about leaving Spotify, and the music industry's futurist visions

    by Judge Russell

    February 06 2022

  • None

    Around the turn of the century, the advent of the internet was being regarded as the "death of the music industry". Record companies were so afraid of what would happen to them that they took down filesharing sites and even went as far as suing individuals for music piracy. Only the artists, and even then, only some of them, were the ones who were open to embracing the internet. In a 2003 article from the New York Times, Moby, System of a Down, Public Enemy, and Ween are listed as being against the policing efforts of major labels against filesharers, and of course, the radical self-publishing of Radiohead's In Rainbows is remembered fondly. But as the dust settled and filesharing lost out to legal downloads, the power of the remaining record companies shrunk as that of digital distributors grew immensely. Their path first rocketed through pay-per-download stores like iTunes, then played with the concept of limited streaming with services like Pandora Radio, and then stumbled upon the big money maker -- unlimited streaming services like Spotify and its competitors. Now, tech companies are the ones who control how we listen to recorded music.


    Streaming services, similar to Amazon, provide a service so convenient for the customer and worthwhile for the company, that it causes severe damage to almost everyone else. For years, Amazon's storefront was operating at a loss while it took control of the market with its convenience and vast product catalog, harming brick and mortar stores who weren't able to compete worldwide. In the realm of music, for example, Spotify was also not turning a profit until recently, operating at a loss as it transformed into virtually the only way people listen to music. Now, music that is not on Spotify is invisible to many music listeners. Of course, there are people like me who bought a 40 dollar Chinese MP3 player and a 128 GB SD card, but even I put my own music on Spotify, knowing that the convenience of having my music so accessible will actually get people to hear it. This is why the fiasco started by Neil Young this January is so interesting -- it serves as a catalyst for a solution to the Molochian problem of streaming services. Here's a definition of "Molochian problem" by Leslie Byvivre Brooks.


        "There exist many situations in which every single person prefers some different, better outcome, but somehow we can’t cooperate to get there, because every action that an individual can unilaterally take towards a greater outcome, only serves to get them removed from the game."


    Every artist would prefer to not be making fractions of cents from Spotify, yet to remove yourself is to become invisible. However, after Neil Young left Spotify, Joni Mitchell left, and then (probably) hundreds to thousands of much smaller artists followed suit. Indie rock band Ovlov wrote on Twitter recently:


        "i never really wanted to have it up there in the first place but was convinced it was "just the way it is now." but now that neil did it, seems like it may even be a good press move for us!"


    However, while I will suggest to consumers to cancel your streaming subscription and download Soulseek, I understand that this doesn't fix the situation at all for artists. Many artists, including myself, think Bandcamp is pretty good -- the portion of the profits they take isn't so high, and they constantly are promoting small artists with their daily articles and sales  reported on the front page in real-time. Others suggest "ethical" streaming services, which may or may not come to exist in the following years.


    However, I fear that artists will not come out of this moment on top, and that music distribution will continue to veer into stranger and more exploitative territory. A Linkedin blog post titled "How to Win the Music Industry of 2019-2025" by Rory Felton, one of the people behind, the site that sold NFTs of thousands of musicians' work without permission, illuminates one of the industry's futurist visions.


    In the piece, Felton champions a yet-non-existent company that will provide digital distribution, "unrivaled cross-platform data analytics", a record-label-like "partnership tier", and "blockchain digital collectible based fan clubs". He believes that this company's strongest service will be its data analytics, which will power an auto-pilot mode for artists, using artifical intelligence to decide the best times to release music or which of the world's markets to cater to. Furthermore, there will also be an artificial intelligence system that can identify "stars" and "hits" before the rest of the industry does. Lastly, he believes this system will be beneficial for artists because of money generated by "brand partnerships", by utilizing corporations' capital to fund artists rather than that of a label. He ends the article with this statement:


        "This company will have the greatest opportunity to carve out a large percentage of the rights to all new music created and build a substantial catalog to [sic] tens of millions of copyrights."


    This shows that the main goal here is copyright ownership, and the money here is in the royalties. This imagined platform has the distribution of Spotify, but with the royalty-collecting power of a label; therefore, this business would be far more profitable than a regular streaming service. Again, until recently, Spotify was operating at a loss. If this kind of company comes into existence, they will be more powerful than any label and more powerful than any streaming service. So, while he is correct that a business model like this would "win" the music industry, it would be a disaster for labels and independent artists. How could they compete with a label holding millions of copyrights supported by an unparalleled AI marketing system? It's easy to imagine artists feeling the same way about this company that they do about Spotify, that without joining, their work will never be heard.


    The AI system is also cause for concern. While deep learning predictive algorithms consistently perform wonders in all sorts of fields, they are also bogged down by issues present in the data, epitomized by the example of racial discrimination produced by AI predictors, for example, those that predict housing approval or risk of illness. Felton imagines the hit-predictor AI system would "analyze audio quality, artist performance quality, social, and streaming metrics" to decide which tracks are on the verge of blowing up. However, the inner workings of such an algorithm would be at risk of the biases present in its data. What about successful outliers? Could an AI algorithm trained on 2012 Billboard hits have predicted the popularity of Baauer's Harlem Shake? Could an algorithm trained on pre-Tiktok hits have predicted the newfound popularity of decades-old songs such as Fleetwood Mac's "Dreams" or Jeanette's "El muchacho de los ojos tristes", or "Castaways" from 00s children's TV show The Backyardigans? Of course, the analytics and machine learning teams at successful tech companies are always hard at work updating their models, and so eventually these aberrations will be factored into future decisions, but there is no knowing what music the future will hold. I do not doubt that a sufficiently trained hit-predictor would be able to discover hit songs. What I do doubt is the willingness of its implementors to allow it to choose the many offbeat and risky tracks that do become hits.


    In any case, successfully determining what songs will be popular before they become popular is a game-changer indeed. The artists that produce these tracks will, if unsigned, most likely sign to this company, who will then add the master copyright to their growing collection. If they are able to do this more swiftly and accurately than their competitors, their industry power will grow until it is unmatched. When this sort of model determines which artists get popular, we will see thousands attempt to game the algorithm, as we see happening with Youtube or Tiktok now. There is a whole genre of Youtube channel dedicated to instructing Youtubers on how to get their own channel recommended by the algorithm. By having a large amount of fame-chasers understanding that there is some way to game the system, no matter how arcane or unknowable that way is, the content is going to become homogenized. Whether the claims of algorithm gurus are accurate or not (it's often hard to tell due to the secrecy of platforms' recommendation algorithms), there will be desperate people who follow them to a T. Plus, with the result of the algorithm picking up an artist being a contract, and not just a recommendation, it's easy to imagine how this would be disastrous for pop music. Perhaps mid-tempo songs, songs with explicit lyrics, or songs shorter than 3 minutes 30 seconds but longer than 2 minutes 15 seconds are more likely to become hits. Not to say that this sort of thinking hasn't already begun; of course, Spotify's official highly curated playlists grant artists exposure, but if the stakes are raised as high as a record contract, artists looking to get noticed by the human or algorithmic curators will plunge the music industry deeper into the problem outlined by Liz Pelly in "The Problem with Muzak". In short, Pelly analyzes the ramifications of Spotify's activity playlists: playlists for chilling, exercising, or gaming. In an industry where music's purpose is to be put on in the background, musicians seeking to profit from this industry will conform to their "muzak" standards in order to be heard. An A&R team powered by deep learning will inevitably create a "sound" that artists will chase in pursuit of fame.


    In summary, despite Felton's current spotlight in music industry discussion stemming from his infamous NFT site (NFTs being based on blockchain's decentralized ledgers), he is heralding the future of the music industry to be hyper-centralized, more than even today. An all-in-one package of promotion, distribution, rightsholding, and data analytics may spell doom for labels of all sizes, and put artists in an even stickier situation than they're already in. The way out of this is unclear, but the mass backlash against Spotify and its ilk is a reminder that the low payouts from unlimited streaming services may not be the only way forward.





    Bowman, Jeremy. “Amazon Just Posted a Record Profit Margin: 3 Reasons It Will Get Even Better.” The Motley Fool, The Motley Fool, 2 May 2021, Accessed 2 Feb. 2022.


    Byvivre-Brooks, Leslie. “Rain World – Introduction: Reaching Enlightenment through Unfairness.” The Experienced Machine, 16 Sept. 2019, Accessed 2 Feb. 2022.


    Covington, Paul, et al. “Deep Neural Networks for YouTube Recommendations.” Proceedings of the 10th ACM Conference on Recommender Systems, 2016,


    Felton, Rory. “How to Win the Music Industry of 2019-2025.” LinkedIn, LinkedIn, 21 Mar. 2019, Accessed 2 Feb. 2022.


    NIHCM. “Racial Bias in Health Care Artificial Intelligence.” NIHCM, 30 Sept. 2021, Accessed 2 Feb. 2022.


    Ovlov. “I Am Gonna Finally Do What I Can to Get Anything of Mine Taken off Spotify Too, FTR I Never Really Wanted to Have It up There in the First Place but Was Convinced It Was ‘Just the Way It Is Now." but Now That Neil Did It, Seems like It May Even Be a Good Press Move for Us!” Twitter, Twitter, 31 Jan. 2022, Accessed 2 Feb. 2022.


    Pelly, Liz. “The Problem with Muzak.” The Baffler, 7 Dec. 2017, Accessed 2 Feb. 2022.


    Salzman, Avi. “Spotify Has Finally Found a Profitable Tune.” Barron's, Barrons, 27 Oct. 2021, Accessed 2 Feb. 2022.


    Sisario, Ben. “Joni Mitchell Plans to Follow Neil Young off Spotify, Citing 'Lies'.” The New York Times, The New York Times, 29 Jan. 2022, Accessed 2 Feb. 2022.


    Sisson, Patrick. “Housing Discrimination Goes High Tech.” Curbed, Curbed, 17 Dec. 2019, Accessed 2 Feb. 2022.


    Strauss, Neil. “File-Sharing Battle Leaves Musicians Caught in Middle.” New York Times, 14 Sept. 2003, Accessed 2 Feb. 2022.


Leave your comment!