Advertising has always been about persuasion. But for most of history, persuasion was an art form—expensive to produce, broadcast to everyone equally, and constrained by the limits of what a creative team could imagine. A good copywriter crafted a message. It reached as many people as the medium would allow. And the same message reached everyone. You and your neighbour saw the same television commercial. You disagreed or agreed on its merits on a shared basis.
Digital advertising dismantled all three of those constraints simultaneously—scale, uniformity, and legibility. And the result is something qualitatively different from anything that came before. Not just advertising made more efficient, but advertising made invisible, individualised, and psychologically calibrated in ways that most people have no framework to detect.
Traditional advertising was persuasion at the level of the market. Digital advertising is persuasion at the level of the individual—tuned to your specific psychology, delivered at the moment you're most susceptible, and invisible to everyone else.
The Science of Influence
Today, an advertiser doesn't create one ad. They create hundreds of variations—different headlines, images, emotional registers, calls to action—and deploy automated testing to determine which version persuades which psychological profile most efficiently. This isn't hypothetical. It's the default workflow on every major digital advertising platform.
AI generated and AI optimised ad campaigns have been shown to increase conversion rates by 30–40% compared to traditionally produced creative. The winning variants aren't chosen by humans exercising judgment. They're selected by machines optimising for conversion metrics, iterating across millions of impressions in real time. When I ran digital advertising operations at lastminute.com, we ran hundreds of simultaneous creative tests. We let the algorithm find the winners. The results were consistently better than what any single creative director would have chosen. The machine's indifference to aesthetics turned out to be an asset in conversion optimisation.
The effects accumulate in ways that should give us pause. Studies show that 38% of people have bought something they had no intention of buying because of targeted digital advertising. And 89% of digital purchases are influenced by remarketing—the practice of following you across the web after you've visited a page, nudging you back until you convert. It's like a very polite but absolutely relentless salesperson who never takes no for an answer.
Real Examples: What Persuasion at Scale Looks Like
Coca Cola's "Create Real Magic" campaign in 2023 showed one face of AI powered persuasion: generative, participatory, and genuinely creative. Coke partnered with OpenAI to let users create their own Coke-inspired artwork using AI tools. Thousands of pieces of user-generated content flooded social media. The brand didn't just reach its audience—it enlisted its audience as co-creators. The campaign racked up over 800 million impressions. The persuasion worked because it felt like play.
Heinz ran an equally clever experiment: they asked an AI image generator "What does a ketchup bottle look like?" The AI consistently produced something resembling a Heinz bottle. This became a global campaign—proof that even machines associate Heinz with ketchup. It went viral precisely because it felt like a genuine discovery rather than an advertisement. That's the sophistication of modern persuasion: it works best when it doesn't look like itself.
Microtargeting and the Fracturing of Shared Reality
The most consequential development in digital advertising isn't performance marketing or programmatic bidding. It's microtargeting—the ability to show entirely different messages to different people based on algorithmically inferred psychological profiles. Social media now influences 47% of online purchases, and among Gen Z, 30% discover new products primarily through influencers—people whose recommendations are themselves part of a paid, targeted system. The categories driving this most strongly: Beauty, Fashion, and Tech.
You and I might see completely different versions of the same political candidate, the same product, the same event—each version engineered to resonate with what the platform believes about each of us. We're not disagreeing about how to interpret shared facts. We're looking at different constructions of reality, each personalised to maximise our individual susceptibility. Democracy depends on the existence of a common information environment. Microtargeting quietly dissolves it—not through censorship or propaganda in the old sense, but through perfect individualisation.
When persuasion becomes invisible and individualised, democratic conversation becomes structurally impossible. You can't argue about what you're not allowed to see.
What Gets Lost—and What You Can Do
Something important disappears when persuasion is individualised at this scale: the possibility of a common information environment. Democracy—messy, imperfect, contested democracy—depends on the idea that citizens can argue about facts, that there is at least some shared substrate of information we're all responding to. Persuasion at scale doesn't just make advertising more effective. It quietly dismantles the conditions that make public deliberation coherent.
But this isn't a counsel of despair. It's a call for literacy. Understanding how the machine works changes how you interact with it. Before sharing an emotionally charged piece of content, ask: why does this make me feel this way, and who benefits from me feeling this way? Before a purchase decision, ask whether the urgency you feel is genuine or manufactured. The persuasion system is extraordinarily powerful. But it only works when its targets don't see it. Seeing it—really seeing it—is the beginning of genuine autonomy.
