In 2011, I co-founded a software company on an emerging best practice: content marketing was a game you won by producing consistently, at volume, without ever going quiet.
DivvyHQ, a content marketing platform, was the tool we built to make that possible. Editorial calendars, production workflows, a command center for teams cranking out more content than they could otherwise keep straight.
We shipped features whose entire job was making sure you never missed a publish date. We grew it from a whiteboard sketch to an acquisition, and for the better part of fifteen years I believed, and sold, the premise underneath the whole thing. More content, published reliably, was how you got found.
I read the new Orbit Media Blogging Statistics Report last week, their annual survey of more than a thousand content marketers, and felt two things at the same time. Vindication, because the numbers backed up a shift I’ve spent all year arguing. And a quiet gut-punch, because those same numbers retired the bet I built a company on.
Two figures did most of the damage.
92.4% of marketers now use AI for their blogging. Three years ago that number was basically zero. So the superpower finally arrived. The thing every content team used to fantasize about, infinite content produced on demand, is sitting on all of our desks right now.
Over that exact same stretch, the share of marketers reporting “strong results” from their content dropped to 13.9%, the lowest in the survey’s history.
For twelve years that number had held somewhere between 20% and 30%. And the AI everyone raced to adopt shows no correlation with results at all. Teams using it were no more likely to report strong results than the 8% still writing without it.
Sit with that one for a minute… I spent years building software to help teams produce more content, faster. The ultimate version of that has arrived, free and instant for everyone. And collectively, the results are worse than they were before AI existed.
We Used AI to Cut the Work That Actually Works
The lazy version of this story writes itself: AI flooded the internet with junk and buried everyone under the pile.
Andy Crestodina, who has run this survey for thirteen years, points somewhere less convenient. The practices and effort that actually drive results like original research, real human editing, collaboration with outside experts, and honest keyword work are the exact things marketers seem to be walking away from.
Every one of those practices is down. The shortcuts are up.
AI made cutting those corners frictionless, so a whole industry cut them, and we lost the parts that were carrying the good stuff.
The teams that kept the effort in are still winning. The programs built on original research, outside expertise, and real depth are the ones still posting strong results.
Publishing original research alone improves your odds by roughly 50%, and it’s the one thing an AI can’t gin up for you.
Depth is What Earns the AI Recommendation
The reason depth pays now goes past effort for its own sake. Your buyer increasingly asks an AI for a recommendation before they ever come looking for you. That system reads everything it can find about you, decides whether you’re credible on the exact problem it’s being asked about, and either puts your name forward or skips you. Get skipped and there’s nothing to recover…no page two and no second look.
The catch is that depth doesn’t stretch across everything. A marketing team of three can’t build original research and corroborated proof across forty topics, but they can absolutely do it across a few. So the strategic act is deciding which handful of problems you’re going to be undeniable on.
Problem Ownership: Going Deep on a Few Buyer Problems
At Forge & Fathom, Problem Ownership is the north star of how we think about AI visibility. You own a problem when AI systems consistently recommend you as buyers describe that problem in their own words, before they’ve named your category, your competitors, or you.
The change it demands shows up in where you start. Instead of asking “what should we publish this month,” you ask “which few buyer problems are we going to own,” phrased in the raw language buyers actually use. Then you narrow your strategic focus to the three to five problems you can credibly and defensibly win.
Most teams still organize their content around topics they want to rank for, or keyword clusters they can blanket. That’s the reflex Problem Ownership breaks. A topic is what you want to be found for. A problem is what your buyer is trying to solve, in the words they’d use to describe it.
Build around the second one.
The Coverage Illusion: The Trap I Lived In
The Coverage Illusion is the belief that a big pile of content on a problem or topic means you own that problem. You’ve published twenty posts on the thing. You’ve got a pillar page, a couple of guides, a respectable traffic line. It feels like ownership.
What you actually have is volume, and an AI system reads volume very differently than depth. I spent a long stretch of my career helping teams gauge their expertise by counting their output, which in hindsight is a bit like measuring how well you know a city by tallying how many times you’ve driven through it on the highway.
If you’re producing more than ever and it’s landing less than ever, the answer isn’t a better prompt or a faster tool. The uncomfortable question sitting under all of it is whether your leadership team has actually decided which problems you intend to own, or whether you’ve been covering ground at speed and calling it strategy.
That’s a harder conversation than “let’s ship more.” It’s also the one that decides which teams get recommended and which ones are publishing into a void.
I don’t regret the bet I made in 2011. It was the right call for that era. Consistency and volume genuinely won the contest that existed, and building software to help teams pull it off was honest work I’m still proud of. That era just closed while a lot of us were still optimizing for it.
So I’m placing a new bet.
The teams that win the next decade will be the ones who decide, on purpose, what they want to be known for, and build deep layers of evidence that make an unequivocal case for the LLMs to recommend.
I spent fifteen years betting on “more”. I’m betting the next fifteen on depth.
How Deep is Your Moat?
If you can’t say for certain whether you own a single problem in the eyes of AI, that’s a gap you can diagnose. Our AI360 Snapshot shows you where you currently surface when buyers describe their problems to AI tools (and where you don’t). So you can see whether your content is building ownership or just adding to the pile.
And if the honest answer is that your leadership team never actually agreed on which problems to own in the first place, that’s the deeper thing worth pressure-testing, before you pour another quarter into producing anything at all.
