The Attention Economy Is Collapsing. Good.

Why click-driven media failed—and what we’re building instead

Here’s a fun thought experiment: If a tree falls in the forest and a Google AI Overview summarizes it, did anyone actually click?

For those of us running local newsrooms, the question isn’t rhetorical—it’s existential. Google’s new AI summaries, which now appear above links in many search results, have decimated referral traffic for publishers. Click-throughs are disappearing into the ether. According to Pew, users click just 1% of the time when AI Overviews are present. That’s not a dip. That’s an extinction event.

And yet, the dashboards still look beautiful. Engagement metrics gleam. “Time on page” sparkles. Impressions balloon like dollar-store parade floats.

But the revenue? The readership? The civic impact? Vapor.

The Illusion of Precision

As Tim Hwang outlines in Subprime Attention Crisis, this isn’t a new problem. It’s just gotten harder to ignore. The digital advertising economy has long been built on shaky math, faith-based metrics, and a lot of very confident PowerPoint decks.

Take Nico Neumann, an academic who stood up at Programmatic I/O—a sort of CES for ad-tech folks—and calmly explained that most of the consumer targeting data in the industry is laughably inaccurate. Some datasets were wrong about 85% of the time. The cost of “precision” ad targeting, he showed, often exceeds any benefit to the advertiser. His conclusion? Old-school mass marketing was more effective.

No one in the audience asked a question. They just moved on to the next panel about “AI-enhanced emotional resonance optimization.”

If You Build It (and Let Google Scrape It), They Won’t Come

At The Charlotte Mercury, we’ve seen the same pattern firsthand: Google gets smarter, our traffic gets thinner, and the analytics tools politely ask if we’d like to spend more on “discovery campaigns.” What used to be a virtuous cycle—good reporting earns trust, trust earns clicks, clicks fund more reporting—has quietly been replaced by a rigged carnival game.

As The Attention Merchants author Tim Wu put it, the platforms don’t care who you are or what you make. They only care how long you can hold someone’s gaze before the next ad slot needs filling.

So what happens when those ads are based on garbage data? When the AI-generated summary cannibalizes the reporting it’s based on? When publishers are both the source and the victim of someone else’s monetization scheme?

You get a lot of beautifully formatted reports showing “record reach” and “strong Q2 engagement.” And a newsroom with no money.

Building Something That Can’t Be Summarized

Our answer isn’t louder headlines or faster SEO pivots. It’s a slower, deeper, more honest conversation with readers—rooted in first-party relationships, not drive-by impressions.

No third-party trackers. No pop-ups begging you to enable notifications. No clickbait headlines about traffic crashes unless the street is actually closed.

We don’t optimize for engagement metrics because we don’t work for engagement brokers. We work for readers. For neighbors. For people who understand that what happens at a city council meeting matters more than the click count.

We’re not anti-technology. We’re just anti-bullshit.

And if the attention economy has to collapse before something better can be built, then let it burn.


About the Author

Peter Cellino is the founder of Mercury Local, often fueled by cold brew from wherever isn’t Google-owned. You can find him causing trouble over on @pc51.bsky.social—DMs open for bug reports, ad metrics rants, or pastry tips. He also blogs frequently at mercurylocal.com/blog, where he’s trying to help rebuild local journalism without feeding the bots.


Privacy Policy | About Us | Terms of Service | Media Kit | Contact


Creative Commons License

© 2025 Mercury Local / Mercury Local
This article, “The Attention Economy Is Collapsing. Good.” by Peter Cellino is licensed under CC BY-ND 4.0.

“The Attention Economy Is Collapsing. Good.”
by Peter Cellino, Mercury Local (CC BY-ND 4.0)