad cheaper. After four or five decades of mythologising the cost of creative — Madison Avenue, Mad Men, the rock-star copywriters of the eighties and nineties — the industry decided all of it had been a colossal waste of money. It is now busy disrupting that waste. And wasting the same money again, somewhere less visible.

The industry had a scarcity of adaptation. It used the most powerful adaptive technology ever built to solve a scarcity of production. Costs came down, output went up, and the variable that drives half the result stayed exactly where it was — one message, one grammar, now manufactured more efficiently. 

The money AI frees up was never meant to leave the system. It was meant to pay for the thousand versions.

I spent twenty years in branded entertainment, and one thing proved true every time: craft is not decoration. Work that lands is work built for a specific medium and a specific moment, and that fit is a skill, not a flourish. What is new is that it could finally scale. The cheap, do-it-yourself creative that AI has made free is not a democratisation of craft. It is its retirement, at the exact moment the technology could have industrialised it.

The creator economy was the answer, bought as a discount

The one genuine attempt at native creative came from outside the industry. Creators speak their duchy's language natively, because they are from there. For a brief moment it looked like the structural fix: stop shipping foreign messages into native contexts, and have natives make the message.

Then the industry bought it the way it buys everything, as inventory, priced per unit, optimised downward.

WARC calls what followed the creator effectiveness gap: a large share of creator investment wasted, returns far more volatile than any other channel, and effectiveness hinging almost entirely on brand-creator fit and the quality of the creative work. Which is another way of saying that the thing being purchased was never the audience. It was the adaptation and adaptation is the one thing you cannot buy by the impression.

The pattern is consistent enough to be a law. Every time fragmentation hands the industry a creative problem, the industry converts it into a procurement problem, and then wonders why the numbers don't move.

And yet the proof that native creative works is sitting in plain sight, in the cases where it is allowed to work. A creator introducing a product nobody has heard of, to people who already trust them, in the format that trust was built in, does something no media plan can buy: they supply the context the product doesn't have yet. The clearest proof is the creators who stopped selling access to their audience and launched their own brands instead. They have no media budget. What they have is a message that is native by construction.

That is the part of the feudal break worth defending. The monoculture made brand-building a privilege of whoever could afford national reach. Fragmentation made it available to anyone fluent in one duchy's language. Feudalism didn't only break the map, it democratised who gets to build a brand on it. The industry still reads that as a loss of reach. It is a collapse in the cost of meaning.

Owned attention requires owned creative

I've argued for a while that brands have to stop renting all of their attention and start owning some of it, building audiences instead of borrowing other people's. The feudal map forces a harder version of that argument, one level deeper.

Renting the format is as fatal as renting the audience.

A brand can buy perfectly native placement, inside a perfectly chosen bubble, from a perfectly matched creator, and still arrive as a foreign power  because the creative it ships is imperial. Built elsewhere, for everyone, in a grammar that belongs to no particular place. Targeting gets you through the gate. Only personalized adaptation gets you taken seriously once you're inside.

That is what owned creative means. Not producing more. Producing differently per territory, with the same seriousness that media planning applies to reach, and with the technology that now makes it trivially possible pointed at the right problem.

The map went feudal fifteen years ago. Advertising is still sending the same dispatch from the capital, in the capital's language, and reading the response rate as a measurement problem.

On contentwithaview.com you will find an “imperial test” useful to understand where your brand lies.

_____

Sources: IPA Databank / Peter Field, The Crisis in Creative Effectiveness; Circana; WARC, Future of Media 2026. "Feudal media" is Alan Wolk's framing.