Clip Farms are the New Street Teams: How to Game the Algo as an Indie Artist
Including Cadence Weapon as the primary video here, since his substack article was the inspiration for this unusually lengthy post.
Including Cadence Weapon as the primary video here, since his substack article was the inspiration for this unusually lengthy post. Linked further down below.
If you spent any time on music Twitter or Bluesky in April 2026, you saw the discourse. A Brooklyn indie rock band called Geese - good band I'm told by folks more fluent in the genre, I'm certainly not - became the unlikely lightning rod for a conversation that had been building for a while: what does it mean when the virality you're seeing isn't real?
The short version: in March 2026, the co-founders of a digital marketing agency called Chaotic Good sat down with Billboard's On the Record podcast at SXSW and more or less admitted, on tape, to manufacturing fake internet buzz for their artist clients. They called it "trend simulation." They said things like "everything on the internet is fake" with the casual confidence of people who had decided this was a feature rather than a bug. Within weeks, a widely-shared Substack post by singer-songwriter Eliza McLamb had connected the dots between Chaotic Good's methods and Geese's TikTok ubiquity, Wired had published a story with a predictably grabby headline, and the music internet had done what the music internet does, which is argue at high volume for a week and then move on.
But the underlying question - how do independent artists, and for the audience of this site specifically independent hip-hop artists, compete in an algorithmic environment where the playing field is actively being gamed by people with money - didn't go anywhere. That one's worth sitting with.
What Chaotic Good Actually Does
Jesse Coren and Andrew Spelman founded Chaotic Good in February 2025, coming out of years of managing indie-pop artists through their company Mutual Friends. The service they built is, depending on your perspective, either a sophisticated form of digital PR or a bot farm in a peacoat. Possibly both.
The mechanics, as they described them to Billboard: Chaotic Good creates networks of social media pages - hundreds or thousands of accounts, primarily on TikTok - and uses them to flood the zone with content featuring a client's music. Songs get dropped into the backgrounds of videos. Live clips get shared. Comments get posted. The goal is to generate enough volume across enough accounts that TikTok's algorithm registers "this artist" as a trending topic and starts pushing them into real users' For You Pages. Once the algorithm catches, real engagement follows because the algorithm is now doing the work of showing the music to people who might actually like it. Spelman called it "trend simulation," which is a tech investory-friendly name for what is essentially coordinated inauthenticity.
Their client list, as disclosed at SXSW, includes Travis Scott, Childish Gambino, Mitski, Tame Impala, Coldplay, and Zara Larsson. Geese and Cameron Winter were specifically discussed as artists whose TikTok presence was amplified through what Chaotic Good described as "distributing clips of them performing and doing some interviews." They denied falsifying streaming numbers in either case claiming that the virality was manufactured, but the streams that followed were real people making real choices after the algorithm served them the content.
This matters for hip-hop for an obvious reason: Travis Scott is on that client list. So is Childish Gambino. These are not obscure artists who needed a leg up. They are established, commercially successful rappers with major label infrastructure behind them. The clip farm isn't a last resort, it's apparently become a standard tool of the industry at every level.
Hip-Hop Has Always Understood Manufactured Buzz
It's worth pausing to note that this is not actually new. Hip-hop has a long history of manufactured virality it just used to involve different mechanisms. Street teams were clip farms for the pre-internet age: actual humans, hired to plaster stickers, hand out mixtapes, and create the impression of a groundswell where one was still being assembled. The Cash Money and No Limit albums of the late '90s weren't blanketing radio because DJs independently decided to play them, rather there were people making calls and writing checks. The "organic" discovery of a new artist has always involved more infrastructure than the mythology suggests.
What Chaotic Good sells is that infrastructure, optimized for the current algorithmic environment. The platforms have changed. The underlying transaction - you pay to simulate demand; if the music is good enough, real demand follows - hasn't.
Independent Canadian rapper Cadence Weapon, writing in his Substack about the Geese situation, put it plainly from the artist's perspective: "As an independent artist currently promoting a new album on a shoestring budget, learning about the jet fuel that other acts have access to can be troubling. Many of these bands already had big labels, booking agents and PR at their disposal before they started using performance-enhancing droids to take them to the next level." The irritation there is real and fair. But McLamb, the artist who originally surfaced the story, arrived at a more pragmatic place: "If a label wanted to contract Chaotic Good on my behalf, I would accept such a deal handily and gratefully."
What It Costs
Chaotic Good doesn't publish pricing, and they've been notably cagey about specifics since the controversy. Their "Narrative Campaign" page disappeared from their website within 24 hours of the discourse peaking. But you can triangulate from what the industry generally charges for comparable services.
A full-service digital marketing agency retainer runs $2,500 to $15,000 a month for ongoing work. Premium agencies with proven track records and celebrity client lists charge more. For a campaign-based engagement - a single release cycle, say, six to eight weeks - you're looking at $5,000 to $50,000 depending on the scope, the network of accounts being deployed, and whether the agency has relationships with real curators and influencers in addition to the manufactured activity. The clip farm infrastructure itself - the accounts, the content creation, the posting volume - is labor-intensive to build and maintain, which is reflected in the cost.
A more modest clip-farming operation, if you were running it yourself or using a smaller boutique service, could theoretically be done for less. Services that specialize in TikTok sound placement, getting your track into the background of lifestyle content, run $500 to $2,000 for a defined campaign. Playlist pitching through services like SubmitHub or Groover is $100 to $500 per campaign. None of this is "trend simulation" in the Chaotic Good sense, but it's the accessible end of the same spectrum: using money to put your music in front of people who didn't seek it out.
The honest number for an independent hip-hop artist with a real shot at the Chaotic Good experience is somewhere north of $5,000 per release cycle, and probably closer to $10,000 to $15,000 if you want something with actual teeth. For most independent artists working without a label, much less a kid recording tracks in his bedroom closet studio, it's probably not money they have.
What You Can Do With No Budget
The thing Chaotic Good is actually selling is volume, consistency, and pattern recognition. Their insight is that TikTok's algorithm responds to signals of momentum: lots of accounts posting content using the same audio, consistent engagement in a short window, the appearance of a thing that is already trending. The algorithm, like any pattern-matching system, can be gamed if you know what patterns it's matching.
The tools to do a version of this yourself exist, and some of them are free.
Content creation at volume. The bottleneck for most independent artists running their own social media is time and creative bandwidth. You can only make so many videos, write so many captions, and respond to so many comments before the whole operation collapses into an anxiety spiral. AI tools - Claude, Copilot, ChatGPT - remove some of that bottleneck. You can use them to generate caption variations, draft hooks for videos, repurpose interview quotes into short-form content, brainstorm video concepts for a specific song, and batch-produce a week's worth of social copy in a single sitting. This is not trend simulation, but it does solve the volume problem on the content side without requiring you to be a full-time content creator.
What this looks like in practice: you drop a new single, and instead of one TikTok and hoping for the best, you use Claude to help you build out fifteen different content concepts around the release: a clip framing the story behind the song, a text-over-video with a lyric that hits, a "this song is for people who..." format, a behind-the-scenes of the recording, a reaction to a reaction. You post them staggered over two weeks. You've created the appearance of ongoing momentum because there is ongoing momentum, even if the algorithm hasn't caught yet.
The burner account question. This is where independent artists have to make a decision about what they're comfortable with. Chaotic Good's approach involves accounts that aren't the artist's, third-party accounts posting content about the artist to simulate organic discovery. An independent artist could, in theory, do a version of this with help from their actual community: ask twenty friends to each post one TikTok using the new single this week, targeting a specific sound or format. This is not fake, these are real people who know you, but it is coordinated, and the result looks to the algorithm like the kind of distributed organic content that suggests a trend is forming. But as many of you have discovered yourselves, even getting your friends to take the split second to just thumbs up your clips is near impossible as it is.
Platform-specific strategy without paying for it. TikTok, Instagram Reels, and YouTube Shorts all respond to different content patterns. A tool like Claude can help you analyze what's actually working in your genre right now not by giving you proprietary data, but by helping you think through what formats and hooks are resonating, and then drafting scripts or outlines that put your music in that context. The creative research that an agency charges for ("what kind of content performs for this genre right now") is a conversation you can have with an AI assistant for free. It won't have real-time data, but it will help you think more systematically than you otherwise would.
The honest ceiling. What Chaotic Good does works (when it works) because of scale that an individual artist can't replicate. Thousands of accounts, posting constantly, flooding an algorithm until it tips: that requires infrastructure. The DIY version of this is closer to what LaRussell, the independent Vallejo rapper, has been doing: consistent content building slowly over time. LaRussell has millions of views because he posted constantly and built a community around the idea that the process itself was the content. That's a different philosophy than trend simulation, and it takes longer, but it's sustainable and it's real. It's just near impossible to do if you have a full time job or, god forbid!, a family to look after.
The Actual Question
For an independent hip-hop artist working in 2026, the takeaway from the Geese situation probably isn't "this is unfair," though it is unfair, or "Chaotic Good is evil" though they're at minimum morally complicated. The takeaway is that the algorithm is a system, systems can be understood and gamed, and understanding them is free. You can't outspend Travis Scott's marketing budget. You can learn to think like the people spending it, and do a smaller version of it with what you have.
The music still has to be there. That part was always true, and Chaotic Good can't fix it when it isn't. But assuming the music is real the system for getting it heard is more transparent than it's ever been. Use that.
Including Cadence Weapon as the primary video here, since his substack article was the inspiration for this unusually lengthy post. Linked further down below.
If you spent any time on music Twitter or Bluesky in April 2026, you saw the discourse. A Brooklyn indie rock band called Geese - good band I'm told by folks more fluent in the genre, I'm certainly not - became the unlikely lightning rod for a conversation that had been building for a while: what does it mean when the virality you're seeing isn't real?
The short version: in March 2026, the co-founders of a digital marketing agency called Chaotic Good sat down with Billboard's On the Record podcast at SXSW and more or less admitted, on tape, to manufacturing fake internet buzz for their artist clients. They called it "trend simulation." They said things like "everything on the internet is fake" with the casual confidence of people who had decided this was a feature rather than a bug. Within weeks, a widely-shared Substack post by singer-songwriter Eliza McLamb had connected the dots between Chaotic Good's methods and Geese's TikTok ubiquity, Wired had published a story with a predictably grabby headline, and the music internet had done what the music internet does, which is argue at high volume for a week and then move on.
But the underlying question - how do independent artists, and for the audience of this site specifically independent hip-hop artists, compete in an algorithmic environment where the playing field is actively being gamed by people with money - didn't go anywhere. That one's worth sitting with.
What Chaotic Good Actually Does
Jesse Coren and Andrew Spelman founded Chaotic Good in February 2025, coming out of years of managing indie-pop artists through their company Mutual Friends. The service they built is, depending on your perspective, either a sophisticated form of digital PR or a bot farm in a peacoat. Possibly both.
The mechanics, as they described them to Billboard: Chaotic Good creates networks of social media pages - hundreds or thousands of accounts, primarily on TikTok - and uses them to flood the zone with content featuring a client's music. Songs get dropped into the backgrounds of videos. Live clips get shared. Comments get posted. The goal is to generate enough volume across enough accounts that TikTok's algorithm registers "this artist" as a trending topic and starts pushing them into real users' For You Pages. Once the algorithm catches, real engagement follows because the algorithm is now doing the work of showing the music to people who might actually like it. Spelman called it "trend simulation," which is a tech investory-friendly name for what is essentially coordinated inauthenticity.
Their client list, as disclosed at SXSW, includes Travis Scott, Childish Gambino, Mitski, Tame Impala, Coldplay, and Zara Larsson. Geese and Cameron Winter were specifically discussed as artists whose TikTok presence was amplified through what Chaotic Good described as "distributing clips of them performing and doing some interviews." They denied falsifying streaming numbers in either case claiming that the virality was manufactured, but the streams that followed were real people making real choices after the algorithm served them the content.
This matters for hip-hop for an obvious reason: Travis Scott is on that client list. So is Childish Gambino. These are not obscure artists who needed a leg up. They are established, commercially successful rappers with major label infrastructure behind them. The clip farm isn't a last resort, it's apparently become a standard tool of the industry at every level.
Hip-Hop Has Always Understood Manufactured Buzz
It's worth pausing to note that this is not actually new. Hip-hop has a long history of manufactured virality it just used to involve different mechanisms. Street teams were clip farms for the pre-internet age: actual humans, hired to plaster stickers, hand out mixtapes, and create the impression of a groundswell where one was still being assembled. The Cash Money and No Limit albums of the late '90s weren't blanketing radio because DJs independently decided to play them, rather there were people making calls and writing checks. The "organic" discovery of a new artist has always involved more infrastructure than the mythology suggests.
What Chaotic Good sells is that infrastructure, optimized for the current algorithmic environment. The platforms have changed. The underlying transaction - you pay to simulate demand; if the music is good enough, real demand follows - hasn't.
Independent Canadian rapper Cadence Weapon, writing in his Substack about the Geese situation, put it plainly from the artist's perspective: "As an independent artist currently promoting a new album on a shoestring budget, learning about the jet fuel that other acts have access to can be troubling. Many of these bands already had big labels, booking agents and PR at their disposal before they started using performance-enhancing droids to take them to the next level." The irritation there is real and fair. But McLamb, the artist who originally surfaced the story, arrived at a more pragmatic place: "If a label wanted to contract Chaotic Good on my behalf, I would accept such a deal handily and gratefully."
What It Costs
Chaotic Good doesn't publish pricing, and they've been notably cagey about specifics since the controversy. Their "Narrative Campaign" page disappeared from their website within 24 hours of the discourse peaking. But you can triangulate from what the industry generally charges for comparable services.
A full-service digital marketing agency retainer runs $2,500 to $15,000 a month for ongoing work. Premium agencies with proven track records and celebrity client lists charge more. For a campaign-based engagement - a single release cycle, say, six to eight weeks - you're looking at $5,000 to $50,000 depending on the scope, the network of accounts being deployed, and whether the agency has relationships with real curators and influencers in addition to the manufactured activity. The clip farm infrastructure itself - the accounts, the content creation, the posting volume - is labor-intensive to build and maintain, which is reflected in the cost.
A more modest clip-farming operation, if you were running it yourself or using a smaller boutique service, could theoretically be done for less. Services that specialize in TikTok sound placement, getting your track into the background of lifestyle content, run $500 to $2,000 for a defined campaign. Playlist pitching through services like SubmitHub or Groover is $100 to $500 per campaign. None of this is "trend simulation" in the Chaotic Good sense, but it's the accessible end of the same spectrum: using money to put your music in front of people who didn't seek it out.
The honest number for an independent hip-hop artist with a real shot at the Chaotic Good experience is somewhere north of $5,000 per release cycle, and probably closer to $10,000 to $15,000 if you want something with actual teeth. For most independent artists working without a label, much less a kid recording tracks in his bedroom closet studio, it's probably not money they have.
What You Can Do With No Budget
The thing Chaotic Good is actually selling is volume, consistency, and pattern recognition. Their insight is that TikTok's algorithm responds to signals of momentum: lots of accounts posting content using the same audio, consistent engagement in a short window, the appearance of a thing that is already trending. The algorithm, like any pattern-matching system, can be gamed if you know what patterns it's matching.
The tools to do a version of this yourself exist, and some of them are free.
Content creation at volume. The bottleneck for most independent artists running their own social media is time and creative bandwidth. You can only make so many videos, write so many captions, and respond to so many comments before the whole operation collapses into an anxiety spiral. AI tools - Claude, Copilot, ChatGPT - remove some of that bottleneck. You can use them to generate caption variations, draft hooks for videos, repurpose interview quotes into short-form content, brainstorm video concepts for a specific song, and batch-produce a week's worth of social copy in a single sitting. This is not trend simulation, but it does solve the volume problem on the content side without requiring you to be a full-time content creator.
What this looks like in practice: you drop a new single, and instead of one TikTok and hoping for the best, you use Claude to help you build out fifteen different content concepts around the release: a clip framing the story behind the song, a text-over-video with a lyric that hits, a "this song is for people who..." format, a behind-the-scenes of the recording, a reaction to a reaction. You post them staggered over two weeks. You've created the appearance of ongoing momentum because there is ongoing momentum, even if the algorithm hasn't caught yet.
The burner account question. This is where independent artists have to make a decision about what they're comfortable with. Chaotic Good's approach involves accounts that aren't the artist's, third-party accounts posting content about the artist to simulate organic discovery. An independent artist could, in theory, do a version of this with help from their actual community: ask twenty friends to each post one TikTok using the new single this week, targeting a specific sound or format. This is not fake, these are real people who know you, but it is coordinated, and the result looks to the algorithm like the kind of distributed organic content that suggests a trend is forming. But as many of you have discovered yourselves, even getting your friends to take the split second to just thumbs up your clips is near impossible as it is.
Platform-specific strategy without paying for it. TikTok, Instagram Reels, and YouTube Shorts all respond to different content patterns. A tool like Claude can help you analyze what's actually working in your genre right now not by giving you proprietary data, but by helping you think through what formats and hooks are resonating, and then drafting scripts or outlines that put your music in that context. The creative research that an agency charges for ("what kind of content performs for this genre right now") is a conversation you can have with an AI assistant for free. It won't have real-time data, but it will help you think more systematically than you otherwise would.
The honest ceiling. What Chaotic Good does works (when it works) because of scale that an individual artist can't replicate. Thousands of accounts, posting constantly, flooding an algorithm until it tips: that requires infrastructure. The DIY version of this is closer to what LaRussell, the independent Vallejo rapper, has been doing: consistent content building slowly over time. LaRussell has millions of views because he posted constantly and built a community around the idea that the process itself was the content. That's a different philosophy than trend simulation, and it takes longer, but it's sustainable and it's real. It's just near impossible to do if you have a full time job or, god forbid!, a family to look after.
The Actual Question
For an independent hip-hop artist working in 2026, the takeaway from the Geese situation probably isn't "this is unfair," though it is unfair, or "Chaotic Good is evil" though they're at minimum morally complicated. The takeaway is that the algorithm is a system, systems can be understood and gamed, and understanding them is free. You can't outspend Travis Scott's marketing budget. You can learn to think like the people spending it, and do a smaller version of it with what you have.
The music still has to be there. That part was always true, and Chaotic Good can't fix it when it isn't. But assuming the music is real the system for getting it heard is more transparent than it's ever been. Use that.