Reader comments, Google ad-blocking and the moderation dilemma for publishers

Google’s recent move to block advertising over reader comments below a Mail Online column should sound a warning to publishers.

A quick recap: Google blocked display adverts from appearing on Piers Morgan’s column about the US gymnast Simone Biles, published at dailymail.co.uk during the Tokyo Olympics.

The piece had some 9,000 reader comments, some of which Google said were “racist” and in breach of its policies.

Mail Online said Google had acted without showing any evidence for its claims and Morgan himself decried Google’s actions as a “disgraceful attack on free speech”. Press Gazette understands the issue was resolved and the article was monetised with ads again.

While it might seem like a small spat on the surface, this incident speaks to the power the tech giant can now wield over publishers.

As the dominant search engine, Google is also the dominant force in search advertising – where ads appear within search results for relevant terms.

Since buying Doubleclick in 2007, Google has also become a powerful force in display advertising – e.g. the banners that appear around articles. News websites have a few options when it comes to monetising their content, but those that don’t choose a paywalled subscription model are reliant on advertising as their main source of income.

Digital ad spend falls into two broad categories: search and display.

As the Wall Street Journal put it: “Google is the major force at every layer between advertisers and websites, providing tools used in the many steps of purchasing and selling online ads – whether they are meant to run on Google’s own platforms or sites around the web.”

The US Department of Justice and the attorney general of Texas have both brought claims against Google over what they claimed was anti-competitive conduct. Google described the US DoJ’s lawsuit as “deeply flawed” and the case brought by Texas – that it was using its power “to manipulate the market” and “destroy competition” – as “misleading”.

[Read more: Regulation: Politicians running the world’s leading economies are planning to tame Facebook and Google]

Google runs two ad networks for publishers: Ad Sense and Ad Manager, which together form a network of websites serving Google Ads. These ads are served programmatically, meaning they are sold in an automated real-time auction that seeks to put them before the eyeballs of relevant users.

When Google blocks adverts from appearing against online content only the advertisers that use its ad server will have their ads blocked. But given Google’s dominance, ad blocking can make a significant dent in revenues from an article, especially one that’s gone viral.

Both sides – Google and news publishers – would agree that explicitly racist comments are not wanted online. All major news websites actively moderate reader comments to remove such terms, but it raises the question of how much Google should be able to influence that process.

News titles often deal with tricky and at times controversial subjects, many of which inspire heated comments below the line. Google is clear that in the case of Mail Online the comments were explicitly racist.

The majority of news websites post-moderate rather than pre-moderate reader comments, which fall under the umbrella of user-generated content. In post-moderation, comments are only removed or edited when it is flagged that they have breached standards.

Publishers prefer this option because it means they are protected provided they down legally contentious comments swiftly once notified.

[Read more: Mail Online search traffic halves after Google algorithm update]

Blocking ads is ‘serious problem’ for publishers

Media law consultant David Banks said Google’s actions to block ads over reader comments “could potentially put publishers in a bit of a difficult position”.

He said: “If a paper now is seeing a threat to its advertising revenue because of the possibility of something racist being put there, which then results in Google withdrawing advertising, then it could potentially push them towards a pre-moderation system for certain stories.

“They might think. ‘okay, this story, because it’s got this kind of content and is likely to attract this sort of comment, we’re going to have to pre-moderate it so that we don’t possibly lose the advertising from it, which would then reduce the libel defence they have for not pre-moderating.

“Or else it’s going to mean they’re going to have to be much more active in post-moderation than they are at the moment. Some people might say well that’s no bad thing given the nature of the types of comment that are made on certain kinds of story.”

The Defamation Act 2013 offers a defence to libel for publisher who can show they did not publish the offending comment, but in order for this to succeed they must be able to identify the poster. This means having accurate identity details for users, which many websites do not have because they allow people to register with minimal details.

The only way to tackle this would be to impose a more rigorous registration process for users before allowing users to comment, said Banks.

“I can see why publishers would be concerned. Columnists are… often there to put forward views that will stir up a reaction. You can see when they’re getting close to the edge of things it might attract the sort of attention that Google is now saying: ‘We’ll cut off your advertising if you do that’.

“That’s a serious problem… Piers Morgan is a figure that has his detractors as well as his fans, but this isn’t just going to apply to him – there are lots of columnists out there who write difficult stuff.”

[Read more: The highs and lows of Piers Morgan’s career in journalism]

Banks said you could end up with a situation where Google blocks ads on a column promoting an anti-racist position because of readers using racist terms to attack the columnist in the comments.

“Are we saying now that Google’s going to remove the advertising from that kind of column as well, where it is an anti-racist column, but nevertheless attracts racist commentary beneath? That’s potentially quite serious,” he said.

NMA: ‘Another example of the enormous power that the tech platforms wield’

Google maintains ad blocking is not a free speech issue, but when you control the technology on which free-to-read news websites rely to make money, it’s hard to argue that you don’t have the whip hand and that this could ultimately have a chilling effect on certain types of editorial content that might inspire offensive comments from readers.

Google’s dominance of online advertising is only set to grows with the move away from a cookies-based internet, in place of which Google is pushing its own solution: the Privacy Sandbox.

[Read more: Enter Sandbox: How Google is building an internet without cookies – and why publishers are concerned]

News Media Association chief executive Owen Meredith said: “This case is yet another example of the enormous power that the tech platforms wield in the digital advertising marketplace and, in this instance, Google’s willingness to use that power according to its own editorial judgements.

“The Digital Markets Unit has been set up to tackle the issue of the tech platforms’ dominance in digital advertising and ensure a fair deal for news publishers. The sooner it is granted, through legislation, the powers it needs to do that job effectively, the better.”

Google: ‘When content breaches our ads policies, we take action…’

Google has said that its policies on user-generated content have been in place for years and publishers are aware of the risks of falling short of these. In a blog post in 2017, it said: “Comments that violate Google policies can also cause your site to no longer be eligible to show Google ads.”

Google also offers guidance on how publishers can avoid falling foul of this. One solution tells publishers to “put comments on their own page, and don’t run ads on that page”, but this ignores the habit of digital readers to scroll down to comments as part of their reading experience, and suggests publishers add a barrier to content it then cannot monetise.

A Google spokesperson told Press Gazette: “We have strict publisher policies that prohibit ads from running alongside harmful or hateful content to protect our advertising partners and users alike.

“Because these policies apply to user-generated content as well, such as comment sections, we’ve also provided extensive guidance on how publishers can best manage the risks associated with user comments to remain compliant with our guidelines.

“When content breaches our ads policies, we take action to prevent ads from running on that page. We alert publishers of these actions so they can address the issue and apply to have ads reinstated.”

Google blocks ads on thousands of pages across the web each month that violate its policies. Publishers are informed when ads are being blocked so they can review and moderate offending content and apply for display ads to be reinstated on the page.

The tech giant has thousands of staff working 24/7 to deliver a safe experience for web users. It uses a combination of technology and human reviewers to spot and enforce its web policy violations.

UK news website reader comment policies:

Every major newspaper website has a warning to users that racist, sexist or homophobic abuse will not be tolerated. Most post-moderate, with a filter to catch offensive terms and a team responding to users’ flagging up posts. Many also pre-moderate comments on sensitive articles, or pre-emptively shut off the comments section altogether.

Reader comments published on news websites that are members of the Independent Press Standards Organisation fall under the regulator’s remit “so long as these comments can be moderated by the publication”.

This means IPSO will adjudicate on complaints about user comments that violate the Editors’ Code of Practice, and indeed it has done so.

In 2020 IPSO found a breach of Clause 1 (accuracy) in a reader comment published on the Oxford Mail website that remained online for a month without being reviewed following a complaint.

In its ruling, IPSO’s Complaints Committee said: “Given that the comment was not pre-moderated, the requirement for the publication to take care did not begin from the date of first publication.

“Rather, the Committee found that the requirement to take care over the accuracy of a reader comment began from when the publication was made aware of the alleged breach and was given the opportunity to post-moderate the comment.”

With so many mechanisms in place, Google’s actions to block ads and so revenue generation could be seen as unnecessarily punitive and these heavily regulated news websites pale in comparison to the Wild West of social media, where standards violations are quotidien.

Mirror, Express and Start publisher Reach said it has “continued to invest in moderation and building positive online communities by partnering with an AI-powered platform [Viafoura] across the regional and national news sites and matching this with in-person oversight at Reach.”

Head of communities Daniel Jackson said: “Our policy is that hate speech and uncivil comments should be reviewed and removed swiftly to allow genuine users to engage in meaningful debate.”

Press Gazette understands Google has prevented ads from running against news, politics and live blog articles on theguardian.com, but that this was not the result of reader comments violating Google’s policies.

Since 2000, Press Gazette understands there have been 43 cases of Google ad-serving being disabled on Guardian content, of which only one was last year (which did not have comments enabled) and none this year so far.

Earlier this year the Mail filed an anti-monopoly lawsuit against Google in the US, accusing it and parent company Alphabet of unlawfully acquiring and maintaining “monopolies for the tools that publishers and advertisers use to buy and sell online ad space”.

Google called the Mail’s claims “completely inaccurate” and said Mail Online authorises dozens of other ad tech companies to sell and manage its ad space, including Amazon and Verizon.

Google Ads. Picture: PixieMe / Shutterstock.com

Email pged@pressgazette.co.uk to point out mistakes, provide story tips or send in a letter for publication on our "Letters Page" blog