The Portal Problem

I was in a CNN boardroom last month when someone asked the question that’s haunting every media executive: “Why do our analytics show people spending 30 seconds on our homepage but three hours on TikTok?”

The room went quiet. Because everyone knew the answer, but nobody wanted to say it.

Our audiences didn’t just change platforms. They evolved into a completely different species of information consumer. And we’re still broadcasting to the old one.

The Day Everything Broke

Here’s what actually happened (and I’ve got the server logs to prove it): Somewhere between 2019 and 2022, human beings fundamentally rewired how they consume information. Not gradually. Violently.

I remember watching Arc XP traffic patterns during this period—we power content delivery for hundreds of major publishers—and seeing something that should have been impossible. Homepage visits were collapsing while individual article shares were exploding. People were reading our clients’ journalism, but they weren’t visiting our clients’ websites.

The math didn’t compute. Until it did.

They’d moved to the feeds.

Three Generations, One Problem

Think of audiences like operating systems. Most people can still boot up Windows 95 if they have to, but nobody chooses to live there.

Audience 1.0 was appointment television. Walter Cronkite at 6:30. The morning paper on the kitchen table. We controlled the schedule, they showed up.

Audience 2.0 gave us the logged-in web. Email newsletters. Bookmarked sites. Push notifications. They let us track them in exchange for some personalization. We built paywalls and recommendation engines for this group. Most media companies are still optimizing for them.

Audience 3.0 broke everything.

They don’t visit websites—content finds them. They don’t trust institutions—they trust individual creators. They don’t read articles—they consume stories across formats, platforms, and contexts we never imagined. A TikTok video explains the Supreme Court decision. A ChatGPT summary covers the election. A Discord thread dissects the economic data.

Same information. Completely different physics.

The Brutal Truth Nobody’s Saying

Most media executives are lying to themselves about this transition. (I know because I sit in their strategy meetings.)

They’ll tell you Audience 3.0 is just young people with short attention spans. That “real” readers still come to homepages. That TikTok journalism isn’t serious journalism.

Bullshit.

I’ve seen the data. Audience 3.0 will watch a four-hour video essay about the history of NATO, then immediately context-switch to a 15-second news update, then dive into a 200-comment Reddit thread for analysis. They’re not shallow—they’re operating at information velocities that would have been impossible five years ago.

They didn’t abandon depth. They abandoned our containers.

What I Learned Building the Machine

For the past decade, I’ve helped build the software infrastructure that runs modern journalism. Arc XP powers digital operations for major publishers worldwide. Before that, I worked on AP’s B2B platforms that distribute news to thousands of outlets globally.

These systems made newsrooms incredibly efficient. Faster publishing. Better collaboration. Streamlined workflows.

But here’s what we missed: We optimized the wrong machine.

We made it easier to publish articles when our audiences were moving to feeds. We improved homepage performance when people stopped visiting homepages. We built better CMSes when the fundamental unit of media consumption was shifting from articles to… something else entirely.

(What’s that something else? Still figuring it out. But it’s definitely not articles.)

The Uncomfortable Math

If you accept that Audience 3.0 represents the future—and demographics guarantee it does—then several things become mathematically certain:

Our core products are obsolete. Articles, sections, homepages? Those are broadcast-era artifacts.

Our brands matter less than our people. Audiences follow individual creators, not institutional mastheads. Casey Newton has more influence than The Verge. Taylor Lorenz drives more engagement than The Washington Post.

Distribution belongs to platforms we don’t control. Facebook, TikTok, YouTube, Discord, ChatGPT—they own the pipes now.

One-size-fits-all content is dead. The same story needs to exist as a newsletter, a video, a thread, a chat response, and formats we haven’t invented yet.

This isn’t a trend to resist. It’s physics.

What They Actually Want

Strip away the platform complexity and technological noise. Audience 3.0 is asking for something surprisingly simple:

Stories that understand them as individuals. Content that finds them instead of hiding behind paywalls. Information that matches their context—sometimes a summary, sometimes a deep dive, always at the right moment.

They want journalism that assumes they’re smart but busy. Experiences that improve through interaction. Voices they recognize and trust.

They don’t hate journalism. They hate how we package it.

The Organizations That Get It

Some companies are building for this reality. The New York Times’ games and cooking apps acknowledge that not everything needs to be an article. Individual journalists building massive audiences on Substack understand that trust is personal, not institutional.

But these are exceptions in an industry that’s still structurally organized around broadcast-era assumptions.

The Portal Closed Behind Them

Every day we spend optimizing for Audience 2.0, we lose more ground to creators and platforms that never hesitated to serve Audience 3.0. Our audiences aren’t waiting for us to figure this out.

They’ve already moved on.

The question isn’t whether to adapt. It’s whether we’ll have any audience left by the time we try.

I believe journalism matters. I chose this industry because stories shape reality. But if those stories can’t reach people where they actually live, we’re just broadcasting to empty space.

The portal has closed. Time to build our own way through.