Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

Publishers Weigh Cutting Google Off as Search Referrals Collapse

Search referrals to publishers have fallen sharply enough that some media companies are now debating whether to block Google's crawlers entirely, according to a Wall Street Journal report summarised by Nieman Journalism Lab on 22 July 2026.

Media & internetAnalysisRachel NwosuPublished: 27 September 20266 min readSources 6
Publishers Weigh Cutting Google Off as Search Referrals Collapse

The numbers have been accumulating for more than a year. Chartbeat data reported exclusively by Axios and covered by Search Engine Journal in March 2026 found search referral traffic to small publishers down 60% over two years, nearly three times the decline at large publishers. Mid-sized sites lost 47%. Large publishers lost 22%.

That fragmentation matters. Most coverage has treated publishers as one bloc. The size breakdown shows the damage is concentrated at the bottom, among sites with the fewest resources to build alternatives.

Chartbeat also logged a 34% fall in page views from Google Search between December 2024 and December 2025, with Google Discover down 15% over the same window. ChatGPT referrals grew more than 200%, but chatbots still account for less than 1% of all publisher page view referrals, so the growth has not come close to replacing what search lost.

Why some publishers are talking about opting out

Google's crawlers serve two purposes: indexing sites for search and training its AI models. Publishers can switch off the bots, but doing so also removes them from traditional search results. According to the Journal, traffic for some publications fell more than 40% between June 2025 and June 2026, based on Semrush data. The Guardian and the BBC saw gains in the same period.

That trade-off, once unthinkable, is now openly discussed. USA Today Co. CEO Mike Reed told the Journal, "It's time to take a stand and say enough is enough." The company, which is suing Google over an alleged ad-tech monopoly, says it may cut off Google's access completely if search traffic keeps falling. "I wouldn't call it a big decision because we're blocking other crawlers," Reed told Adweek. "For those with licensing agreements, they get our content. For those without, we block them."

Reddit is reevaluating its $60 million per year deal with Google after seeing traffic declines, the Journal reported. Politico and Reuters are both considering limiting Google bot access and putting up registration walls. Penske Media sued Google over its AI summaries last year. People Inc. CEO Neil Vogel told the Journal that "Turning [Google] off and blocking them entirely is 100% on the table."

Cloudflare is adding pressure from the infrastructure side. According to Nieman Lab, bots now make up more than half of all web traffic, and from 15 September the company will block multi-purpose crawlers, those that scrape for both search indexing and AI training, unless site owners opt out. "We want a technical solution that allows you to be discoverable without having to give your content away for free," Cloudflare chief strategy officer Stephanie Cohen told Adweek earlier in July.

The regulatory lever in Britain

Last month a British regulatory court ruled that Google must allow publishers to opt out of AI features without affecting traditional search results. Google is testing that capability in the UK and has nine months to make the change, but has not rolled it out globally.

The Economist, based in the UK, has been "actively discussing" opting out of AI search, per the Journal. Its VP of generative AI said opting out of Google entirely is not an option, because "traffic can be choppy and declining, but traffic is still traffic." Time is trying a different route: creating text-based ads that regular readers never see but AI crawlers might, in the hope that chatbots cite the ads in responses without knowing what they are.

None of this happens in a vacuum. The Reuters Institute for the Study of Journalism surveyed 280 media leaders from 51 countries and found they expect search engine referrals to fall by 43% over three years. Search traffic to news sites has already dropped by a third globally in a single year, according to Chartbeat data for more than 2,500 news sites cited by the Guardian in January 2026. Google's AI Overviews appear at the top of about 10% of US search results, and referrals from ChatGPT remain, in the report's phrase, "little more than a rounding error."

"It is not clear what comes next. Publishers fear that AI chatbots are creating a new convenient way of accessing information that could leave news brands, and journalists, out in the cold."

That is Nic Newman, senior research associate at the Reuters Institute, describing the end of what he calls the "traffic era" that has sustained online publishers since the internet began. He added that tech platforms do not hold all the cards, and that reliable news and expert analysis remain important.

Google's counter-argument, and the gap in the data

Google disputes the aggregate picture. In a blog post, Liz Reid, the company's head of search, said the introduction of AI in search was "driving more queries and quality clicks," and described third-party reports of dramatic declines as based on "flawed methodologies, isolated examples, or traffic changes that occurred prior to the rollout of AI features in search." She also acknowledged that traffic is shifting between sites, resulting in decreases for some and increases for others.

Google has a product in this fight too. On 26 June 2025 it launched Offerwall, a free tool in Google Ad Manager that lets readers access content through micropayments, surveys, watching ads, or newsletter sign-ups, after more than a year of tests with 1,000 publishers. TechCrunch reported that Google shared one case study: India's Sakal Media Group used Offerwall on esakal.com and saw a 20% revenue boost and up to 2 million increased impressions within three months. Google confirmed to TechCrunch that publishers with Offerwall saw an average revenue uplift of 9% during testing. Early reports during the trial period cited an average revenue lift of 9% after 1 million messages on AdSense for viewing rewarded ads, and 5% to 15% for Ad Manager customers.

Micropayments have failed before. TechCrunch noted that the economics rarely work, and that the Twitter-like startup Post, backed by a16z, shut down after trying to make them work for publishers.

What the Offerwall announcement did not include was any broader evidence that such tools can offset search losses at scale. The revenue percentages come from publishers who opted into a Google product, not from the wider population of sites watching referrals fall. And the most consequential number in the debate, how much traffic a publisher would lose by blocking Google's crawlers, is still mostly estimated from before the AI rollout, which is precisely the period Google says produces misleading comparisons.

Mark Stenberg of Adweek framed the standoff simply: "No publisher wants to make do without Google, but many are increasingly prepared to do so." His conclusion was blunter still. "Google, on the other hand, has no such backup plan."

Comments 0

Sources

6
  1. 01As AI kills search traffic, Google launches Offerwall to boost publisher revenueEN
  2. 02Google's AI search features are killing traffic to publishersEN
  3. 03Publishers fear AI search summaries and chatbots mean 'end of traffic era'EN
  4. 04Search traffic has declined so much that some publishers are considering opting out of Google entirelyEN
  5. 05Search Referral Traffic Down 60% for Small Publishers, Data ShowsEN
  6. 06Search traffic to the FT down 25-30%, Daily Mail down 89%EN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.