Algorithmic music discovery: What’s behind this?

In today’s digital world, discovering new trends and developments in music largely relies on algorithms managed by streaming platforms such as Spotify or Pandora Internet Radio. These algorithms, based on data about individual user behavior and preferences, aim to deliver personalized music recommendations. This is not surprising, given that companies offer full customization of their services because the customer is paramount and their needs come first. However, a question arises: do algorithms genuinely help discover music, or do they primarily serve the interests of streaming platforms?

Personalization vs. commercialization

Streaming platforms offer personalization based on user data analysis. Algorithms process this data to match music to individual listeners’ tastes. Users often discover new music through algorithmic recommendations that are accurate and aligned with their preferences. Many appreciate the ability to create personalized playlists based on a single song or artist, leading to the discovery of new, similar performers.

Spotify describes the recommendation process as follows:


„At Spotify, we aim to create great and unique experiences for each user. Our goal is to connect everyone with what they love and help them discover something new. No two listeners are the same, so everyone’s Spotify experience, and many of our recommendations, are personalized. When asked what they like about Spotify, most listeners cite our personalization as their top feature. (…).
At Spotify, people and technology work together to deliver relevant recommendations. Some recommendations are based on editorial curation, like a pop playlist created by music editors. Other recommendations are tailored to each listener’s unique taste, like a personalized playlist powered by our expert-designed algorithms. (…).
Spotify offers algorithmic recommendations that are relevant, unique, and specific to each user. Our algorithms select and order content across each listener’s Spotify experience,

including in Search, Home, and in personalized playlists“.

At first glance, everything seems fine. We could say that Spotify cares about the listener and strives to deliver exactly what they want. However, Spotify does not fully automate the music selection process. It is a mix of editorial work and algorithms, introducing an additional layer of opacity.

Kitchin, R.

2017. Thinking Critically About and Researching Algorithms. Information,
Communication & Society 20 (1), p. 15.

„We are now entering an era of widespread algorithmic governance, wherein algorithms will play an ever-increasing role in the exercise of power, a means through which to auto­mate the disciplining and controlling of societies and to increase the efficiency of capital accumulation.”.

However, this process is not without criticism. Frank Pasquale, a professor of law with expertise in artificial intelligence (AI), algorithms, and machine learning law, is among the critics. In his book The Black Box Society: The Secret Algorithms that Control Money and Information (2015, Cambridge: Harvard University Press.).

Pasquale noted that algorithms are often “black boxes,” their operations hidden from users and regulated internally by private companies. We can observe data input and output, but we cannot say how one becomes the other. We are unsure of the potential spread of this information, its intended use, or its potential repercussions. Commercial interests, which aim to increase the platform’s profits by promoting certain content, can also drive the recommendation process. This introduces recommendations that are not free from the deeply hidden internal interests of the platform.

Have you heard of The Echo Nest?

It’s worth mentioning a significant event here. In 2014, Spotify made a significant investment in the importance of data analytics in the streaming sector. The platform acquired “The Echo Nest,” a little-known startup, for $58 million. This company was (and is) a leading firm in music intelligence. Herein lies the key to gathering data on music listening.

The Echo Nest has a knowledge base that includes over a trillion data points, about 37 million tracks, and 3.3 million artists. The company processes and classifies music according to many acoustic factors, from pitch and tempo to danceability. The system downloads and analyzes every mp3 file, paying attention to every single “event” in a track, such as a note in a guitar solo or the way two notes are connected. An average song contains about 2000 such “events” that the system can analyze. It then creates connections between this song and other songs with similar progressions or structures.

Pasquale, F.

2015. The Black Box Society. The Secret Algorithms That Control Money and Information,
London, Harvard Univesity Press, pp. 3-4.

„Knowledge is power. To scrutinize others while avoiding scrutiny oneself is one of the most important forms of power.8 Firms seek out intimate details of potential customers’ and employees’ lives, but give regulators as little information as they possibly can about their own statistics and procedures.9 Internet companies collect more and more data on their users but fight regulations that would let those same users exercise some control over the resulting digital dossiers“.

In addition to analyzing each track, The Echo Nest analyzes online conversations about music—millions of blog posts, music reviews, tweets, and discussions on social media. The Echo Nest platform collects keywords from music and creator descriptions and links them to other artists and tracks described with similar words and key phrases. These data help determine the similarities between tracks on a more cultural level.

Modern streaming services track and feed into their algorithms every song we listen to, every song we skip, and every thumbs up or down.

After mapping the world of music, there remains only one task: to determine where each listener fits on this map and their individual movements within the musical space. Thus, in real time, the Echo Nest algorithms identify you as a listener, what type of music fan you are, and define your musical preferences (artists and songs) and musical behaviors (favorites, ratings, skips). This is called a “Taste Profile.”

How does this relate to each of us as listeners? It’s very simple: all analyses have practical applications in song or artist recommendations. That would be fine if it weren’t for ad targeting. If Spotify knows what type of listener you are, it also knows very well what kind of ads the platform should show you! This is because the vast majority of Spotify listeners choose the “free” version with ads.

Pasquale, F.

2015. The Black Box Society. The Secret Algorithms That Control Money and Information,
London, Harvard University Press, pp. 61.

„«Better user experience» is the reason the major Internet companies give for almost everything they do. But surely their interests must conflict with ours sometimes— and then what? Disputes over bias and abuse of power have embroiled most of the important Internet platforms, despite the aura of neutrality they cultivate so carefully. It would be reassuring to have clear answers about when conflicts happen and how they’re handled. But the huge companies resist meaningful disclosure, and hide important decisions behind technology, and boilerplate contracts. What happens, happens out of our sight.”

As early as 2013, when The Echo Nest launched the “Music Audience Understanding” service, the company’s CEO, Jim Lucchese, said the company was transitioning from “What’s the next song you want to hear?” to “What ads do you need, which you are likely to respond to?” Nothing seems to be free, and there is a strong connection between the two. Was Brian Whitman, co-founder of The Echo Nest, right when he said, “Music preferences can predict more about you than anything else“?

Moving on, if streaming platforms can identify statistically significant relationships between musical tastes and non-musical information (the listener’s age, gender, and dozens of lifestyle categories), this directly translates to ad targeting. This, in turn, translates to the ability to increase the rates at which advertisers have the right to access their listeners.

We can ask: What is the purpose of algorithmizing music? Every coin has two sides. When it comes to streaming platforms, it is no different!

Conclusion: The Future of Music and Algorithm Management

Algorithms have undoubtedly revolutionized the way we discover and consume music. Their role will grow in the future, but it is important to maintain a balance between personalization and commercialization. Transparency in algorithmic operation and greater control by users over their data can contribute to a more authentic discovery of music.

Spotify assures:


Spotify prioritizes listener satisfaction when recommending content. In some cases, commercial considerations, such as the cost of content or whether we can monetize it, may influence our recommendations“.

But can we trust these words, considering the operation of the platform’s tracking and behavior analysis tools? Are we only talking about “some cases”? Or is each of us’s commercialization the key to algorithmizing music?

In the context of the future of music, it will be crucial to understand and accept the role that technology plays in our daily musical lives. If we use algorithmic recommendations ethically and prioritize the listeners’ needs over the commercial interests of profit-driven streaming platforms, they can prove to be a valuable tool.

Based on: