Off Kilter 203: Plumbing Not Prophets.

tl;dr: We need to stop treating Big Tech rhetoric as truth.

The earth’s most valuable commodity isn’t gold, oil, rare earth metals, data, or Bitcoin—it’s human attention. And Big Tech is its gatekeeper.

The so-called “Magnificent Seven”—Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta, and Tesla—are now worth more than entire national stock markets. Their aggregate value floats around $17 trillion, which is more than the entire GDP of the European Union. Their dominance is built not just on software and silicon, but on the algorithmic intermediation of human attention at a global scale. And yet, marketing—the corporate function that pays Big Tech for access to this attention—has become little more than a bystander. This needs to change.

This is a story with two sides. On one side, it’s a story of economic and regulatory capture: how the false promises of algorithmic oracles seduced an entire profession with tall tales of efficiency as salvation. On the other hand, it’s a story of how marketing itself made two foundational errors that opened the door to this reality, and why the solution isn't just walking away from Big Tech (plot spoiler, you can’t)—it's about treating them as plumbing on a journey toward increased business performance.

But first, a question: We all know that marketing today is working under extreme efficiency constraints, but why such an obsession with efficiency? It’s not as simple as recent economic uncertainty. While that doesn’t make anything easier, marketing budgets have been cut by one-third over the past four years.

Here’s a simple perspective. If marketing were delivering the business outcomes it's supposed to deliver, efficiency would be secondary. Therefore, if CFOs are demanding even more efficiency, it’s probably because marketing isn’t meeting corporate expectations for effectiveness. When the function tasked with growing the business fails to do so, cost reduction becomes the inevitable booby prize. And that's precisely where we seem to be.

As a result, marketing's efficiency crisis isn't a strategic choice. No CMO chooses to slash budgets by a third—it's a tacit admission of defeat.

How we got here isn’t just that Big Tech ate the marketing functions’ lunch. Nothing so prosaic. In actuality, it bought the restaurant, transformed the menu, and then started charging its diners rent. Take advertising as a single example. Approximately 50-60% of all digital advertising globally flows through just two platforms: Google and Meta. This means they aren’t just hosting the market; they are the market. Their standards have become the profession's standards. Their language—impressions, clicks, conversions, ROAS—has become the lingua franca of the industry these two firms have so solidly captured: Cannes revolves around their parties. The trade media subsists upon their sponsorships. Conferences sow their narratives. The metrics marketers are told to worship were largely invented by them. The outcome? Marketing's impact and influence have been slowly minimized and colonized, while its budgets are systematically extracted to be converted into wealth for Google and Meta shareholders.

Google has already been ruled an illegal monopolist that artificially raises the cost of advertising by U.S. courts. For proof, its US search advertising costs are rising faster than any other advertising medium.

Meta's behavior—tying products, copying competitors, squeezing publishers—would likely meet the same threshold. However, legal status is almost beside the point, especially when considering how soundly antitrust enforcement has been weakened in recent decades or the glacial pace at which court-ordered remediation occurs.

No, what matters is functional power. These platforms don't just connect buyers and sellers; they extract margin at every turn. They take a cut of creation, distribution, targeting, measurement, and optimization. Their algorithms optimize for their own revenue, not yours. And if you want to reach customers, there's often no alternative.

Digital advertising is no longer a market. It's an ever-more-expensive toll road.

Here's why all this matters as we enter the cusp of the AI era: the same platforms that disrupted marketing by promising perfect efficiency, but then raised prices once they became dominant, are now selling AI-driven efficiency to address the ongoing inefficiency of marketing. First, they convinced marketers to build expensive, complex tech stacks in pursuit of algorithmic optimization and programmatic precision across their platforms, which systematically eroded brand strength and long-term effectiveness. Now, as their extractive behavior forces marketing costs to rise faster than the growth delivered, these same platforms are offering AI-driven efficiency as the solution. But the solution to what? Yup, you guessed it, to the problems they themselves created.

Make no mistake, this is a Catch-22 filled with chutzpah designed by and for the benefit of Big Tech and nobody else.

This didn't happen by accident. Big Tech simply exploited the landscape as they found it. In truth, it was marketers who built their own trap by making two critical errors that handed Big Tech the keys to the kingdom:

The first was a measurement error. In an attempt to be taken seriously as an investment rather than a cost, marketing adopted ROI as its primary financial metric. But ROI is a financial concept that assumes a linear relationship between input and output, over a defined timeframe. Marketing doesn’t have those same qualities. Its most powerful effects are often emergent, compounding, and culturally embedded. The moment marketing agreed to be judged by quarterly returns, it forfeited its long-term strategic voice. As a result, it became vulnerable to any system that promised measurable, immediate results, no matter how fantastical those promises ultimately turned out to be.

The second was a category error. Under the false promises of customer surveillance, marketers began treating markets like closed, predictable systems—machines to be optimized—rather than open, complex systems—ecosystems to be nourished.

In truth, customer markets are ecosystems. They’re often unpredictable, adaptive, shaped by social forces, stories, feedback loops, and context. They cannot be optimized in the way a bidding algorithm can. But tech platforms sold the illusion that they could. And marketers, being promised control over factors that had historically been chaotic, took the bait.

So, how did tech convince an entire profession to worship at these algorithmic altars?

Because it dressed itself in the clothes of magic.

The disruption of marketing wasn't about software or data; it was a new truth. Digital Natives. Machine learning. First-party data. Predictive analytics. Real-time bidding. Performance Marketing. Product Marketing. Account-Based Marketing. Growth Hacking. The promise wasn't just to increase reach or target more precisely—it was a whole new reality of how people consume, presented in dashboard form. Click here, spend there, get this many conversions. Never mind that the metrics were largely self-reported by the platforms themselves. Or that most of the promises being made—like last-click attribution—were a lie. Or that the platforms won’t allow you to run an independent audit. Or that fraudulent impressions are a feature rather than a bug. Marketers, while pursuing what they believed to be rational value-maximizing behaviors, instead became useful idiots for tech firms hellbent on industry domination. At some point, they stopped trying to hold the machine accountable and instead started thanking it for the scraps.

This is where we are today: a profession designed to create mutual value exchange between corporation and customer, to drive preference, protect margins, and grow demand has instead become a high-speed execution layer for tools it doesn't own, chasing KPIs it didn't define, in a system it no longer has visibility into. The result isn't just dependency—ongoing budget declines suggest it is decay.

Optimization flattens the differences between brands. Automation accelerates sameness. Algorithms reward convergence. The tools that promised to make marketers better at their jobs have instead made them more interchangeable and replaceable.

Now, to fulfil the promise of replacing marketers comes AI, wrapped in the exact same efficiency promise that failed to pan out the last time tech firms made it.

McKinsey says 20% of marketing tasks could be automated. But don’t be suckered by the headline statistic. 20% task automation doesn’t equal 20% cost-out. Here's the math that should prompt every executive to pause: a 20% task reduction in a function where total costs (including areas where there's no task reduction benefit) account for roughly 7.7% of revenue, equals savings of roughly 0.4% of total revenue as a best-case scenario. More likely, the savings would be in the 0.2-0.3% range, and that doesn’t include any switching costs, teething troubles, or areas where the technology may or may not perform optimally.

For a growth function, such an efficiency saving isn’t a transformation. It’s a rearranging of the deck chairs. Meanwhile, ironically, the platforms offering you this AI efficiency are the same platforms that are actively inflating your costs as we speak. They're not handing you leverage. More likely, they're handing you a leash.

Here's what the efficiency narrative misses entirely: the real AI opportunity isn't 0.4% cost savings—it's using AI to do better marketing that drives superior business outcomes.

Unfortunately, the Big Tech firms have little interest in helping their clients do better marketing. Why? Well, that’s simple. Being better at marketing would give these clients increased self-determination, and that’s the last thing Big Tech wants. To drive ever-increasing share price growth, what Big Tech needs is dependence rather than self-determination.

This is why I often joke that marketing competence is inversely proportional to your distance from Silicon Valley. There’s a reason the Ehrenberg Bass Institute is in Australia, folks.

Anyway, if you are a marketer looking at AI today, the question shouldn't be how to automate marketing tasks that are already experiencing diminishing returns. Instead, it should be about how you utilize this revolutionary new technology to gain a deeper understanding of your customers, increase their desire, boost margins, build stronger brands, and create more meaningful and memorable experiences for people.

Getting there will require marketers to have their own distinct strategies, hypotheses, and goals, and to stop treating Big Tech as prophets and start treating them as plumbing: Infrastructure and pipes. Useful for most, essential for many, they aren’t prophets because the only future they see is the one that works best for them, not their clients.

This distinction—treating tech as utilities rather than oracles—should become the organizing principle for every marketing decision going forward. Marketers need to stop treating the rhetoric of technology platforms as gospel and start treating them like what they are. Plumbing.

AI can be part of the solution rather than the problem—but only when we view it as part of a more integrated system. In complicated systems like media planning or asset resizing, AI is a good optimizer. In complex systems, such as customer behavior and cultural context, AI must play a different role. It should not be treated as a control system, but as a sensemaking tool. A pattern detector. A probe for insight. Not a driver, but a lens. Used effectively, the pattern recognition and unstructured data capabilities of Generative AI can help us identify emerging behaviors, test new ideas, surface cultural tensions, and expand our understanding of our customers.

To build effective brands using this plumbing-not-prophets framework, marketers will have to relearn how to use the tools of tech without being used by them. This means understanding not only how a tool works, but also why it works the way it does. It means embracing old-fashioned ideas, such as distinctive value propositions and compelling experiences. It means investing in memory, not just immediacy. It means thinking in years, not just quarters. It means being different, not just efficient. Treating platform data as a signal, rather than as scripture. Resisting the urge to outsource strategy to code. And above all, refusing to get sucked into the exact same magical, fantastical thinking that created the mess we already find ourselves in.

It’s time to stop optimizing for games that were rigged from the start. It's time to remember what marketing is for: mutual value exchange, creating desire, protecting margins, and building value over time. The job of marketing isn’t to become more efficient or be a patsy for Big Tech. The job of marketing is to drive better business outcomes by mattering more to more people, more often.

This requires imagination, not just automation. Judgment, not just logic. And above all, the courage to reject the magical thinking of the Big Tech landlords that no longer serve us, no matter how shiny the interface, or how overwhelming the rhetoric.

I’m going to be bold and state that marketers don't need AI to make them faster or more efficient at what they’re already doing, because what they’re already doing doesn’t seem to be working very well.

They don’t need to increase velocity; they need to change trajectory.

What marketers need is AI that improves marketing’s contribution to business performance.

And for that, we’ll need to look beyond Big Tech.

---

---