Hi everyone, this is Neo.

If your LinkedIn feed this month has been nothing but “the algorithm killed my traffic” and “the machines finally won,” congratulations — you’re not alone.

The overseas marketing world really is in collective mourning: Google organic clicks collapsing, LinkedIn reach halved, Substack discovery traffic dead. Every number reads like an obituary for content marketing.

But earlier this month, Tony Uphoff — former UBM TechWeb CEO and now a writer on agentic AI in B2B — published a long piece titled Rage Against the Machines, and his core argument stings: the numbers are all real, but you’re aiming your rage at the wrong target.

SEJ’s Greg Jarboe followed up with an analysis that checked Uphoff’s data, found an error in it, and turned the essay into something practitioners can actually act on. Today I’m unpacking the whole discussion — because the conclusion directly affects how independent site owners should spend their budget next year.

1. First, the numbers: how bad is it, really?

Between Uphoff’s original post and Jarboe’s follow-up, you get a fairly complete picture of the carnage:

On the Google side:

• Organic clicks down roughly 42%: that’s Define Media Group’s number, pulled from Search Console data across 64 sites — organic clicks have fallen 42% since AI Overviews scaled across search results;
• AI summary vs no AI summary: click-through roughly halves: Pew Research, tracking the real browsing behavior of 900 U.S. adults in March 2025, found users click a traditional result 8% of the time when an AI summary appears, versus 15% when it doesn’t;
• Links inside the AI summary almost never get clicked: Pew found users clicked a summary link in just 1% of visits;
• Users are more likely to just leave: 26% of visits to pages with an AI summary ended the browsing session entirely, versus 16% on pages without one.

On the LinkedIn side:

• Organic reach down 47% year-over-year (Uphoff’s own post says roughly 50%): LinkedIn rebuilt its feed ranking around 360Brew, a 150-billion-parameter model that ranks posts by expertise and dwell time instead of reaction counts;
• Posts with outbound links get throttled 30%–50% against native content — the harshest cut of all for anyone doing content marketing;
• 41% of LinkedIn long-form posts are fully AI-written: Pangram Labs scanned over a million posts and found LinkedIn has the highest share of AI-written long-form content of any major platform — and 62% of all AI-flagged content across platforms came from LinkedIn;
• On July 30, LinkedIn quietly added a “seems like AI slop” report button.

The bigger backdrop:

• Mercer’s 2026 Global Talent Trends survey puts fear of losing a job to AI at 40% of workers worldwide, up from 28% two years ago;
• Challenger, Gray & Christmas has tracked AI as the cited reason for more than 101,000 announced job cuts in the first half of 2026 — roughly double the full-year 2025 pace.

2. Uphoff’s core argument: decline vs. structural disruption

The numbers are brutal. But Uphoff says: you’re all reading them wrong.

His key distinction:

• Decline means the thing you’re doing no longer has a version worth saving;
• Structural disruption means the function still matters — the form it takes is changing underneath you.

He’s lived through this before: at ThomasNet when print directories went digital, and at UBM TechWeb when the 2008 financial crisis hit in the same month as the print-to-digital collapse. Both times, the loudest reaction in the building wasn’t strategy — it was people angry at the channel that was disappearing, rather than at the decision to bet the whole business on it.

The line from his post worth highlighting in bold:

“The machines aren’t taking anything from you that you haven’t already agreed to rent out.”

Translation: you parked all your traffic on someone else’s platform, built your business model on reach that platform allowed you to have, and now they changed the rules — so you’re mad at them? The person who agreed to rent instead of buy deserves a share of the blame.

3. A lesson in AI-era data literacy: how 85% became 43%

Jarboe’s analysis contains one detail worth remembering on its own — about data literacy in the AI era.

Uphoff’s original post cited “Business Insider lost 85% of its Google referral traffic,” a figure from a Wall Street Journal chart. But the Journal corrected that chart on July 22 — three weeks before Uphoff’s piece ran. The corrected trailing-12-month numbers put Business Insider’s decline at 43%, not 85% (USA Today down 28%, The Washington Post down 44%).

The catch: the wrong 85% version had already been circulating across the SEO industry for weeks, cited in posts viewed tens of thousands of times.

Here’s the takeaway: AI-era numbers move fast enough that the version you saw shared last month may already be stale. Spending 30 extra seconds to check the original source before quoting a figure is basic hygiene in 2026.

4. So what do you actually do? Uphoff’s framework, plus my practical add-on

Uphoff ends his essay with a gut check rather than a strategy, which is fair for a leadership piece — but practitioners need next steps. Jarboe supplied them, and I’ve reworked them for independent site owners:

1. Keep “channel metrics” and “function metrics” on separate ledgers

Channel metrics: organic clicks, LinkedIn reach, Facebook impressions. If they drop, they drop — anxiety doesn’t change it.

Function metrics are the real thermometer for your business:

• Is AI still citing your content? (mentions / citations)
• Are email subscribers growing and converting?
• Is direct traffic growing?
• Is your content being saved, bookmarked, and re-shared?

Jarboe’s point: “Those are the numbers that tell you if the function you perform is intact even while the old channel numbers fall.”

2. Audit what you actually own

Email lists, first-party data, direct client relationships — none of those get repriced by an algorithm update or a report button. If 90% of your visibility depends on one platform’s mood, that’s not a topic for your next LinkedIn rant — it’s a business risk to fix this quarter.

3. Stop measuring content by rankings, start measuring it by citations, saves, and re-engagement

We’ve said this before, but today’s data gives it new force: when clicks are structurally shrinking, obsessing over rankings and CTR is like counting fish on a shoreline that’s receding.

Neo’s take

The thing that hit me hardest in this whole debate is Uphoff’s “renting” metaphor.

For the past decade, independent site owners lived with the illusion that platform traffic was “theirs.” The truth is Google Search, LinkedIn, whatever — they’re all storefronts you rented with content and time, and the landlord can change the rent, the rules, or the floor plan whenever they want. What happened in 2026 is just the landlord finally doing a big renovation.

So the real dividing line isn’t “to SEO or not to SEO.” It’s: how much of your business sits on rented reach, and how much sits on assets you own?

Email lists, owned audiences, branded search, AI citation assets, verifiable industry trust — that’s the “buying land” part. Moving your energy from “raging at the algorithm” to “buying land” is the biggest takeaway I got from this debate.

One last note: Uphoff’s original post ran on August 3, and it only exploded across SEJ this week — a two-week gap. In the AI era, information often travels in two waves: the source, then the breakout. Being able to digest a story before everyone else sees it is itself an information advantage. Next time you see a panic post about “traffic is dead,” ask three questions: where did the number come from? Has it been corrected? And is this rage aimed at the wrong target again?