What if digital marketing was never a real profession?

Posted on: 3 December 2025

I'm taking my cue from a LinkedIn post by Luciano Giuliani, which references Mark Zuckerberg's intention for Meta to replace marketing agencies. In an interview given in May 2025 to Ben Thompson of Stratechery, Zuckerberg said:

"You're a business, you come to us, you tell us what your objective is, you connect to your bank account, you don't need any creative, you don't need any targeting demographic, you don't need any measurement, except to be able to read the results that we spit out."

AI-generated creativity, automated targeting, continuous optimisation. Connect your bank account, state your objective, and the machine does the rest. "You don't need any creative, you don't need any targeting demographic, you don't need any measurement." His words, not mine.

Industry reactions oscillate between panic and indignation. One agency CEO called the prospect "terrifying". Others speak of brand safety, the loss of human touch, the devaluation of creativity.

But perhaps they're looking in the wrong direction. Perhaps the question isn't whether Zuckerberg is right or wrong. The question is: what exactly were all these people doing who now fear being replaced?

I can afford to ask this question because I was there. I was there before digital marketing existed, when working in communications meant knowing how to communicate. I was there when the internet arrived and everyone said it changed everything. I was there when social media arrived and everyone said it changed everything. I'm here now as AI arrives and everyone says it changes everything. And I've watched, each time, the exact same pattern: those who knew how to do real things adapted, those who only knew how to ride the wave of the moment drowned in the next one.

Over the past twenty years, digital marketing has generated an ecosystem of specialisations that defies all organisational logic. Retargeting specialists, funnel optimisation consultants, bid management experts, attribution modelling analysts, programmatic buying strategists, tag implementation technicians, audience segmentation managers.

Each with their own minuscule territory. Each with their own proprietary terminology. Each with their own monthly invoice to submit.

To an observer who was there before, this stratification appears for what it is: an artificial multiplication of roles made possible by two converging factors.

The first: the genuine complexity of digital platforms in their early years. Clunky interfaces, sparse documentation, constant changes. Understanding how Facebook Ads worked in 2012 genuinely required time and dedication. There was a real informational advantage for those who specialised.

The second: the extraordinary margins that digital advertising generated when the cost of attention was still low. When a click costs a few pence and converts at three per cent, there's economic space to pay three, four, five intermediaries along the chain. Nobody looks too closely when everyone's making money.

But the informational advantage eroded as platforms became more intuitive. And margins compressed as more advertisers entered the market. What remained is a cost structure built for an era that no longer exists. And an army of people telling themselves they have skills when they only have familiarity with interfaces.

The problem with hyperspecialisation isn't just economic. It's cognitive. And it's devastating.

When you fragment a trade into twenty sub-specialisations, you get twenty people who know how to do one twentieth of the job and none who can do the whole thing. Everyone knows their segment perfectly. No one has the overview necessary to understand whether what they're collectively doing makes any sense.

I've seen this pattern in other sectors, in other eras. It happens whenever a system becomes complex enough to permit the proliferation of coordination roles, supervision, sub-process optimisation. At some point you find yourself with more people managing the process than you need to do the actual work. And everyone convinces themselves they're indispensable because no one remembers what it was like when the job was done by one person who actually knew how to do it.

In digital marketing, this phenomenon has reached farcical levels. Teams of fifteen people to manage campaigns that one competent person could handle in half a day. But that competent person, the one who can do the whole job, has become increasingly rare, precisely because the system incentivised fragmentation.

Why learn to do everything when you can specialise in a niche, get paid as an expert in that niche, and let someone else worry about the rest?

The paradoxical result: more "experts" in the system, less actual capacity to produce something that works. More hours billed, less value generated. More procedural complexity, less strategic clarity. An entire sector that confused activity for results, process for product, presence for contribution.

Now AI compresses these twenty roles into an algorithm. Not because AI is magically intelligent, but because those twenty roles should never have been twenty separate roles. It was artificially diluted work. And diluted work, when a tool arrives that does the same thing without dilution, evaporates.

Bid optimisation? Automatable. Audience segmentation? Automatable. Creative variant testing? Automatable. Performance analysis? Automatable. Everything that could be reduced to process, and that had been artificially expanded into careers, returns to being process.

And suddenly the question that had been buried for twenty years under layers of procedural complexity emerges: but you, concretely, what can you actually create?

Not "which segment of the process do you oversee". Not "which platform do you know better than others". Not "how many acronyms can you use in a presentation". But: are you capable of starting from a blank page and producing something a human being would find interesting?

It's the question David Ogilvy, Leo Burnett and Bill Bernbach asked before tracking pixels existed. It's the question anyone working in communications asked when "working in communications" meant knowing how to communicate.

Why should anyone dedicate even three seconds of their attention to this message?

It's not a technical question. It doesn't require certifications, doesn't require familiarity with interfaces, doesn't require continuous updates on the latest platform features. It requires something rarer and less teachable: the ability to understand what interests people and articulate it effectively. Something you don't learn on a course, don't certify with a badge, don't measure with a KPI.

At this point, someone will already be thinking about the obvious way out. If technical execution gets automated, just move up the value chain. Become strategic consultants. Do what the Big Four do, on a smaller scale.

It's the instinctive repositioning, almost obligatory. And it's a dead end for most agencies.

To do strategic consulting requires exactly that integrated competence that twenty years of hyperspecialisation has systematically eroded. It requires overview, understanding of the client's business, capacity to think beyond the next quarter. It requires knowing how to do the whole job, not one twentieth of it.

Someone who's spent ten years optimising bids or segmenting audiences hasn't developed those competencies. Not their fault, but because the system didn't require or reward them. The system rewarded efficiency in the fragment, not understanding of the whole. And now those people discover that the fragment is worth nothing and they never learned the whole.

So you end up with agencies wanting to sell strategy but with teams trained for tactical execution. And on top of that, they find themselves competing with McKinsey, Accenture, Deloitte, who've been in that territory for decades, have the brands, the relationships at board level, the track records with CEOs.

It's the classic squeeze: you can't do execution anymore because AI does it at marginal cost, you can't do strategy because you don't have the competencies and you're competing with enormously better-equipped players. You're trapped in territory that's shrinking from both sides.

Some will try to differentiate on service. "Right, execution gets automated, but we have the client relationship, the understanding of their business, the human touch". The problem is everyone says that. It's a point of parity, not differentiation. And if your value is "I understand the client", why should the client pay you instead of hiring someone internal who understands them even better?

What happened to travel agencies when Booking arrived? Some survived by specialising in complex travel: corporate, personalised luxury, exotic destinations where real expertise is needed. Most disappeared. There wasn't a repositioning possible for everyone. And no one today mourns the travel agencies that sold airline tickets with a markup.

The same logic applies here. Not all digital marketing agencies will become something else. Many will simply close, or be absorbed, or become empty shells reselling AI services with an ever-thinner markup. Until even that markup no longer makes sense.

Viable paths exist, but they're narrow and require prerequisites most agencies don't have.

The first is genuine verticalisation, not just declared. Not "we specialise in B2B" but "we're the only ones who understand how the sales cycle works in the industrial valve sector". Sector knowledge so deep that the client can't replicate it internally and AI can't replace it because the training data doesn't exist. But this requires years of genuine immersion in a sector, not a website rebrand with new keywords.

The second is returning to being artisan workshops. A few very good people doing quality work for a few clients willing to pay the premium. The model of pre-digital creative agencies, where the founder was the talent and the work went out with their signature. It requires having real talent, which is exactly what hyperspecialisation stopped cultivating.

The third is genuinely putting skin in the game. Instead of selling hours, tie compensation to results. Revenue sharing, equity stakes, substantial success fees. If you're really good, you earn more. If you're not, the market tells you immediately. The problem is this exposes you to risk, and most agencies have structured themselves precisely to avoid risk. They sold certainty to clients while offloading uncertainty onto results. Now uncertainty comes knocking at their door.

All three paths have one element in common: they require real competence, not procedural competence. They require knowing how to do something that has value in itself, not knowing how to manage a piece of a process someone else designed.

The irony is that "going backwards" in this case means going forwards.

Back to when people knew how to do whole things instead of fractions of things. Back to when value lay in creation, not in managing complexity that others had created. Back to when an advertising person was someone who could write, think visually, understand consumer psychology. Not someone who knew how to set campaign parameters.

It's not boomer nostalgia. It's the facts. The digital agency economy was built on an arbitrage: high technical complexity, margins sufficient to sustain the cost structure. Both factors are reducing simultaneously. AI lowers complexity, competition compresses margins. The model doesn't hold. It's not an opinion, it's the numbers. And numbers don't get emotional.

Those who see apocalypse in this scenario probably never had a profession in the full sense of the term. They had a role in a system that permitted that role. It's not the same thing. And the difference is about to become painfully clear.

AI eliminates alibis. It doesn't eliminate talent.

Zuckerberg thinks he's invented something new. He thinks total advertising automation is the future. In reality, he's accelerating a reset that was inevitable, a return to a truth that digital marketing had temporarily obscured.

Advertising has never been, in its best moments, a matter of technical optimisation. It has always been a matter of human understanding. Understanding what people want, what frightens them, what makes them laugh, what convinces them. Then finding the right words and images to say it.

Everything else, the funnels, the pixels, the segments, the automations, was infrastructure. Useful, sometimes necessary, but always secondary to the fundamental question: have you got something interesting to say?

For twenty years, infrastructure ate substance. It generated careers, agencies, entire industries dedicated to managing processes that should have been ancillary. Now infrastructure is being automated, and substance returns to the foreground.

It's not the death of advertising. It's the end of its bureaucratic phase.

And you reading this, what will become of you?

No, really. Stop for a second. What's the thing you can do that an algorithm can't do better, faster, at zero cost? What's the reason a client should pay you tomorrow, when today they pay you only because they didn't have better alternatives?

If the answer is "I know the platforms", you have a problem. If the answer is "I have client relationships", you have a bigger problem, because those relationships are based on value that's evaporating. If the answer is "I'll reinvent myself in strategy", ask yourself honestly: when was the last time you thought strategically instead of executing tactics dressed up as strategy?

You can keep telling yourself that AI doesn't understand nuances, that the human touch is irreplaceable, that clients will always want someone to talk to. You can keep attending webinars on "how to integrate AI into your workflow" as if the problem were adding a tool instead of rethinking everything from scratch.

Or you can look at the facts. The model that's paid your salary for ten years won't exist in five. Perhaps three.

The question isn't whether to change. It's whether you still have time to do it, and above all whether you have something to rebuild on.

Because if beneath the technical skills there's nothing else, no repositioning will hold. And the only person who can answer that question is you, now, before the market answers for you.