Clickbait isn’t what it used to be. That might sound odd given how much of it still floods our feeds, but the techniques have evolved significantly from those early “One Weird Trick” ads that cluttered the internet in the early 2010s.

Understanding how we got here helps explain where we’re going. And where we’re going isn’t necessarily better.

The First Wave: Curiosity Gaps

The original clickbait formula was simple: create a gap between what you know and what you want to know, then promise to fill it. “You Won’t Believe What This Celebrity Looks Like Now” works because it creates artificial curiosity about something you didn’t care about seconds ago.

BuzzFeed and Upworthy perfected this in the early 2010s. They A/B tested headlines obsessively, figuring out exactly which phrasings maximised clicks. “This Is What Happens When…” and “The Reason Why…” became templates because they worked.

The backlash came quickly. By 2014, people were sick of it. Facebook changed its algorithm to penalise obvious clickbait. The low-hanging fruit had been picked clean.

The Second Wave: Emotional Manipulation

When pure curiosity gaps stopped working as well, clickbait evolved. The next generation weaponised emotion instead of curiosity. Outrage, fear, and moral superiority became the primary drivers.

Headlines like “This Company Is Destroying the Planet and Nobody’s Talking About It” or “Everyone’s Getting This Important Issue Wrong” hit different emotional buttons. They weren’t promising unknown information; they were promising confirmation of what you already felt.

This version is more insidious because it feels more legitimate. There’s often a real story underneath, even if the headline oversells it. You’re not being tricked into reading about celebrity weight loss—you’re being manipulated into clicking on a story that validates your existing worldview.

The Algorithm Generation

The current wave of clickbait is shaped by platforms, not publishers. TikTok, YouTube, and Instagram algorithmically promote content that keeps people watching. The clickbait isn’t in the headline anymore—it’s in the content structure itself.

Videos now open with “Wait for it…” or “You have to see what happens at the end.” Tweets are written as threads that require clicking through for the payoff. Instagram posts hide the punchline in the caption continuation. The platforms reward this behaviour, so creators adapt.

I’ve watched this happen with technology coverage. Team400 has written about how recommendation algorithms change content creation incentives, and it’s playing out exactly as predicted. Content gets shaped by what the algorithm rewards, not what serves readers best.

The Trust Erosion

Here’s what makes this evolution concerning: each wave has eroded trust a bit more. The first wave taught people to distrust headlines. The second wave taught them to distrust news sources. The current wave is teaching them to distrust content itself.

When you can’t tell whether a video is building to something meaningful or just wasting your time for engagement metrics, you start approaching all content with suspicion. That’s not a healthy relationship between creators and audiences.

The metrics don’t capture this erosion. A platform can show growing engagement numbers while trust steadily declines. You only notice when people start abandoning the platform entirely.

What’s Next

The next evolution is already happening, and it’s more subtle than what came before. AI-generated content is enabling personalised clickbait at scale. Instead of one headline tested on everyone, systems can generate variations targeted at different audience segments.

You might see “Why Remote Work Is Failing” while your colleague sees “The Remote Work Revolution Is Just Beginning”—both leading to the same article that argues a nuanced middle position. The headline isn’t lying, exactly, but it’s been optimised for your particular biases.

This is more sophisticated than previous generations. It’s harder to spot, harder to criticise, and harder to resist. You can’t call it out as obviously manipulative because it’s not obvious—it’s tailored.

The Economics Won’t Change

None of this will stop unless the underlying economics change. Publishers need clicks to survive. Platforms need engagement to sell ads. Creators need views to make rent. As long as those incentives exist, someone will figure out how to game them.

The question isn’t whether clickbait will exist—it’s whether we’ll build systems that reward quality alongside engagement. Some platforms are trying. Medium’s reading time metrics, Substack’s subscription model, and Patreon’s direct support all try to align creator incentives with audience value rather than just attention.

But they’re fighting against much larger forces. The free, ad-supported internet is built on attention extraction. You can’t fix clickbait without fixing that fundamental model.

Reader Responsibility

There’s also the uncomfortable question of reader responsibility. Clickbait works because we click on it. We know “This Simple Diet Trick” probably isn’t going to change our lives, but we click anyway. We recognise outrage bait, but we engage with it anyway.

Part of the solution has to be readers getting better at resisting these techniques. That doesn’t mean blaming victims of manipulation—these tactics are designed by smart people with lots of data. But it does mean developing some immunity to obvious manipulation.

The headlines that make you feel strongly before you’ve read anything should be red flags. The content that promises emotional satisfaction should be approached skeptically. If it seems designed to make you angry or superior or afraid, ask why.

The Cycle Continues

Clickbait will keep evolving because the people creating it are smart and motivated. Every time we develop resistance to one technique, they’ll develop a new one. Every time platforms crack down on one approach, creators will find another that technically complies while achieving the same goal.

This isn’t necessarily pessimistic. Evolution means adaptation, and adaptation can go in better directions if we build the right incentives. But it requires actually changing those incentives, not just complaining about the symptoms.

Until then, expect more of the same. Just more sophisticated, more targeted, and harder to spot. That’s progress, of a sort.