The last two weeks, we have been talking about the importance of metadata, and about how to use AI as a tool to enrich your digital music files with the correct metadata (see How to save time with AI). This approach helps to get the music in the right ears and the royalties in the right pockets. But these features can be used by both world-famous artists and independent musicians and will not necessarily provide any advantage to independent artists. Nonetheless, there is something very cool AI can do that can benefit the not-yet-famous artists in particular.
As you might know, AI based algorithms such as Spotify's one not only consider objective metadata information (music style, instruments, language, type of vocals, moods, etc.) for music suggestions. They also take into account the number of streams the track has and audience listening preferences. In an ideal world, works of excellent quality and vibe that many people would like would be the most streamed ones. But this is not the case in the real world.
The amount of money that an artist can invest in promoting their music (directly coming from the artist’s pocket, manager, publisher, or label they are signed to) makes a significant difference in the number of times that their music is played. There are thousands of incredible songs lost on the streaming platforms that many listeners would love... if they could only find them.
This working pattern of the algorithm is just a choice of the platform and not a requirement for the music suggestions. The AI doesn't need to consider the number of streams and the listenership information to make suggestions to the users. It can be trained to only account for objective data from the track, such as music style, groove, type of instruments, overall vibe, etc.
If the former were the case, a debuting independent artist would be given the same priority as a big music star if both artists' music has a significant number of traits in common. No more being hidden in the background of the streaming platform: if a listener is searching for your kind of music, they would get it suggested! This is an unexploited potential of AI in the democratization of music discovery. Wouldn't it be awesome if listeners could select the type of algorithm they want in order to get their music suggestions?
In musicube we believe in that, and we are tirelessly working on that path: No wonder our motto is, and will always be: "Good music deserves to be found."
Would you jump in with us?