Streaming service Twitch has come under fire recently due to changes it is making that could impact how creators earn revenue using the platform. The biggest change is adding an automatic Brand Safety Score to every streamer, which measures and rates them based on numerous factors such as chat behaviour, ban history, manual ratings from Twitch staff, and the games they play.
The Brand Safety Score was first noticed in Twitch’s code last week by Daylam ‘tayari’ Tayari, who explains the exact changes in a comprehensive Twitter thread. Despite the fact that Twitch responded to his thread explaining that it is “exploring ways to make sure ads are appropriately matched to the right communities,” and reassuring streamers that “nothing has launched yet,” many streamers and viewers are up in arms.
As of March 11, two days after tayari revealed the system, he explained that Twitch had “removed all elements” of the Brand Safety Score from their API. Tayari says the backlash was “unjustified,” and that the Brand Safety Score could have resulted in creators earning more.
We sat down with two of the UK’s biggest streamers, Simon ‘Blue’ and Rowena ‘Queenie’ Flynn, who stream together, to see how they feel the changes will impact their streams and their revenue.
Blue explains that, as Twitch Partners, the duo have an insight into new systems on the platform that sometimes even Twitch can’t talk about. “[People who work for Twitch] can’t always say what they feel,” he says. “Or they’re not allowed to explain any of these new things and why they’re there. So as Twitch Partners on the inside, it’s kind of a nice outlet for them and they trust us to explain this.”
As for the Brand Safety Score, the pair is “confident” that something similar is already in place but it might not be automated currently. “[Twitch has] rolled it back now, or at least they’ve made it hidden,” explains Blue. “But this is kind of something that they’ve been working on for a long time.”
Blue continues to explain that having an automated system, or at least a standardised set of rules that set brand-friendly channels apart from those that aren’t, seems like a sensible idea for a service like Twitch. “The whole idea makes sense to moderate that many streamers with only a small moderation team,” he says. “It kind of makes sense that they would have some background rating, some notes.”
Weird and wonderful: our favourite Twitch clips of 2020
As one of the top UK channels, Blue and Queenie focus on being brand friendly, but that is not necessarily the same as being family friendly. They don’t have any fears that they would have to water down their content in order to get brand sponsorships through Twitch – something that many of the Brand Safety Score opponents have warned may happen.
We were never going towards family friendly as such, but we're brand friendly,
Blue and Queenie tell The Loadout that they don’t stop anyone from swearing and they’re not afraid to talk about real world situations and news which may not be appropriate for children. However, at the same time they maintain that they aren’t controversial, which keeps their channel attractive to brands who may want to work with them. They’ve worked on their content and behind the scenes to maximise their partnerships, and don’t see the Brand Safety Score changing anything that Twitch’s Bounty Board system already does.
“We were never going towards family friendly as such, but we’re brand friendly,” says Blue. “We are one of the top UK channels and we’ve worked internally with Twitch as well to make that happen, so we’re on like an exclusive list to work with companies.”
Blue and Queenie maintain that, due to the fact that Twitch’s own revenue model is based on subscriptions, cheers, and bits, rather than relying on adverts like YouTube does, the platform will never push for family friendly content from its top streamers.
“Kids aren’t the ones subscribing,” says Blue. “They don’t have they don’t have the means to subscribe, so it makes no sense for Twitch to turn it into a family friendly [service].”
Twitch has unfortunately removed all elements of the Brand Safety Score from their GraphQL API.
Hopefully it is not a permanent change caused by the unjustified backlash from people completely misinterpreting that this was simply an ad targeting score. pic.twitter.com/WGnjoTWipo
— Daylam 'tayari' Tayari (@tayariCS) March 11, 2021
It makes sense that companies who want to advertise on streams don’t want their advert being placed alongside the stream of a controversial creator, because viewers may associate their product with that ban-worthy sentiment. Adding the Brand Safety Score to Twitch’s API seems like a way of standardising the criteria for a brand friendly channel across the board, rather than relying on manual reports which could allow inherent biases to creep in.
However, Blue and Queenie also think that the change to the coding could have been to aid discoverability, something that Twitch regularly experiments with.
“I think the coding change is something they’re working on to help discoverability,” says Blue. “Every fortnight Twitch has been changing the way the discoverability works, to trial how to improve it.”
Out and about: who are the best IRL streamers?
The streamers believe that not only will having a good Brand Safety Score – which may or may not still be in place – improve your chances of getting ad sponsorships and brand partnerships, but it may also give you a better shot at being hosted as a “recommended streamer” on Twitch’s front page.
Twitch’s changes are positive
As for the future of advertising on the platform, Blue and Queenie are very happy with the way things are going. As streamers who regularly work with brands off their own backs, a recent promotion with food delivery service Just Eat – organised through Twitch – was effortless in comparison. “It was pretty straightforward and really easy,” says Queenie. But the streamers acknowledge that while the rules they had to follow were “regimented,” their trial of a new targeted advertising system on the platform was an overwhelming success.
“Twitch basically sorted out everything for us,” she explains. The service lifted their analytics and sent them over to Just Eat, and the pair assume that they were chosen because of their commitment to brand safety, and the fact that they have worked with Twitch before on different advertising rollouts. As it happened, Just Eat gifted 100 subscriptions during their pizza-eating stream, and Blue and Queenie hung out with about 50 subs on Discord afterwards, all eating Just Eat-ordered meals.
“We were chosen because we were a safe bet,” Blue adds. Queenie says that she worries that brands may stereotype her because she is Northern Irish, and they may be hesitant to work with her in case she swears. But she believes that Twitch reassured Just Eat, likely with an overview of the channel’s history of positive, brand-friendly interactions, and that helped assuage any fears the delivery company may have had.
However, the streamers were also surprised by chat’s adverse reaction to auto-mod, software which automatically moderates the chat, and something that they needed to have on its strictest level for their stream with Just Eat. “I’m surprised by how many people really condemned having auto mod on as a plus point to being brand friendly in this algorithm,” says Blue.
Again, sponsors and advertisers want to be assured that their brands will not be associated with anything that could harm their reputation. Using brand-friendly streamers is one part of that, and having a keenly moderated Twitch chat is another. If viewers want the creators they follow to earn a more sustainable revenue, then perhaps they should get behind the changes that reward streamers for brand-friendly content that will likely result in more, and better, deals.
“Twitch’s changes are positive, it’s better than them not doing anything,” says Blue. “And most likely, the thing from the coding is just for appearances, and has been happening anyway.”
Additional reporting by Aaron Down.