“Whatever sins this platform has boring is not one of them,” billionaire Twitter owner Elon Musk posted this week, an apparent ribbing of the new app Threads from billionaire rival Mark Zuckerberg’s Meta, which was created as an anti-Twitter and has faced its own criticism for being almost too sanitized.

But for Twitter, being not boring has alienated some advertisers, and for those still on the platform, has required radical acceptance of the platform's rowdier corners and Musk's free-tweeting ways.  

The 2023 FIFA Women's World Cup, which kicks off today, will be a test for Twitter to see whether it can retain its dominance as the platform of choice for real-time brand marketing during live events, even as it comes up against new potential rivals and concerns over brand safety. 

Initially, it appears Twitter is being used the way the platform has been for years around global sporting events. Fox Sports is flooding Twitter with highlights and live videos from the tournament, and brands can sponsor the coverage. Fox Sports declined a request to comment on its Twitter strategy during the Women's World Cup, and the nature of its deal with Twitter.

Also read: Twitter and TikTok battle for World Cup brands

Fox Sports broadcasts the games in the U.S., and its content deal with Twitter is one of the first to play out under the leadership of new CEO Linda Yaccarino.

For her part, Yaccarino has argued that attempts to paint Twitter as a cauldron of harmful content are false, noting in a tweet this week that “more than 99% of content users and advertisers see on Twitter is healthy.”

Twitter declined to comment for this story. A person close to Twitter said that major brands were active on the site around World Cup, including Taco Bell, Chevron, Frito-Lay, Visa, Dick's Sporting Goods, Apple, Microsoft, Coke, Samsung, Volkswagen, Google Pixel, Gatorade, McDonald's and Budweiser.

Brands have been interested in sports, and other highly curated content on Twitter, as a way to create a more controlled setting on the platform to run ads. For example, a pre-roll commercial may run before a video of a World Cup recap. There are also brand integrations and sponsored hashtags, which have been utilized by Frito-Lay and State Farm, respectively. 

Brands strike deals with Twitter and Fox Sports to take part in real-time marketing around the World Cup, and advertisers see it as a respite from the often negative chorus in the usual Twitter timeline.

But even the World Cup can get divisive on Twitter these days. One ad from Fox Sports posted last week was flooded with replies that blasted the U.S. Women’s National Team. Fox asked fans to hit the heart button to get reminders about when the games were on—there were some heart emojis in the replies, but also plenty of criticism. One Twitter commenter called the U.S. team “spoiled brats.” “Nah, I’m good,” said another reply. Other comments were more critical and demeaning of the women's team. 

Stress test

The Women’s World Cup is an important moment under Yaccarino’s leadership, which has already been tested in her first weeks on the job. Meta launched Threads to provide an alternate site for real-time conversation around events and become an outpost for online cultural commentary. Some brands that had fled Twitter already started popping up on Threads. Meanwhile, Musk, who hired Yaccarino to pump up ad sales, recently acknowledged, again, that Twitter’s revenue has plummeted since he took over last year. “We’re still negative cash flow, due to [about a] 50% drop in advertising revenue,” Musk tweeted last week.

Then this week, a Bloomberg News report re-raised one of the looming question marks brands have about Twitter: How safe is it? Yaccarino has been touting a robust strategy to deploy brand safety tools, some of which have been released already and some that are still in the works. And she has cited stats about the relatively minuscule level of “hate speech” on Twitter.

But advertisers are not convinced, and Bloomberg News reported that there has been a surge in “harmful content” on Twitter, citing research from a group called the Center for Countering Digital Hate. Since Musk bought Twitter for $44 billion in October, the Center for Countering Digital Hate has issued periodic reports about harmful content such as racial slurs and hateful remarks directed at LGBTQ+ communities. For advertisers, the picture is cloudy, and it’s hard to know exactly how much offensive material permeates any given platform.

“Twitter will tell you that hate speech is at an all-time low, that it's extremely low,” said Jason Lee, senior VP of data and digital strategy at Horizon Media, the independent media buying firm. “They’ve had third parties they partnered with, a company called Sprinklr, to do an investigation and overall an audit to take a look at hate speech.”

In response to the Bloomberg News report about harmful content, Yaccarino tweeted that “99.99% of tweet impressions are healthy.” This week, Twitter Safety said that Sprinklr detected a negligible level of “hate speech,” at a rate of .0003%, meaning about 3 out of every 100,000 posts could be classified as containing hate speech. For comparison, Meta reported in its latest quarterly community standards report that it detected a .02% level of hate speech on Facebook, or 2 out of every 10,000 posts. However, methodologies and definitions are different for how both platforms collect these reports, and their frameworks for validating their figures with independent third parties. For one, Meta issues a quarterly community standards enforcement report, while Twitter under Musk has not set a consistent plan to report similarly in-depth research into content on the service.

On Thursday, Twitter received an assist from Brandwatch, the social analytics platform, which had been cited in Bloomberg’s “harmful content” report. “Recently, we were cited in an article about brand relevance that relied on incomplete and outdated data,” Brandwatch posted to Twitter. “It contained metrics used out of context to make unsubstantiated assertions about Twitter.”

Yaccarino retweeted the post, calling it “further validation from a third party regarding our progress.”

Never 100% safe

It’s hard to know, for sure, Lee said. “We’re never at 100% safety but the question is ‘how big is the crack,’” Lee said, “and I think that is what the industry is worried about [with Twitter], we don’t know what we don’t know and whether there are things behind the scenes that we don’t have access to.”

Twitter has been working on a version of brand safety tools that also mesh with the dictates of Musk, who has bristled at being constrained by advertisers. Many advertisers, though, are not fully sold on Twitter’s claims about the health of its service and say that guardrails could be stronger. Advertisers also say it’s hard to know how Twitter is performing with its tools that are designed to keep ads away from bad content.

“There’s, generally speaking, a hesitancy to be on Twitter because of the brand safety concerns,” said Lee. “There’s brand safety concerns on all platforms, or most platforms, it’s just a matter of how the platform pivots, how communicative they are, how quickly they jump on it and update their technology.”

Twitter has been working with DoubleVerify and Integral Ad Science as third-party validators to report to brands about the context in which ads appear. Twitter has promised brands they can avoid adjacencies to unacceptable tweets. Lee said avoiding harmful tweets is harder than it sounds, though. Twitter has given brands the ability to avoid up to 2,000 terms—words that would raise red flags for the brands’ campaigns. And advertisers can exclude up to 2,000 accounts from their campaigns, meaning their ads would not appear next to any of those Twitter users. “I don’t think 2,000 is enough,” Lee said, adding that in other programmatic advertising platforms brands can sometimes create exclusion lists of up to 30,000 terms.

As for the reporting from third parties, DoubleVerify and Integral Ad Science do not have visibility into every part of Twitter. Ads that run in replies under a tweet can’t be validated for quality, for instance, and more ads are popping up in comment sections and on the account timelines of highly followed Twitter accounts. In fact, last week, Twitter implemented an ad-revenue sharing program with prolific tweeters, some of whom might be considered on the risqué side of brand suitability.

“I still recommend that brands do not spend ad dollars there,” said one top agency exec within a major holding company, speaking on condition of anonymity. This person said that the 2,000-keyword limit on exclusion lists was a “drop in the ocean.” “New accounts are popping up by the minute, so once you block one, there’s another.”

While Twitter declined to comment, a spokesperson pointed to the extensive brand safety roadmap that Twitter has put forth, which includes a plan for “pre-bid” controls. “Pre-bid” controls would be an important tool because they help advertisers avoid unsuitable placements before a campaign runs and don’t rely solely on reports about whether campaigns avoided offensive material after the fact.