Cybersecurity researcher Daylam Tayari has found evidence in Twitch’s internal API that the site plans to implement something called a “brand safety score” for its streamers. That score would depend on a number of criteria: the streamer’s age, a rating given by Twitch staff, their ban history, the relationship the streamer has with Twitch, their automod settings, their partnership status, the ESRB rating of the game being played, and whether the stream is set to mature.
If Tayari’s research is indeed accurate — and here I should note that I have reached out to Twitch and will update this story if I hear back — this would represent a shift in the way that advertisers interact with streamers on the platform. Presently, one of the main ways that marketers work with streamers on Twitch is the site’s bounty board; select partners and affiliates in the US, UK, Germany, and France can choose from a list of paid opportunities to either play games or watch branded videos with their communities for automatic payouts. It’s a pretty slick system, one that streamlines and automates the (occasionally arduous) process of working with a brand and getting paid for it.
In a series of replies to Tayari’s tweets, Twitch confirmed that nothing has launched yet and that personal information had been shared — but also said that the company is “exploring ways to make sure ads are appropriately matched to the right communities on Twitch.”
It’s not hard to imagine that if Twitch does in fact implement a brand safety score for streamers that it would be used to expand the bounty program; it seems like a useful thing for brands to be able to compare streamers on that specific axis, at least. Then again, for streamers, it does mean the site is tracking you on yet another metric that may or may not be available for you to see. I can’t help but think about Uber and Lyft ratings, and how quietly insidious they can be: if your rating gets screwed up for whatever reason, you lose your ability to drive — to make money.