Publishers, news industry bodies and consultants have reacted with scepticism to the results of a Google experiment which concluded news has “no measurable impact” on its ad revenue.
The randomised control trial removed some 13,000 news websites from search, Google Discover and Google News for 1% of users in eight EU countries and observed a negligible drop in daily active users and ad revenue among those in the trial compared with a control group.
But a range of news industry experts have questioned the design of the trial, suggesting it was not well constructed to assess the value of news to Google’s platforms.
And others expressed open distrust of the company and its findings, which come amid a series of international regulatory tussles over whether tech platforms should be compelled to pay publishers for the content they distribute.
The research blocked more than 13,000 European Economic Area “press publications” for the affected users in Belgium, Croatia, Denmark, Greece, Italy, Netherlands Poland and Spain , but not all news content. Publications outside the European Economic Area, for instance, would still have been indexed.
‘The experiment does not show any decline in users’ interest in finding reliable information’
Thomas Höppner, a competition lawyer at Hausfeld – one of the law firms leading a collective action lawsuit against Google on behalf of UK publishers – argued on Linkedin that “the experiment’s design deprives it of any evidentiary value”. His legal claim alleges Google owes UK publishers £13.6bn after abusing its dominant position in the online advertising market.
“If a de-facto monopolist for 20 years replaces original news with other content sources without informing users or offering them any alternative, the lack of switching behaviour within just ten weeks isn’t surprising.”
Instead, Höppner argued, the result “only confirms Google’s entrenched monopoly: even significant quality degradations no longer lead to users switching the platform. Instead, they accept and click at less relevant results, unaware of what they’re missing, and unable to turn elsewhere.
“The experiment does not show any decline in users’ interest in finding reliable information on a search engine.”
David Buttle, a former director of platform strategy and public affairs for the Financial Times who now runs media and technology consultancy DJB Strategies, said: “If Google was trying to prove that news isn’t valuable to it by publishing this research then at best it’s failed and at worst it’s been counter-productive.
“The ad revenue figures are an irrelevance; as we all know Google’s monetisation engine relies on far more than serving ads against search queries (it’s still collecting user data!).” (Google says the experiment looked at all revenue, not just that associated with search.)
“The engagement figures are interesting but Google’s choice to publish [daily average users] rather than search volume is likely to underplay the effect.”
Research published in December by Datos, a subsidiary of Semrush, found that across a panel of 130,000 devices in the US between 2023 and 2024, news and media topics made up 2.4% of Google search queries. For comparison, social media represented 4.2% of queries, “adult” topics 3.6% and vehicles 1.3%. The most popular search topic was arts and entertainment (17.5%), followed by computers and technology (13.4%) and e-commerce and shopping (7.1%).
Google has previously said that news-related queries account “for under 2% of total queries on Google Search globally”.
In common with some other commentators, Buttle questioned the decision not to inform users that they were in the trial.
Randomised control trials, which are often described as the most rigorous way to design a study, require participants to be unaware of whether they are in the control group or the test group. However, Buttle said, this design meant there was no effect from users who “knowing this, would have chosen to go elsewhere”.
“Given this, the 1% decline feels fairly significant to me. Especially when taken in the context of Google’s market share hardly moving for around 15 years.”
‘Self-serving narrative’
Chris Duncan, the former chief executive of UK publishing at Bauer Media Group and a former managing director of Times Newspapers, said he appreciated the study had been “carefully constructed” but that its conclusion “should be treated carefully”.
“The experiment shows that removing a subset of news from a subset of the population without informing them has a number of interesting effects. To take that all the way to ‘news has no impact on ad revenue’ is to stretch the data too far, imho [in my humble opinion].”
Duncan suggested that because the experiment was carried out on 1% of users its ability to capture the significance of news to Google’s platforms was limited.
One well-placed news industry source argued there had been “no data to suggest that users found the alternatives as useful as news articles – not least because they did not know they were alternatives. They would simply assume that The Times or The Guardian did not cover the story they were searching for if a link was not provided after searching”.
They also questioned Google’s reading of the data, saying: “Google conveniently claims that the 0.77% fall in DAU [daily activities views] for web search, and 5.47% DAU decline for Discover are statistically significant, but the 1.54% rise in Google News DAU is not – it is unclear why, and in fact this could demonstrate that users who could not find news in search or Discover were seeking it out on Google News.”
SEO expert Jacob Sawa said: “The report completely overlooks the economic asymmetry between Google and publishers, the value of user data collected, and country-specific variations. This appears to be a self-serving narrative designed to avoid fair compensation for news content rather than an honest assessment of the ecosystem.”
‘Pure propaganda’ from Google
Several commentators flat out refused to believe the results.
SEO consultant Barry Adams claimed the experiment was “pure propaganda designed to strengthen Google’s negotiation position and flatly contradicts proper scientific research about the trust factors that news brands shown in Google’s results instil”.
He said: “I don’t believe a single word Google says about news in its results. They’ve been proven utterly unreliable time and again, very recently when they said AI Overviews would send more traffic to websites and the exact opposite has happened.” (Another SEO consultant, Kristine Schachinger, agreed, saying the results were “total bovine faeces”.)
Angela Mills Wade, the owner of Brussels-based public affairs consultancy Europe Analytica, similarly claimed the experiment was “paving the way to refuse to licence publishers for their content for use in Google’s AI”.
“The only thing it really proves is the depth of Google’s monopoly,” she said. “People didn’t switch platforms because they had no idea the quality dropped. They kept clicking, unaware.
“It completely ignores what was lost — traffic, visibility, and credibility for actual publishers, but confirms Google’s dominance on advertising.”
Owen Meredith, the chief executive of Industry group the News Media Association said: “Google knows perfectly well that the value of trusted journalism extends far beyond its ad revenues, but this highly partial and incomplete picture of the digital ecosystem happens to suit its current lobbying objectives.
“Plenty of authoritative and independent studies, across different territories, have shown otherwise. Regulators must not be fooled.”
James Rosewell, the co-founder of Movement for an Open Web, was similarly dismissive.
“What matters is the proven fact that Google control the ‘gates’ that enable publishers to make money,” he said. “That choke point needs to be removed so that publishers can control their own monetisation.
“If Google takes content and then choose not to make money from it, or remunerate the content owner fairly for their loss as a result of Google’s derived free product, then that is unacceptable.
“Google’s monopoly in search means that they have been able to damage newsbrands’ businesses with impunity. Thankfully, the US search case – which is due to come to a conclusion shortly – alongside actions in the UK and EU are recognizing this and will, hopefully, take action to create a more balanced marketplace.”
Google hits back at publishers sharing ‘false statements’
Press Gazette shared publisher criticisms with Google and received the following statement: “It is not surprising that some publishers choose to share false statements because they don’t like the outcome of this experiment. We developed a robust and economically sound experiment to provide objective and accurate data that publishers and regulators have asked for, and we’re now sharing that transparently. We are not saying that EU news is not important; just that it does not generate significant revenue for Google.”
Google also noted the following points:
- It said that fewer than 2% of all searches globally are news queries and that it does not run ads or make money on the vast majority of them
- The significance of news to revenue, not search volume, is what mattered most in the experiment. Google said that search volume wouldn’t have altered the outcome. Revenue didn’t significantly decline even though daily average users on Search declined by 0.8%, which it said indicates that any lost usage was from queries that generated minimal or no revenue
- On the fact not all sources of news were blocked, Google said other forms of news would still show other than the blocked sites. The report notes: “The top click gainers in the experiment were youtube.com, infobae.com, facebook.com, wikipedia.org, and pinterest.com“
- Google said that other studies suggesting news was worth a great deal to the search giant have been “debunked”.
Email pged@pressgazette.co.uk to point out mistakes, provide story tips or send in a letter for publication on our "Letters Page" blog
