The pattern runs like this.
A new platform emerges. It changes the economics of attention. Publishers and brands rush toward whatever the new metric is — clicks, time-on-site, video views, engagement, shares. Editorial judgment gets subordinated to the metric. The content degrades. The audience leaves. Everyone agrees, afterward, that chasing the metric was a mistake.
Then the next platform emerges.
This has happened, with remarkable consistency, at every major inflection point in the history of online media. Understanding the history is not nostalgia. It is the most useful thing a content strategist can study, because the pattern has not changed and will not change, and the organizations that survive platform shifts are not the ones that chased the right metric — they’re the ones that maintained editorial judgment when everyone else was abandoning it.
The content farm era.
The first iteration was pure SEO optimization. The insight was correct: Google ranked pages, pages got traffic, traffic could be monetized. The implementation was the problem. If Google ranked pages with keywords, you needed pages with keywords. If long-tail queries drove traffic, you needed thousands of pages targeting long-tail queries. Quality was irrelevant because quality wasn’t being measured — keyword density was.
Demand Media built a business model on this. At its peak, eHow was producing 4,000 articles per day. The articles were written for search engines, not humans. They answered queries in the most technically compliant way possible while providing approximately zero value to the person asking.
Google eventually adjusted its algorithm. Demand Media’s stock collapsed. The lesson everyone took was: don’t do keyword stuffing. The lesson they should have taken was: content produced to satisfy an algorithmic metric rather than a human need will always be bad content, and eventually the algorithm will figure this out.
The pivot to video.
In 2016, Facebook announced that video was the future of its platform. The engagement numbers they shared were extraordinary — video, they said, received far more engagement than text. Publishers, already struggling with declining display ad revenue, pivoted.
BuzzFeed laid off writers and hired video producers. Mic pivoted to video entirely. Mashable was acquired at a fraction of its previous valuation after pivoting to video. Vice, which had built its editorial identity on first-person documentary journalism, refocused around Facebook video. Dozens of mid-sized digital publishers cut their editorial staffs and redirected resources toward video production.
The Facebook numbers were wrong.
In 2019, Facebook settled a lawsuit with advertisers who had been misled by the inflated video metrics. The company had been calculating average video watch time by dividing total watch time only by users who had watched at least three seconds, rather than by all users who had seen the video. This inflated the numbers by somewhere between 60 and 80 percent.
Publishers had fired editorial staff based on metrics that were false. Most of them never rebuilt. The audiences they had developed over years of distinctive editorial work didn’t wait for them to come back.
The engagement trap.
The pivot to video was the most dramatic version of a recurring mistake, but the underlying logic appears everywhere. Optimize for shares, and your content becomes optimized for outrage — because outrage shares more than information. Optimize for time-on-site, and your content becomes optimized for confusion — because something confusing keeps people on the page longer than something clear. Optimize for page views, and your content becomes optimized for slides — because ten slides generate ten page views where one article generates one.
Your content calendar is almost certainly lying to you about this. The metrics you’re watching are not proxies for value. They are proxies for whatever the platform incentivizes this quarter, which will change.
What the survivors have in common.
The organizations that have navigated multiple platform shifts without losing their editorial identity share a characteristic that is not primarily technical. It is structural.
They have — or had — someone whose job was to protect editorial judgment from the demands of the metric. An editor whose authority over what got published was not subordinate to what the analytics said was performing. A person who could say: this content is performing well by the platform’s standard and I won’t publish it anyway, because it’s not what we’re for.
This is not sentiment. It is organizational architecture. The fractional managing editor role exists precisely because most content operations don’t have this person, and the absence is expensive in ways that compound slowly and then suddenly.
AI slop and the current version of the mistake.
We are running the pattern again right now.
The new metric is AI-generated content volume. The insight is correct: AI can produce content faster and cheaper than humans. The implementation is the problem. Content produced at scale without editorial judgment is content produced for the algorithm, not the reader. It will, eventually, be adjusted for. The organizations that have gutted their editorial functions to run AI pipelines are making the same bet that Demand Media made in 2010 and Facebook video publishers made in 2016.
The AI content agency pitch is snake oil for the same reason those bets were snake oil: they’re optimizing for a metric that isn’t actually what readers want and that the platform will eventually stop rewarding.
The organizations that will be standing after this shift are the ones maintaining editorial standards right now, when the economics of doing so look disadvantaged. They are building something that cannot be automated. Everyone else is building something that already is.
The history of online media is the history of the same mistake, told over and over in different fonts. The metric changes. The editorial abandonment is always the same. So is the outcome.
Jacob Clifton is the principal of Clifton Creative.