Twitch, Amazon’s gaming focused live streaming platform, found itself in hot water with advertisers this month, after a number of brands took issue with the prevalence of content which they considered to be explicitly sexual.
The issue has shone a fresh light on the difficulties for user-generated content (UGC) platforms when it comes to enforcing community guidelines, due to the inherent subjectivity of terms like “sexually suggestive”. Where there’s wriggle room within community guidelines, it becomes harder for advertisers to fully understand the kinds of content their ads will appear alongside.
An abundance of aubergine emojis
Twitch prohibits sexually suggestive content within its community guidelines. And while it acknowledges that there is subjectivity within this term, it provides a number of examples of what might constitute sexually explicit content. These include ‘content or camera focus on breasts, buttocks, or pelvic region, including poses that deliberately highlight these elements’, ‘fetishizing behaviour or activity, such as focusing on body parts for sexual gratification or erotic role play’, and ‘erotic dances, such as those involving stripping or flashing’.
The platform also has rules on attire, which essentially prohibit streamers from wearing particularly revealing clothing.
The recent issues stemmed in part from exceptions to these rules on attire, which allow streamers to wear clothes appropriate to their real world setting. This allows streamers to wear swimwear in appropriate contexts – for example on a beach, or in a hot tub.
These rules have been in place for a long time, but the last few months saw an explosion in so called ‘hot tub streams’, where streamers would wear swimsuits and film from hot tubs or paddling pools, usually while chatting with viewers. These streams became so popular that Twitch users declared the platform had entered a ‘hot tub meta’ – meta being a gaming term which describes a tactic which is so successful it has become predominant.
These hot tub streams quickly became controversial within Twitch’s community, as streamers and viewers felt they fell foul of Twitch’s rules on sexually suggestive content. Many of these streams’ own titles suggested sexual content too, often including ‘18+’, ‘adult stream’, and peach/aubergine emojis.
And it seems brands took issue with it too. No major brands publicly complained about the hot tub streams. But Twitch indicated last month that advertisers had privately asked for change, saying that “community and advertiser feedback made clear that we need to offer more ways to control the content that’s recommended as well as where ads appear.”
In response, Twitch said it will continue to allow hot tub content on its platform, but requires it be hosted in a separate ‘Pools, Hot Tub, & Beaches’ category. Brands keen to avoid this content can simply choose to opt out of advertising on this new category.
Treading a thin line
Twitch, and other UGC platforms, have to tread a thin line around sexually suggestive content.
If rules are set too tightly, platforms might be seen as somewhat draconian and misogynistic. Twitch is clearly keen to avoid this perception, saying in its statement on hot tub streams that “being found to be sexy by others is not against our rules, and Twitch will not take enforcement action against women, or anyone on our service, for their perceived attractiveness”.
And some advertisers won’t have any problems with this sort of borderline content. The Global Alliance for Responsible Media (GARM), a coalition of industry players aimed at solving issues around harmful content. GARM distinguishes between content which should never be monetised, defined by its ‘brand safety floor’ guidelines, and content which has issues around brand suitability.
“From what I have seen, this is very much a brand suitability issue,” said Stevan Randjelovic, director of brand safety and digital risk in EMEA for GroupM. “Brand safety is about content which should never be monetised, brand suitability is about content which might be suitable for some, and not for others.”
But it seems many advertisers are keen to avoid it. Hot tub streams have all but died out since being forced into a separate category – suggesting advertisers have been steered clear of this content. At the time of writing, only two streams in the Pools, Hot Tubs, & Beaches category had more than a thousand viewers – and one of these was a live stream of an otter enclosure, run by Marine Mammals Rescue as a parody of the hot tub streams.
Theoretically, the way for UGC platforms to tread this line is to make it easy for brands to control the types of content their ads appear next to. This lets brands avoid any content they find to be inappropriate, without forcing those standards on the whole platform.
GARM is trying to help facilitate this, by creating a standardised brand suitability framework which separates content into high risk, medium risk, and low risk across eleven different content categories.
“Ultimately the idea behind GARM is to standardise this discussion,” said Neal Thurman, co-founder of the Brand Safety Institute. “The idea is to create levels of risk around a specific dimension of questionable content which allows marketers to set their preferences, and then use those same levers to control content across different platforms.”
Twitch has effectively taken this approach. By forcing this controversial content into a separate category, the platform has made it easier for brands to opt-out if they have concerns.
“[The Twitch hot tub content] has come up with a few advertisers,” said GroupM’s Randjelovic. “But the ability to exclude that content and manage the adjacency to it has generally addressed their suitability concerns.”
No silver bullet
But this approach isn’t perfect. Just as there is ambiguity in Twitch’s community guidelines, there is also an element of ambiguity in brand suitability standards. GARM defines high-risk adult content as containing “suggestive sexual situations requiring adult supervision/approval or warnings,” – it’s not fully clear whether Twitch’s hot tub content would fall within this category or not.
And it’s very hard to design any sort of standardised rules which can easily translate across platforms, and fit with each platform’s unique culture. Some of Twitch’s hot tub streamers, for example, claimed to be parodying others, or to be using tongue-in-cheek titles which weren’t reflective of their content.
“I think for everyone in the industry involved in these conversations, their jobs are really hard and I don’t think we should minimise that,” said the Brand Safety Institute’s Thurman. “There’s nuance to satire, there’s nuance to tongue-in-cheek content. And there’s nuance to what advertisers might want to avoid because they don’t like it, and what we think shouldn’t be monetised through advertising at all. None of those are easy questions, and I suspect we’ll be in for a bit of trial and error with developing tools to handle these things.”
We’ve seen some of this “trial and error” already on Twitch. Users have complained that hot tub streamers have simply moved to other sections of the site, namely the ‘ASMR’ section, to create similarly borderline content.
But while it’s likely to be an ongoing battle, GroupM’s Randjelovic said the tools available to advertisers are still valuable, and that it’s important that platforms work to make these tools as effective as possible.
“We are pushing all platforms to be effective in enforcing their community guidelines, and to provide more comprehensive controls to advertisers so they can exclude channels they don’t want to be next to, or to implement the GARM brand suitability standards, so clients can choose the level of risk they’re comfortable with,” he said.
