Twitch Sponsorship Segments Priced on Average Minute Audience

Peak concurrent viewers is the number a channel's media kit leads with, and it overstates the audience a sponsored segment actually reaches by two to three times. We price against average minute audience over the trailing thirty streams, pulled from public tracking and cross-checked against the streamer's own dashboard. A channel averaging 800 concurrent across four hours delivers more attention than one that spiked at 4,000 for eleven minutes during a raid. Why this is priced differently to everything else we run: Live is unrehearsed and cannot be recalled. There is no approval step between the sentence and the audience. So the vetting moves entirely to the front: we watch a minimum of six hours of a candidate's back catalogue before we approach, including at least two streams where something went wrong, because how a streamer handles a failing build or a hostile chat is the only preview available of how they will handle your product breaking on camera. The segment: Ten to fifteen minutes of integrated use, not a thirty-second read between games. For developer tools that means installing it live in the Software and Game Development category, in front of chat, with the failures left in. For family brands it means using the product the way a parent would and answering chat about it. The tracked link goes into a chat command through Nightbot or StreamElements and into the panel below the player, both live from the start of the stream and both left up for thirty days. The kill fee: Every deal we write carries a fee the streamer can take at fifty percent to walk away after using the product and deciding they do not want to endorse it. We tell them this in the first email. It triggers about one deal in eight. It is the reason the other seven are believed. What we will not ask a streamer to do: Read a script live. Hold a talking-points card. Cover chat criticism of the product. Guarantee a sentiment. Accept category exclusivity beyond the campaign. Play a game or use a tool they have said on stream that they dislike. Move their schedule to hit your launch date, because a stream pulled forward two days runs to a smaller and more irritated audience. What you should bring: Somebody from your team in chat for the duration, named, answering questions as themselves. An engineer in chat during a developer tools segment converts at a rate that nothing in the brief can match, and a channel notices when a brand sends nobody. Reporting and release: Average minute audience for the sponsored window specifically, not for the whole stream. Chat message rate during the segment against the stream's own baseline. Command redemptions. Clip creation, which is the nearest thing live has to a save. Promo code use and conversions from your analytics. VOD views at 7 and 30 days, since a developer tools segment keeps pulling through YouTube exports and clip channels for weeks. Escrow releases against delivered average minute audience, with a pro rata refund where the stream underdelivers. Not a fit: Brands that need a specific sentence spoken exactly. Anything requiring an audience under thirteen; Twitch's own minimum is thirteen and we buy adult and parent audiences only. Launch-day sponsorships bought as reach, where a single stream on a channel nobody in your category watches is money spent on a number.