Mitch Stoller is Chairman of Literate AI and co-founder of Group SJR, the first pure-play content marketing agency (acquired by WPP).
The brands that won the last twenty years of digital marketing didn’t win because they were smarter.
They won because they moved first. They built expertise in a new platform before the market understood what that platform was. They captured audience, built data advantages, and compounded those advantages while their competitors were still reading think pieces about whether the shift was real.
This has happened twice. The pattern is identical. And it is happening again right now.
Era One: The Intent Machine (2006–2014)
When I sat in early Googler offices in San Francisco in 2006, managing $150,000 to $200,000 in daily paid search spend, the thing that struck me wasn’t the scale. It was the signal.
A person typing ‘buy running shoes’ into a search box was doing something unprecedented in the history of advertising: they were raising their hand in public and announcing a need. You didn’t have to interrupt them. You didn’t have to persuade them they had a problem. The intent was declared. You just had to be there.
For a decade, the marketers who understood this signal earliest built extraordinary businesses. CACs that seemed impossible to anyone still buying print ads or TV spots. Attribution that was clean, measurable, and direct. Marketing that actually worked.
And then, slowly, it stopped working. Google matured its quality score system to favor legacy brands. Auction prices rose. Startups found themselves paying a premium CPC just to compete against household names with structural advantages baked into the algorithm. Venture capitalists started calling Google a margin-killing addiction.
Era One didn’t end with a disruption. It ended with a squeeze.
Era Two: The Audience Machine (2015–2022)
By 2015, Meta arrived with a fundamentally different model. Not intent marketing. Audience marketing. Instead of waiting for a user to declare a need, Meta lets you find the exact demographic pockets most likely to have that need — and reach them before they even knew they were looking.
I managed food-space portfolios spending $3 to $4 million a month in those years. On Google, our customer acquisition cost hovered around a brutal $40 to $50. On Meta, with lower CPCs and aggressive creative testing, our CAC fell to $5 to $10. We were suddenly profitable on the very first order.
Multiple brands in our portfolio went from absolute zero to $400 million in annual revenue on the back of this model. For nearly a decade, the creative-first, audience-based playbook was the best-kept secret in marketing. Until it wasn’t a secret anymore.
In April 2021, Apple deployed its App Tracking Transparency framework with iOS 14.5. Overnight, Meta’s ability to track users across third-party apps disappeared. The attribution engine went dark. CACs spiked. Margins collapsed. DTC businesses built entirely on the Meta backbone went under.
TikTok filled some of the void. Snapchat filled almost none of it. Brands were left trapped in a high-priced duopoly, paying more for less, desperate for a third channel that could actually perform.
Era Two didn’t end with a disruption. It ended with a policy update.
Era Three: The Intelligence Machine (Now)
The third era doesn’t have a single launch date. It has been accumulating for the past three years as consumer behavior quietly shifted.
Consumers stopped searching and started asking. They moved from search boxes to conversations. From ranked lists of links to synthesized, contextual answers. From Google’s blue links and Meta’s feed to ChatGPT, Perplexity, Gemini, Grok, and Copilot.
And here is the thing that should stop every marketer reading this: the early campaigns running inside LLM platforms feel exactly like Era One and Era Two felt in their first years. Low competition. High contextual relevance. CACs that look impossible to anyone still anchored to the old metrics.
I’ve sat across from the data. The deja vu is real.
The Pattern Is the Point
Every era of digital marketing follows the same arc. A new platform creates a new signal — intent, audience, intelligence. The marketers who decode that signal earliest build compounding advantages. The platform matures, prices rise, the house restructures the game. Then a new platform emerges with a better signal.
The brands that spent 2004 optimizing Yellow Pages while their competitors were buying AdWords spent the next decade playing catch-up.
The brands that spent 2015 doubling down on Google while their competitors were building Meta audiences spent the next decade playing catch-up.
The brands that spend 2026 optimizing the old duopoly while their competitors are building AI Marketing capability will spend the next decade playing catch-up.
The pattern doesn’t change. Only the platform does.
Next week: Jay Bhatti on what AI Marketing actually is — and what most agencies are getting dangerously wrong. seoisdead.com