Marketing Is No Longer About Persuasion, It's About Being Chosen by Algorithms?

Not long ago, marketing was largely about persuasion. A brand's job was to convince people — that its product was better, its service was faster, its offer more valuable than the alternatives. That challenge hasn't disappeared. But is it still where marketing starts? Or does something else have to happen first — does a brand have to be chosen by an algorithm before it ever gets the chance to persuade anyone at all?

Think about how we discover things today. Products appear in social media feeds. Videos are recommended by platforms. Search engines decide which results we see first. Online marketplaces determine what gets featured. AI assistants increasingly decide which answers, brands, and recommendations reach us at all. In most of these cases, consumers don't see everything that's available — they see what a system has already selected for them. This is sometimes called algorithmic curation: the selective surfacing of content based on inferred interest, running quietly underneath products we experience as simple feeds or search boxes. That may sound like a subtle distinction. But if visibility itself is being filtered, ranked, and distributed by machines before a human ever sees it — is persuasion still where the marketing battle begins? Or has something upstream of persuasion quietly become the real contest?

For most of marketing history, the central challenge was getting into people's minds. That still matters. But increasingly, brands need to get into feeds, recommendation engines, search results, and AI-generated answers first. You can see the effect almost everywhere online: articles are written with search rankings in mind, videos are structured around platform preferences, brands chase trending keywords, engagement signals, and watch time. Sometimes it feels as though the first audience for a piece of content isn't the customer at all. It's the system deciding whether the customer ever sees it.

Marketers have always adapted to gatekeepers, of course. Newspapers decided which advertisements appeared on their pages. Television networks controlled airtime. Radio stations determined what listeners heard. So is this really new, or just the same old gatekeeping wearing a different face? There's a case for "same old": brands have always needed someone else's permission to be seen. But there are a few things about this gatekeeper that feel different. Could you ask an editor why a story ran? Usually, yes. Can you ask an algorithm the same thing, and expect a real answer? One day a post reaches thousands of people; the next day something almost identical disappears into digital silence, with no explanation available even to the advertiser paying for it. Is that a difference in degree, or a difference in kind?

The result is an odd version of marketing where visibility itself becomes a competitive advantage almost independent of the message. A brilliant campaign that nobody sees may end up less effective than an average campaign an algorithm happens to favour. That sounds unfair until you remember that discoverability has always mattered to some degree — the difference is that the gatekeeper used to be a person you could, in theory, identify and appeal to. Now it lives inside code that even its own engineers don't fully control end to end.

It's worth being careful here, though, about how separate "the algorithm" really is from human judgment. These systems are trained on the aggregate behaviour of real people — what they click, watch, and linger on — so in one sense an algorithm is a compressed, automated version of the crowd's own past preferences. But researchers studying recommendation systems have also found that this relationship runs in both directions: these systems don't just reflect what people already want, they can also shape what people come to want, by narrowing the range of what's easy to discover in the first place. So which is it? Is the algorithm just a mirror of the crowd, or is it quietly training the crowd to want what's easiest for it to serve? Maybe the honest answer is that it's both, feeding into each other in a loop nobody fully designed. What would it mean for a marketer to be persuading not a person, but one half of that loop?

Sometimes I wonder how many products, businesses, or ideas I've never encountered simply because they failed an algorithmic test I didn't know existed — not because they were bad, and not because they lacked creativity, but because they weren't selected. In that sense, consumers are often no longer choosing from the entire market. They're choosing from a version of the market that has already been curated for them. And if that's true, what does persuasion even look like in a world where selection happens first?

The rise of generative AI makes the question stranger still. Imagine searching for a product a few years from now: instead of scrolling through pages of results, you ask an AI assistant for a recommendation, and it responds with two or three options, maybe with reasons attached, never mentioning the hundreds of alternatives that also existed. At that point, who is the marketer really trying to persuade — the customer, or the system generating the recommendation on the customer's behalf? This isn't a purely hypothetical worry. Industry research from Forrester has already found that among business buyers, generative AI tools are becoming one of the most common ways people research a purchase, ahead of vendor websites and even peer recommendations. A new discipline — sometimes called generative engine optimization — has sprung up specifically to help brands get mentioned inside AI-generated answers rather than just ranked in a list of links. Is that a durable new field of marketing, or a gold-rush industry that will look quaint in five years, the way some early SEO tactics do now? Nobody writing about it today really knows. But the question underneath it doesn't depend on the answer: if a system can hand a customer three options and never mention the other three hundred, what exactly is a brand supposed to be optimizing for anymore?

It's a slightly unsettling thought, though not because algorithms are necessarily bad. Most of us appreciate useful recommendations, nobody wants to sort through ten million options every time they need a phone charger or a hotel room. The tension is more philosophical than practical. Marketing has traditionally been understood as a conversation between brands and people: brands make claims, consumers evaluate them, choices get made. Increasingly, there's a third participant in that conversation, one that filters, ranks, recommends, and suppresses before the human conversation even begins. And unlike consumers, algorithms don't respond to humour, emotion, or clever copywriting directly. They respond to signals, patterns, and predictions, which is a different kind of audience to write for entirely.

Which means marketers now find themselves persuading in two directions at once. One message for humans. Another, quieter one for the systems deciding whether humans will ever see it. Is marketing still about persuasion, then, or has persuasion quietly become the second act rather than the first? Before a customer can choose a brand, a system increasingly decides whether that brand gets considered at all — so is winning attention still the hard part, or is the real challenge now earning the chance to compete for it in the first place? I don't think there's a clean answer yet. What do you think, is this a problem marketing already knows how to solve, just wearing new clothes, or is it something genuinely different that the field hasn't caught up to?