Google Gave Out Free Traffic for Years — Now It's Squeezing Every Publisher Dry

The Deal Nobody Signed, But Everyone Relied On

For twenty years, the arrangement was simple, even if nobody wrote it down: you built something worth reading, Google sent you readers, and everyone made money off the ads in between. Publishers built entire businesses on that handshake. Millions of independent sites — recipe blogs, gear reviewers, travel guides, local news outlets — were built, staffed, and funded on the assumption that if you wrote something genuinely useful, Google would find it and send people your way.

That deal is over. Not quietly winding down — over, in the way a body count is over. And unlike most disputes about search rankings, this one isn’t just publishers complaining on Twitter anymore. As of 2026, it comes backed by two federal court rulings declaring Google an illegal monopolist, an EU antitrust investigation, a UK regulatory designation, and multiple lawsuits from named plaintiffs with financial statements to back up their claims. This is the documented story of how that happened — what changed, who got hurt, what Google says in its own defense, and what the courts have actually found so far.

A note on how this piece was built: every figure, quote, and date below is drawn from named, checkable sources — court rulings, Pew Research, the Wall Street Journal, Digital Content Next, and on-the-record statements from the publishers and Google executives involved. Where the evidence is contested, that’s stated explicitly, not smoothed over.

Part 1: The Update That Publishers Call a “War”

In August 2022, Google rolled out something called the Helpful Content Update, framed around a simple-sounding promise: reward “content by people, for people,” and demote content built primarily to rank rather than to help. Nobody objected to the stated goal. What happened next is why the update has its own graveyard of Reddit threads titled “My site is dead.”

Google expanded the update in September 2023, then folded it entirely into its core ranking systems in March 2024 — meaning, as SEO analyst Lily Ray of Amsive put it, “the helpful content system doesn’t exist on its own anymore.” It’s just part of how Google ranks everything, permanently, with no separate lever to isolate or appeal against.

The damage, by the numbers: an analysis of 671 independent travel publishers found that 32% of them — 213 sites — lost more than 90% of their organic search traffic. Amsive’s research found 73% of surveyed news publishers saw their Google Discover traffic drop to zero after the September 2023 update. Lily Ray’s own characterization of what she was seeing from publishers was blunt: Google, she said, was “committing war on publisher websites.”

The named casualties are not abstractions. These are specific businesses, run by specific people, who documented their own numbers publicly:

• HouseFresh, an air-purifier review site, reported losing roughly 91% of its search traffic.
• Travel Lemming, run by Nate Hake, reported a 94% decline. Hake’s own summary of the mechanism has become one of the most quoted lines in this story: “There’s a problem when a parasite gets too greedy: it can kill its host.”
• Retro Dodo, a retro-gaming site, lost “about 85%” of its traffic according to founder Brandon Saltalamacchia.
• TechRaptor reported losses over 80%.
• GAMURS Group, a gaming media network, laid off 30 staff in September 2024 and explicitly named the Helpful Content Update as the cause.
• Giant Freakin Robot, an entertainment site, collapsed from roughly 20 million monthly visitors to “a few thousand a month” — and shut down.

This wasn’t confined to small, scrappy operations either. Per a June 2025 Wall Street Journal report citing Similarweb data: Business Insider’s organic search traffic fell 55% between April 2022 and April 2025, a decline that preceded a roughly 21% staff cut in May 2025. HuffPost lost about half its search traffic over a similar window. The New York Times’s share of traffic coming from search fell from 44% in 2022 to 36.5% by April 2025. Chegg, the online education company, reported non-subscriber traffic down 49% year-over-year between January 2024 and January 2025 — a decline severe enough that it became the basis of a federal lawsuit (more on that below).

A necessary caveat, stated as plainly as the source material states it: these are largely correlational findings, not laboratory-controlled experiments. Google has repeatedly and correctly noted that some of these declines predate AI Overviews, and that other platform shifts — Facebook and X both deprioritizing outbound links, for instance — hurt referral traffic across the board during the same window. Attributing 100% of any single site’s decline to any one Google update is more than the data alone can prove. What the data does support is a pattern, repeated across dozens of independently-verified cases, that is too consistent to dismiss as coincidence.

Part 2: Google’s Own Employees Admitted It, in the Room

In October 2024, Google invited roughly 20 hand-picked, visibly affected publishers to a “Web Creator Summit” at its Mountain View headquarters — most of whom paid their own way to attend. Charleston Crafted’s Morgan reported spending over $400 out of pocket just to be in the room.

According to multiple attendee accounts and reporting by SEO journalist Marie Haynes, Google’s own search personnel did not dispute that real damage had occurred. Search quality lead Pandu Nayak reportedly acknowledged that affected creators had produced “great content” that Google’s own ranking systems had simply struggled to surface correctly. Google’s public liaison, Danny Sullivan, told the room that many of the affected sites had “nothing wrong” with them — an extraordinary admission, given that the sites in question had lost the majority of their traffic to a Google algorithm change.

Mike Hardaker of Mountain Weekly News — down 91% — told the room he was, at that point, eating at a food bank. Rutledge Daugette of TechRaptor said he had submitted a 26-page document at Google’s request, detailing his site’s situation, and then heard nothing back for 17 months.

Publishers were told a new update addressing the problem was coming “very soon.” They were also told, in the same breath, that full recovery for most of them was unlikely, because “the whole format of search results has changed” — an acknowledgment that whatever ranking adjustments might follow, the broader shift toward AI-generated answers over ranked links was permanent, not a bug to be patched.

Part 3: When the Rich Get Richer — Selectively

If the Helpful Content Update had simply penalized thin, low-effort content across the board, it would be a hard story to tell sympathetically. What makes it a genuine grievance rather than a business complaint is the documented pattern of who got to keep ranking despite thin content, and who didn’t.

In February 2024, HouseFresh’s Gisele Navarro and Danny Ashton published an investigation titled “How Google is killing independent sites like ours.” It documented, with screenshots and side-by-side comparisons, large media brands — Forbes, Better Homes & Gardens, People, Real Simple, many under the Dotdash Meredith umbrella — outranking hands-on product reviewers with “reviews” that paraphrased manufacturer marketing copy, cited experts who did not appear to exist, and reused the same stock photography across sister publications. By later in 2024, HouseFresh said it had “virtually disappeared” from Google search results entirely, and alleged a practice it called “keyword swarming” by larger competitors.

Google’s own SearchLiaison, Danny Sullivan, responded directly on social media, taking the criticism seriously rather than dismissing it: he said he’d like Google to “do more to ensure we’re showing a better diversity of results that does include both small and large publications.” That’s a meaningful admission from Google’s own spokesperson — not a denial, an agreement that the problem was real.

The independent SEO investigation that named a number. Consultant Lars Lofgren published an exposé in September 2024 on what he called “Forbes Marketplace: The Parasite SEO Company Trying to Devour Its Host,” estimating that Forbes’s affiliate-marketplace content generated roughly $236 million a year in revenue, largely on the strength of Google’s willingness to rank Forbes highly regardless of whether individual articles reflected genuine hands-on expertise. Lofgren’s assessment, quoted widely since: “Google has decided that Forbes is the authority on everything,” and — the line that has become something of a rallying cry in SEO circles — “Google’s ranking algorithm is out of control.” Follow-up reporting extended similar scrutiny to CNN Underscored and USA Today’s affiliate operations, and this reporting is widely credited with pushing Google toward expanded enforcement.

Google did eventually act — unevenly. A new policy against “site reputation abuse” (commonly called parasite SEO) took effect in March 2024, with manual enforcement beginning in May. CNN, USA Today, the LA Times, and Fortune lost rankings for coupon-related keywords; Forbes and the Wall Street Journal preemptively pulled their own coupon directories rather than risk penalties. In November 2024, Google formally penalized Forbes, the Wall Street Journal, Time, and CNN, closing a loophole that had previously exempted content published under a brand’s “first-party oversight.”

Search Engine Land’s assessment of the enforcement effort was pointed: it called the policy “a band-aid for a bullet wound” — manual, slow to apply, and inconsistent in practice, while the underlying dynamic that favored high brand-authority domains over lesser-known expert sites remained largely intact. One data point makes that dynamic concrete: analysis of sites affected by the Helpful Content Update found the average Moz Brand Authority score among “losers” was 37, versus 50–52 among sites that held or gained rankings — suggesting the update disproportionately harmed sites with genuine topical expertise but lower brand-name recognition, exactly the pattern HouseFresh had alleged.

Google’s Chief Scientist, Pandu Nayak, later defended the policy against EU regulatory scrutiny, arguing that loosening enforcement “risks rewarding bad actors and degrading the quality of search results” — a defense of the goal that doesn’t fully answer the criticism of how unevenly it was applied.

Part 4: AI Overviews and the Zero-Click Search

If the Helpful Content Update was the first blow, AI Overviews is the one publishers describe as existential. Launched in the US in May 2024 (evolved from Google’s earlier “Search Generative Experience” experiment) and expanded to more than 100 countries by October 2024, AI Overviews places an AI-generated summary directly at the top of search results — often answering the user’s question well enough that there’s no reason to click through to any source at all.

The data on what this does to click-through rates is unusually consistent across independent research groups with no relationship to each other:

• Pew Research (July 2025), analyzing real browsing data from 900 US adults across nearly 69,000 searches, found that when an AI Overview appeared, users clicked a traditional search result link only 8% of the time — versus 15% of the time when no AI summary was shown. Users clicked on a citation link inside the AI summary just 1% of the time. Users also ended their entire browsing session immediately 26% of the time when shown an AI summary, versus 16% without one.
• Digital Content Next, representing roughly 40 major publishers including the New York Times, Condé Nast, and Vox, found a median year-over-year decline in Google referral traffic of about 10% overall (7% for news content, 14% for non-news) over an eight-week window in mid-2025. DCN’s CEO, Jason Kint, called the losses “a direct consequence” of AI Overviews.
• The UK’s Professional Publishers Association, in a 2025 submission to the country’s competition regulator, cited a lifestyle publisher whose click-through rate on a “how to get rid of [insect]” query fell from 5.1% to 0.6% — while the article continued to rank on the first page.
• Ahrefs found, as of February 2026, that AI Overviews now correlate with a 58% reduction in click-through for content ranking in the number-one search position — up sharply from 34.5% just ten months earlier.

Zero-click search — where a user gets their answer and never visits any website at all — rose from 56% of news-related searches in May 2024 to nearly 69% by May 2025, according to Similarweb. For a publisher whose entire revenue model depends on a reader landing on their page and seeing an ad, that’s not a decline in a metric. That’s the removal of the transaction the business was built around.

Part 5: What Google Actually Says

Fairness requires stating Google’s defense with the same rigor as the criticism, so here it is in full.

Google’s VP and Head of Search, Liz Reid, published a company blog post in August 2025 stating that “total organic click volume from Google Search to websites has been relatively stable year-over-year,” and that “average click quality has slightly increased.” The post characterized third-party research showing declines as relying on “flawed methodologies, isolated examples, or traffic changes that occurred prior to the roll out of AI features.” Google SVP Nick Fox separately stated: “We’re now sending billions of clicks to websites every week through AI features in Search alone.” Google spokesperson Jose Castaneda has repeatedly said AI Overviews “send traffic to a greater diversity of sites,” and has called the publisher lawsuits discussed below “meritless.”

The core problem with evaluating this defense: Google’s blog post presented conclusions but no underlying dataset. SEO consultant Aleyda Solis’s public response captured the frustration of researchers trying to verify the claim: “It’s sad how they try to manipulate our perception of what’s happening using half-baked half-truths… we know better than that.” Without the raw numbers, Google’s “stable clicks” claim isn’t falsifiable by outside researchers — which is itself one of publishers’ core grievances: that Google is the only party in this dispute with access to the ground-truth data, and it hasn’t shared it.

In fairness to Google, not every data point favors the “collapse” narrative. Semrush found the zero-click rate specifically on AI-Overview-triggering keywords actually fell — from more than 45% in January 2025 to 38% by October 2025. And ChatGPT alone sent an estimated 1.2 billion referral visits to publisher websites between September and November 2025, a real and growing alternative traffic source, even if researchers agree it hasn’t yet come close to offsetting what search has taken away.

Part 6: The Courts Have Started Ruling — And Google Is Losing

This is the part of the story that moves it from “publishers are upset” to “a federal judge agrees something illegal happened,” and it’s worth being precise about what has and hasn’t actually been decided.

United States v. Google (Search). On August 5, 2024, Judge Amit Mehta ruled that “Google is a monopolist, and it has acted as one to maintain its monopoly,” finding Google in violation of Section 2 of the Sherman Antitrust Act. The court found Google held roughly 90% of the desktop search market and 95% of the smartphone search market. On September 2, 2025, Mehta issued behavioral remedies: banning exclusive default-search-placement contracts, requiring some search-data sharing, and imposing six years of technical-committee oversight. He declined to force a breakup of Chrome or Android, citing — notably — the disruptive, fast-moving rise of generative AI as a reason to avoid a structural remedy that might not fit the market a year later. Both sides are expected to appeal. It’s also worth noting precisely what Mehta did not do: he explicitly rejected proposed remedies that would have forced changes to Google’s publisher-facing policies, finding them unrelated to the specific monopolistic conduct at issue in that case.

United States v. Google (Ad Tech). On April 17, 2025, Judge Leonie Brinkema ruled that Google illegally monopolized the market for publisher ad servers and ad exchanges, finding conduct that “substantially harmed Google’s publisher customers, the competitive process, and, ultimately, consumers of information on the open web.” New York Attorney General Letitia James stated that “everyone from major news organisations to small independent bloggers has taken a financial hit because of Google’s conduct.” The Department of Justice sought a forced divestiture of Google’s ad exchange and open-sourcing of its ad-server auction logic; at closing arguments in late 2025, Brinkema expressed skepticism that a forced breakup was practical, partly because no clear buyer for the ad exchange exists. A remedies ruling is expected in early 2026.

Chegg v. Google, filed February 24, 2025, is the first major antitrust suit specifically targeting AI Overviews. Chegg alleges AI Overviews “crushed” its traffic and that Google used Chegg’s 135 million question-and-answer pairs to train its AI models without adequate compensation. An amended complaint in June 2025 added the allegation that Google effectively coerces publishers into supplying content for AI training under threat of reduced search visibility if they refuse. Chegg’s public statement was direct: Google is “killing the digital information ecosystem.”

Penske Media v. Google, filed September 12, 2025, is the first major U.S. news publisher lawsuit over AI Overviews — Penske owns Rolling Stone, Variety, Billboard, and the Hollywood Reporter. The suit alleges Google effectively forces publishers to accept AI-generated summarization of their content or risk being buried in search results, that roughly 20% of searches linking to Penske properties now trigger an AI Overview, and that Penske’s affiliate revenue has fallen more than a third from its peak. CEO Jay Penske said Google’s conduct threatens “the future of digital media.” Penske filed a second, related lawsuit over advertising practices in January 2026.

European Union and United Kingdom. A coalition — the Independent Publishers Alliance, Movement for an Open Web, and Foxglove Legal — filed a formal EU antitrust complaint on June 30, 2025. The European Commission opened a formal investigation on December 9, 2025 into Google’s use of publisher content and YouTube videos to train its AI systems. Regulatory adviser Tim Cowen offered a memorable framing: “Google has broken the bargain that underpins the internet… Gemini is Search’s evil twin.” Separately, the UK’s Competition and Markets Authority designated Google with “strategic market status” in search on October 10, 2025 — its first such designation under new UK digital-competition law — and, acting on the same coalition’s complaint, has moved to require that publishers be able to opt out of AI Overviews without losing their ordinary search rankings, and that Google properly attribute and link to the sources it summarizes.

Part 7: What This Actually Means If You Run a Website

This isn’t purely academic if you or someone you know depends on search traffic for a living. A few grounded, non-hyped takeaways:

If more than half your traffic comes from Google search, that’s now a single point of failure, not a stable foundation. The publishers weathering this shift best — per reporting from Digital Content Next and Digiday — are the ones with genuine direct-brand recognition: people who type your name into a browser bar, or open your app, or are on your email list, rather than finding you fresh through a ranked link each time. AI interfaces have quietly eroded the old assumption that ranking #1 signals trustworthiness, because increasingly, the user never sees the ranking at all.

Watch the “crawler separation” issue specifically. Right now, Google’s AI-training crawler (Google-Extended) is technically separable from its search-indexing crawler — but blocking the AI crawler currently also risks your visibility in AI Overviews, which increasingly is the search results page. Forcing a genuine separation — so a publisher could opt out of AI training or AI summarization without being penalized in ordinary search — is one of the specific asks in both the DOJ case and the UK regulatory action, and it would be the single most consequential structural change if regulators actually deliver it.

Document everything, publicly, if you’re affected. The publishers who successfully moved this conversation — HouseFresh, Chegg, Penske — did it with hard Search Console data and public write-ups, not just anecdotes. That material is exactly what feeds regulatory investigations and lawsuits.

Frequently Asked Questions

Q: Is Google’s search algorithm actually “out of control,” or is this publisher sour grapes?

A: Both framings oversimplify a genuinely mixed picture. The financial self-interest of publishers and SEO consultants in amplifying this narrative is real and worth remembering — traffic loss is existential for their businesses. But two independent federal courts have now ruled, using the ordinary evidentiary standards of antitrust law, that Google illegally maintained a monopoly in search and in ad tech respectively. That’s a higher bar than a publisher blog post, and it was cleared twice.

Q: Has Google actually done anything wrong regarding AI Overviews specifically?

A: No court has yet ruled on AI Overviews directly — the Chegg and Penske Media lawsuits are ongoing, not decided. What’s established is the click-through data (Pew, Ahrefs, DCN) showing a large, consistent reduction in traffic when AI Overviews appear, and Google’s own defense of that data has not included any published dataset of its own.

Q: What would actually change this?

A: Per the source reporting: a forced divestiture in the ad-tech case, or an EU/UK ruling mandating that publishers can opt out of AI Overviews without losing ordinary search visibility, would meaningfully shift leverage back toward publishers. Absent that, most analysts assume no structural relief before 2027 at the earliest, given the pace of appeals.

Q: Is any of this offset by AI chatbots sending traffic instead?

A: Partially, and it’s growing — ChatGPT alone sent an estimated 1.2 billion referral visits to publishers in a recent three-month window. But researchers who track this, including those cited by Digital Content Next, agree it is not yet close to replacing what search referrals have taken away.

Conclusion: A Bargain, Broken in Slow Motion

The uncomfortable version of this story is the one that resists a single villain. Google’s stated goals — reward genuinely useful content, reduce AI-generated ranking spam — were reasonable on their face, and some of what got demoted probably deserved it. But the documented pattern is unambiguous: independent sites with real, hands-on expertise lost the majority of their traffic and, in several cases, their entire business, while large brands with weaker individual content continued to rank — until public pressure and, eventually, manual policy enforcement caught up, unevenly and years too late for some of the sites that didn’t survive.

And underneath the algorithm dispute sits a harder problem that content updates can’t fix: even a perfectly fair ranking algorithm still sends a shrinking fraction of users anywhere at all, because the product Google is building now increasingly answers the question itself. Two federal courts have already found Google broke the law to protect its position in search and ad tech. Whether that finding translates into a search results page that still sends people to the humans who wrote the answer — rather than just the answer itself — is the open question regulators, courts, and the publishers still standing are all fighting over right now.

Disclaimer: This article synthesizes publicly reported facts, named on-the-record statements, and court rulings from sources including the Wall Street Journal, Pew Research Center, Digital Content Next, Search Engine Land, Amsive, and official court filings and rulings in United States v. Google LLC (search and ad tech cases), Chegg, Inc. v. Google LLC, and Penske Media Corp. v. Google LLC. Figures describing traffic declines are correlational third-party estimates, not court-established facts, unless specifically attributed to a ruling. Google’s public statements are represented as stated by the company; where Google’s position and third-party research conflict, both are presented rather than resolved in either direction.