Algorithmic Culture and Big Data: Spotify’s recommendation algorithm

By:
On: September 28, 2020
Print Friendly, PDF & Email
About


   

Have you ever listened to music on Spotify and wondered how it is possible that the recommended music perfectly suits your personal taste and sometimes even introduces you to music you didn’t even know you enjoy? Most people don’t pay attention to these defining features of streaming services and underestimate the importance of critically engaging with this new technology. The last years have shown that the usage of cloud music services has rapidly increased, introducing a new era of music listening. Today, Spotify is the world’s most popular audio streaming subscription service with 299m users, including 138m subscribers, across 92 markets (Spotify 1). With over approximately 150 million active users, Spotify is the global leader in music streaming services and has had an enormous impact on the way we listen and find music online. Within her work Werner explains that Spotify is more than simply a service that offers music, but rather functions as ‘’an intricate network of music recommendations governed by algorithms, displayed as a visual interface of photos, text, clickable links, and graphics’’ (78). This shows how Spotify’s recommendation algorithm is an emerging new media object that needs to be examined when looking at contemporary new media debates around topics like big data and privacy. The platform is well-known for its algorithms that customize the listening experience based on a variety of factors including likes, playlist formulation, listening history, and amount of time spent on a song (Gershgorn 1). The question raises how this algorithmic culture is regulated and impacts the users freedom of choice and privacy. As far as I’m concerned the cultural importance of Spotify’s algorithms is not to underestimate. 

If you used Spotify before you will likely be familiar with the ‘Discover Weekly’, a playlist with personalised tracks that is presented to the users every beginning of the week. While many competitors have promised to offer a recommendation algorithm that can match Spotify’s , none have come close to the quality of Spotify’s Discover Weekly list. (BBC 1). This is one of the main reasons for Spotify’s success as more than only being a service for streaming music the platform developed into a hybrid service that offers a platform to discover and listen to new and old music. 

Prey explains that Spotify significantly improved its music data analytics capabilities when it purchased The Echo Nest in 2014. The company utilizes acoustic analysis software to process and classify music according to a range of aural factors and analyzes the audio files in order to understand every single event in the song, such as a note in a guitar solo or the way in which two notes are connected’, explained Brian Whitman, co-founder and CTO of the company (Prey 1090). At the same time, the company conducts semantic analysis of daily online conversations about music all over the world in an attempt to turn both dialogs about music, and the music itself, into quantifiable data. (1091). It already becomes clear that efficiently working with music listeners’s data is at the heart of Spotify’s success.

Furthermore,The Echo Nest developed an analytics and visualization tool called the ‘Taste Profile In which every interaction between the user and music, including the listener’s music tastes and music behavior is recorded and saved in real time. The Discover Weekly’ is built around the Taste Profile, but ultimately functions as a hybrid recommender system as it combines content-based filtering of the Taste Profile with collaborative filtering methods (Prey 1091).

Spotify’s taste profile visualized

While Discover Weekly includes both professionally curated playlists and user-generated playlists in its algorithm, it gives preference to playlists with more followers and to playlists that are made by Spotify (Prey 1091). This already shows how the algorithm has tendencies to discriminate and prefer certain playlists and tracks over others. Furthermore, this reveals that Spotify has the power to dictate what does and what doesn’t get recommended to you, ultimately possessing the power over the listeners data. This issue is reflected in critical debates about the impact of big data and algorithmic culture. Within their work ‘ Critical Questions for Big Data, Danah Boyd & Kate Crawford explain that because our contemporary world increasingly relies on automation of data collection and analysis, as well as algorithms that can quantify and illustrate large-scale patterns in human behavior ,it is now the time to ask which systems are driving these practices and which are regulating them. (646). Clearly, the issue of automated music recommendation raises a range of philosophical problems and offers a breeding ground for critique of technology in culture.

Ann Werner supports this idea by revealing that Spotify materially organizes gender, nationality, and race in music culture through connections in the interface, and discursively by representations of artists and genres (88). This means that the algorithm actively contributes to reconstructing dominating genres like rock music as male-focused, masculine, and white, further enhancing already common gendering of music genres. This supports the idea that algorithmic culture within Spotify offers points of critique concerning the algorithms discrimination and how it impacts our contemprorary online culture.

 Undoubtedly, data analytics have enabled businesses to better predict a variety of consumer behaviors while the increasing dependence on Big Data for business decision-making has led to consumer privacy concerns that require strict business compliance ( Yang & Kang, 89), which is reflected in Spotify. Im am convinced that the recommendation algorithm is on one hand a great tool for discovering new and underappreciated music while on the other hand raises questions about privacy and what and what is not recommended. This is reflected by Boyd and Crawford when they claim that on one side, Big Data is seen as a ’’ powerful tool to address various societal issues, offering the potential of new insights into areas as diverse as cancer research, terrorism, and climate change. (664) Spotify has enabled millions to access old and new music and has as many say ‘’ saved the music industry’’ during the age of digitalisation while at the same time extending their platform to more than just a streaming service. Whilst Spotify undoubtedly had a positive impact for the listener’s experience, critique of Spotify’s policies, especially their payment of artists as well as privacy concerns are issues that are addressed frequently. Boyd and Crawford explain that on the other hand ‘’ Big Data is seen as a troubling manifestation of Big Brother, enabling invasions of privacy, decreased civil freedoms, and increased state and corporate control (664), which can directly be applied to Spotify’s case. In 2015 Spotify faced a heavy backlash and harsh critique after introducing new terms of service that include information stored on users mobile devices, such as photos, contacts and other media files ( Mason, 1), which emphasises how the company has to rely on monetizing consumer data. Concluding this shows how Spotify is a great example to investigate when concerning debates around Big Data, Privacy and the impact algorithmic culture has on our everyday lives. 

Comments are closed.