- Dec 28, 2024
- 396
- 368
Interesting setup @djomar, you’ve built there... If you’re clearing around 2k with Apple and Tidal already, it sounds like you’ve already solved most of the hard parts of stream farming. I started in automation years ago as well and eventually ended up building my own tools, mostly focused on YouTube and Spotify, so I’ve spent quite a bit of time comparing how those ecosystems behave.
One thing that confuses a lot of people when moving from DSPs like Spotify, Apple or Tidal into platforms like YouTube or Amazon is that they don’t measure activity the same way.
With Spotify and most DSPs, the system mainly looks at listener behavior. As long as streams look like they come from real users with believable listening sessions, the system mostly evaluates things like completion rates, listener-to-stream ratios and long-term consistency. That’s why farms that are structured well can run for quite a long time if the patterns look natural.
YouTube works very differently.
YouTube is not really a streaming platform at its core it’s an engagement platform. The algorithm doesn’t care much about “streams” the way Spotify does. It cares about sessions. Things like how a viewer entered the video, how long they stayed on the platform, what they watched next, and whether the session creates more watch time for YouTube overall.
So if someone tries to apply a classic DSP farming strategy to YouTube, the numbers often look strange. You can generate views, but the payouts or growth don’t scale the same way, because the system is optimizing for session behavior rather than raw plays.
Amazon sits somewhere in between but with a different problem. Their royalty verification tends to look more at backlog data before payouts, which is why some people see streams counting during the month but royalties getting adjusted later. They’re very cautious when it comes to abnormal patterns across a catalog.
That’s also why a lot of operators who do well on Spotify or Apple feel like YouTube is unpredictable. It’s not necessarily harder it’s just a different game with different signals.
If your concern is protecting the catalog, the safest approach is usually treating YouTube less like a stream generator and more like a traffic ecosystem around the music. When sessions look like normal viewer behavior, the platform tends to be far less aggressive.
Out of curiosity when you tested YouTube before, were you focusing mostly on YouTube Music streams or regular video watch sessions? Because those two paths behave very differently in terms of both detection and payouts.
One thing that confuses a lot of people when moving from DSPs like Spotify, Apple or Tidal into platforms like YouTube or Amazon is that they don’t measure activity the same way.
With Spotify and most DSPs, the system mainly looks at listener behavior. As long as streams look like they come from real users with believable listening sessions, the system mostly evaluates things like completion rates, listener-to-stream ratios and long-term consistency. That’s why farms that are structured well can run for quite a long time if the patterns look natural.
YouTube works very differently.
YouTube is not really a streaming platform at its core it’s an engagement platform. The algorithm doesn’t care much about “streams” the way Spotify does. It cares about sessions. Things like how a viewer entered the video, how long they stayed on the platform, what they watched next, and whether the session creates more watch time for YouTube overall.
So if someone tries to apply a classic DSP farming strategy to YouTube, the numbers often look strange. You can generate views, but the payouts or growth don’t scale the same way, because the system is optimizing for session behavior rather than raw plays.
Amazon sits somewhere in between but with a different problem. Their royalty verification tends to look more at backlog data before payouts, which is why some people see streams counting during the month but royalties getting adjusted later. They’re very cautious when it comes to abnormal patterns across a catalog.
That’s also why a lot of operators who do well on Spotify or Apple feel like YouTube is unpredictable. It’s not necessarily harder it’s just a different game with different signals.
If your concern is protecting the catalog, the safest approach is usually treating YouTube less like a stream generator and more like a traffic ecosystem around the music. When sessions look like normal viewer behavior, the platform tends to be far less aggressive.
Out of curiosity when you tested YouTube before, were you focusing mostly on YouTube Music streams or regular video watch sessions? Because those two paths behave very differently in terms of both detection and payouts.
