Music Stream – Algorithmic Individualization of Masses

Today, when digital methods “strive to follow the evolving methods of the medium” (Rogers 2013:1), music streaming becomes the largest segment of global recorded music industry revenue. Modern music streaming services enable customers to access a huge library of songs and compositions. Spotify is the epitome of streaming’s digital Zeitgeist, shaping the future. With over 60 million subscribers, this music, podcast and video streaming service is considered to be a significant success. Founded as a technological solution for record companies, struggling with piracy, Spotify made music free but legal, available to consumers at no cost, while advertising provided all revenues (or for a fixed monthly payment). It’s both a tech and a media company: first only distributing content produced by others (Fleischer, Snickars. 2017: 133), now Spotify is acting more “as a consumer subscription businesses in media” (Mähler, Vonderau. 2017: 212). From the consumers’ experience, Spotify can be as a medium for the transmission of information from one location to another (Hagen. 2016); a music storage; a content platform; a place to “hang out”; or environment creating atmosphere. Music streaming also can be as a way of being socially (creating a social network of users); and personally, intimately (building your own identity, values and world view) (Sinclair, Green. 2015: 15). Therefore the increasing popularity of legal streaming platforms has an impact on the consumer experience in regards to cultural identity, communication, bonding, shaping their consumption, spending behaviors as well as manipulating their socio-political opinions and worldview. Built-in features – such as social networking and recommendation – allow consumers to access music they may never have heard about before. As a meeting place of free choice and algorithmic personalization, music streaming platforms offer both automated listening and opportunities for participation and individual music management. But does it allow audiences to consume its media content any way they like? What limitations such music streaming platforms as Spotify, put upon its users?

Today media are often deeply integrated into the users’ everyday routines. Internet experiences described with tool metaphors examine technology as extensions of our senses or bodies, allowing us to magnify or amplify certain capacities (Hagen. 2016). As a part of access, music achieves an increased position in everyday life and therefore music-streaming services have become mainstream in everyday music listening. Music immersion by individual streaming practices paved the way for intuitive and effortless music experiences to arise in whatever everyday contexts users find themselves. The services offered a low threshold for individual music management, with minimal effort or attention on the users’ part. The informants hence emphasized service features that provided immediacy, flow, and direction, optimizing listening on the move, in brief in-between moments, in the background, and alongside other daily activities (Hagen. 2016).

Spotify, as all the other streaming and media platforms, works in a principle of algorithm, thus automatically narrow as depending on keyword search, presenting limited choice results, with examples of most streams, not the most similar titles at the top. The company’s algorithm chooses which updates appear higher up in users’ newsfeeds and which are buried. To work to be noticed by the algorithm (as to use certain words that algorithm likes, tagging updates with unrelated brand names, gather a certain amount of streams or pay for the company to advertise you on a front results page) has all become a more and more important part of the music-making and distributing. Also, algorithms not only stream, but produce and create music played on the streaming platforms. In such platforms as Spotify algorithm shapes not only the creative process and business structure, but also who chooses music that the algorithm suggests, either by recommendations, playlists, shuffle playing or suggested search results. You can search only from the catalog they have and choose only from the results they present. Usage of music streaming platform doesn’t leave much for free choice, it just allows to freely choose the direction of the music streaming. Content geared toward these algorithmically fueled bubbles is financially rewarding. In this way, algorithm has a direct impact on our mood, actions, worldview, values and identity, updating it all to fit the more modulatory, data-driven reality of personalized media (Prey. 2018: 1096).

Industrialization and early broadcast media played in the transmutation of ordinary people into the masses. Today, personalized media appear to dissolve the mass, bringing into view the long-concealed individual media consumer (Prey. 2018: 1087). Spotify, whose etymology is of a combination of “spot” and “identify”, is selling music as a personalized experience (Fleischer, Snickars. 2017: 159). Music taste provides a particularly fertile means through which to ‘see’ the individual. One’s taste in music has thus long been seen as a window into one’s sense of self, and place in society (Frith. 1998: 90). In everyday mode, music listening is a context- and identity- sensitive way of being. Individual sense-making of music streamed primarily revolves around the self, and how the self-interacts with and realizes the world in the present – expresses and negotiates self-identities, ongoing life projects and relationships. Therefore it comes as no surprise that music streaming practices are often seen as part of identity work.

With distinctions between technology, everyday life, self, and others beginning to break down, music streaming platforms directly engage with the users‘ self – how they look at and experience themselves and how they interact with and make sense of the world (Hagen. 2016). By translating their daily experiences into data on music streaming platforms, people create data value for their experiences and allow algorithms to store and interpret it in regards to users’ identity. Since these processes are managed by algorithm, it allows the system to get to know the user more than the user himself. In attempting to ‘know’ the individual media consumer, recommendation services are committing the cardinal sin of reification: reifying both the subject and the object of media consumption (Prey, 2018: 1095). Those new distribution channels generate and store valuable data about users and their listening behavior. It allows them to provide a personalized listening experience that is individually and genetically fitted to each user, according to their taste profile . From the songs they listen to, to the products they buy, users produce our identity and modulate themselves as individuals.

The bigger questions about user‘s identity rise up since the human itself is known for science to be a set of many different independent algorithms formed by genes and environmental factors. People are known to take decisions deterministically or accidentally, but not according to their free will. In this case, when people are not immutable subjects to be modeled, algorithm cannot present any ‘real’ individuality. Therefore by streaming algorithms the individual music listener is understood as having many music identities, rather than one stable identity. John Cheney-Lippold describes it as a ‘cybernetic relationship to identification’ whereby essentialist notions of identity are replaced by ‘pliable behavioral models’ (Cheney-Lippold, 2011: 168). Advertising’s “targets” are not individuated human beings but inferred ones: sets of demographic, psychographic, and other data points (Mähler, Vonderau. 2017: 213). Therefore it is based on processual identity: of the perpetually ‘becoming individual’. The algorithmic individuation process remains hidden from view from the user (Prey. 2018: 1096). In this way, modern identity becomes a small data piece in the algorithmic system that is not fully understood by anyone.

The identity, formed by the music streaming platform algorithm , allows making personal recommendations and playlists to the user. As more and more music is available to be consumed, it gets difficult for the user to navigate through it and recommender systems help the user to identify music he or she wants to listen to without browsing through the whole collection (Pichil, Zangerle, Specht. 2015: 1). Music streaming services increasingly incorporate different ways for users to browse for music: taxonomy choice (i.e., mood, activity, or genre), individual differences (e.g., personality traits and music expertise factors), and different user experience factors (i.e., choice difficulty and satisfaction, perceived system usefulness and quality) (Ferwerda, Yang, Schedl. 2019: 20157). New search creates new struggles to formulate what the user is searching for and use correct keywords (Hagen. 2016). There is an argument that people are able to process only about seven items at a glance, for a short span of attention. More options can cause confusion, difficulty to make a choice and decreased choice satisfaction (Ferwerda, Yang, Schedl. 2019: 20161). Also, only recommendations for items the user already interacted with can be evaluated. Therefore the recommender system is very limited and not fit for recommending new artists . (Pichil, Zangerle, Specht. 2015: 4). Even though Spotify Radio claims to be offering a personalized and infinite avenue of discovery, yet the music seems to be delivered in limited loop patterns, playing the same artists over and over (Fleischer, Snickars. 2017: 184). Since the year 2015, the table of the most streaming artists remains the same . Irrelevant search taxonomies on music streaming platforms distract the user, complicate the search process, and increase the search effort because of conflicting attention. Also, streaming platforms change the way users are searching for information about music. Cultivation of traditional music reviews, newsletters, and magazines as sources of information had been marginalized amid the rapid and abundant flow of online music and replaced by service features that supplied him with listening suggestions, such as “related artists” and news flashes (Hagen. 2016).

Since algorithmic identification (a socio-technical process engaged in enacting the individual) is ever-changing, recommendation system relies on the context and environment of the user. It is known that ‘people have more in common with other people in the same situation, or with the same goals, than they do with past versions of themselves’ (Pagano et al., 2016: 1). Context-based recommendation systems personalize to users’ context states, placing the individual into the system of reality in which the individuation occurs. Today mobile devices permit the data points like location, motion, time of day, and nearby contacts (Prey. 2018: 1092), making streaming platforms able to collect and aggregate data to recommend music that matches a listener’s current context. Therefore most of Spotify‘s playlists are assembled together for specific occasions and edited according to musical intuition and everyday routines. Playlist listening experiences suggest a multitude of moods, moments, and genres (Fleischer, Snickars. 2017: 139), with titles serving as hooks for memories or simulated moods. Music becomes a background element, blending into the users’ larger environments without calling attention or an element of the activity. Therefore the shuffle functionality, providing tracks in a random order, is the most often accompanying everyday activities. Music can be streamed more frequently, effortlessly, and unconsciously. Though the music was nominally organized, it affected the listener‘s surroundings and impacted inner states (Hagen. 2016).

Determined users’ context is more useful for one’s ‘musical identity’ not just for predicting what song to play next but also for determining what ad to serve a listener . Spotify promised to ‘identify – in real-time – what a listener is doing, and give brands an opportunity to own that moment’ . In this way, brands can directly engage with Spotify listeners ‘during life activities, major life moments, real-time mood states, and seasonal events’ (Prey. 2018: 1094). According to them, Spotify audiences are sorted into categories, constructed by data (Prey. 2018: 1088). ‘Data subject’ categories are modulated according to contexts that will be attractive to advertisers and mainly defined by brands. Since the system of categorization is deemed the most economically meaningful (Prey. 2018: 1096).

Music has always been both a personal and communal experience (Van Dijck, 2007). Music is intimately connected to the social groups that people identify with and has the power to intimately bond individuals. Therefore the sociocultural value is assessed in terms of both identity and social presence, relationships (Hagen. 2016). The social aspect of music comes through the acts of sharing music and following others. Therefore it came as no surprise when in 2010–11 Spotify established itself as a social media platform, making a strategic partnership with Facebook . Both companies use data to build robust users’ profiles and place recommendations based on their friends’ behavior. Often Facebook algorithm can guess its users’ personalities and inclinations more accurately than their contacts, even the close ones. Spotify users follow what their friends are listening to, thus highlight both its social networking and profile display features (Sinclair, Green. 2015: 15). In this way Spotify was created as a socio-cultural place that accommodated interactions and activities, using music as social objects (Hagen, Lüders. 2017: 1).

The social aspect of music streaming also creates new problems that hinder the free choice of an audio media user. Since ‘we derive our sense of self from the image of our self that others reflect back to us in interaction‘ (Prey. 2018: 1096), listeners are socially aware that posting music demands an internal reckoning regarding what is appropriate to share and what ought to be kept private (Jones. 2011: 213), choosing between non-sharing, selective-sharing and all-sharing approaches. The chief executive of Facebook stated that people “have complete control” over everything they share on Facebook. But the power to create and present an autonomous identity with regards to your music taste and preferences was stripped away from the user due to the Spotify – Facebook link, that has a Terms of Service condition that ‘This app may post on your behalf, including songs you listened to, radio stations you listened to and more‘ (Facebook Apps Centre, 2014). Thus, Spotify automatically posts user’s songs ‘and more’ to their Facebook network without their knowledge or consent. An outcry from users forced Spotify to introduce new options for protecting the privacy of musical preferences and current settings allow users to listen to music in either a ‘private session’ or a default option – ‘public session’, in which a user’s song choices are publicized (Kant. 2015: 41). The Spotify app automatically switches from a ‘private’ to a ‘public’ listening session after twenty minutes and if the user leaves his streaming gadget on shuffle, his social circle is informed on random unheard songs, connected with the name and identity of the user (Kant. 2015: 41). In this way, rather than functioning simply as a tool for self-expression, music streaming platform was revealed as a powerful, algorithmic ‘socio-technical actor’ (Gillespie, 2014: 179), having the power to actively (re)write users’ Facebook history to suit the operational imperatives of Spotify, as well as (re)shape the users’ intentional representations of identity and perform sometimes even unwanted utterance of selfhood. A person, wanting to present a socially acceptable form of selfhood, is forced to control self-performance and orient his listening actions to suit the algorithmic protocols of the two connected apps (Kant. 2015: 44). It is called a Model’ of the socio-technical organization when ‘the grammar of action’ causes the individual ‘to orient their activities towards the capture machinery and its institutional consequences’ (Agre. 1994: 110) as if when Spotify forces its users to regulate music choice to adhere to a normative ideal of publicly acceptable music. In the struggle between the autonomy of the user and the app, the latter holds the potential to tell users ‘who they are’, not vice versa.

Another threat comes from the privacy-invasive collection of users’ personal data with little transparency or outside oversight . A novel algorithm for automatic playlist generation in Spotify utilizes Facebook information on liked bands by a user and his/her listening history (Germain, Chakareski. 2013: 26). Music recommendations are based on the users’ friends‘ music tastes and listening histories. Also, Spotify offers the option to share music through Facebook Messenger . Using private information, both companies micro-target users based on their emotional states (Mähler, Vonderau. 2017: 213), making the human experience a proprietary data and emotion a commodity , even though feelings are unsaleable because they are inalienable. In this way, digital data is mixed into a person’s view of the real world thus creating augmented reality. When technology calculates, models and outwits people’s feelings , users never have a free choice of will .

During most of the modern age, citizens of democratic societies have regarded a person’s experience as inseparable from the individual. People decide if and how to share their experience, with whom and for what purpose. Both collaborators Spotify and Facebook, named “ruthless surveillance capitalists”, are mainly making money by providing content to audiences, collecting gobs of their personal data and monetizing the audiences by selling them to advertisers (Fleischer, Snickars. 2017: 134). In this way, on streaming platforms, it is not just the music that is being streamed, but the listener as well and users have limited means to opt-out of their data being used (Angwin. 2016).
On both platforms, all data is collected as equivalent, though not all of it is equal. By proving that all people’s experiences are equally valuable, human progress will be converted to the values of social equality and humanity will be pushed into degradation (Harari. 2016: 236). Social media is not a public square but a private one governed by machine operations and their economic imperatives, incapable of distinguishing truth from lies or renewal from destruction. Chaos was caused by coordinated schemes of corrupt information, produced by the same tools as used in music streaming platforms: profit-driven algorithmic amplification and micro-targeting. That makes media and consumed public information out of people’s control .

Music is one of the very few things that billions of people around the world enjoy every day. Access to it digitally has redefined the processes of production and commercialization of cultural communication industries (Pedrero-Esteban, Barrios-Rubio, Medina-Ávila. 2019). Algorithmic data-based music streaming creates a new structure, meaning and form for everyday music consumption. Immersed in multimedia culture, the audience bases music listening on display devices (Díaz-Nosty, 2017.) and take part in exchanges within media channels as their personal spaces, transferring intimacy to increasingly mainstream interaction settings (Pedrero-Esteban, Barrios-Rubio, Medina-Ávila. 2019). This makes music-streaming services as lifeworld resources, able to confirm, challenge, mold, establish and endorse listeners’ notions of sociality, corporality, environments, time, and self-identity. Algorithmic individualization according to customer data, intertwines with consumer categories and commercial imperatives, demanded by advertisers (Prey. 2018: 1086). Personal data has become the most prized commodity, traded on a vast scale by some of the most powerful companies (Dance, LaForgia and Confessore, 2018). A modern person doesn’t even have the primal choice whether to consume media (and be used by media companies) in the first place. Therefore new charters of rights, legal frameworks and institutional forms are necessary to ensure a digital future that is compatible with the aspirations of a “democratic information civilization” (Zuboff. 2021), where the conscious choice of cultural experiences within the usage of media would be possible.

List of Literature:
• Agre, P. (1994): “Surveillance and Capture: Two Models of Privacy.” Inf. Soc. 10. 101-127.
• Angwin, Julia et. al. (2016): ‘Facebook Doesn’t Tell Users Everything It Really Knows About Them,’ ProPublica, December 27. Accessed August 19, 2017. www.
• Cheney-Lippold, J (2011): A new algorithmic identity: soft biopolitics and the modulation of control. Theory, Culture & Society 28(6): 164–181.
• Dance, Gabriel J.X. LaForgia, Michael. Confessore, Nicholas (2018): „As Facebook Raised a Privacy Wall, It Carved an Opening for Tech Giants” By Gabriel J.X. Dance, Michael LaForgia and Nicholas Confessore. New York Times. Dec. 18, 2018.
• Díaz-Nosty, B. (2017): Coexistencia generacional de diferentes pra?cticas de comunicacio?n. In B. Díaz-Nosty, (Coord.), Diez an?os que cambiaron los medios: 2007-2017 (pp. 7-26). Barcelona: Ariel.
• Ferwerda, B., Yang, E., Schedl, M. et al. (2019): Personality and taxonomy preferences, and the influence of category choice on the user experience for music streaming services. Multimed Tools Appl 78, 20157–20190.
• Fleischer, Rasmus; Pelle, Snickars (2017): “Discovering Spotify – A Thematic Introduction” Culture Unbound: Journal of Current Cultural Research. Volume 9, Issue 2. Linköping University Electronic Press.
• Frith, S (1998): Performing Rites: On the Value of Popular Music. Cambridge, MA: Harvard University Press
• Germain and J. Chakareski (2013): “Spotify Me: Facebook-assisted automatic playlist generation,” 2013 IEEE 15th International Workshop on Multimedia Signal Processing (MMSP), Pula, Italy, pp. 025-028.
• Gillespie, Tarleton (2014): ‘The Relevance of Algorithms’, in Tarleton Gillespie, Pablo Boczkowski, and Kirsten Foot (eds.) Media Technologies: Essays on Communication, Materiality, and Society (Cambridge: MIT Press), 167–194.
• Hagen, Anja. (2016): The metaphors we stream by: Making sense of music streaming. First Monday. 21. 10.5210/ fm.v0i0.6005.
• Hagen AN, Lüders M. (2017): Social streaming? Navigating music as personal and social. Convergence. 2017; 23(6): 643-659.
• Harari, Yuval Noah (2016): Homo Deus: A Brief History of Tomorrow. 528 psl. Random House.
• Heath, A (2015): Spotify has a secret ‘taste profile’ on everyone, and they showed me mine. Business Insider, 14 September.
• Jones, S (2002): Music that moves: Popular music, distribution and network technologies. Cultural Studies 16(2): 213–232.
• Kant, Tanya. (2015). FCJ-180 ‘Spotify Has Added an Event to Your Past’: (Re)writing the Self through Facebook’s Autoposting Apps. The Fibreculture Journal. 30-62.
• Mähler, Roger & Vonderau, Patrick. (2017). Studying Ad Targeting with Digital Methods: The Case of Spotify. Culture Unbound: Journal of Current Cultural Research. Volume 9. Linköping University Electronic Press.
• Pagano, R, Cremonesi, P, Larson, M, et al. (2016): The contextual turn: from context-aware to context-driven recommender systems. In: Proceedings of the 10th ACM conference on recommender systems, Boston, MA, 15–19 September, pp. 249–252. New York: ACM.
• Pedrero-Esteban L., Barrios-Rubio A., Medina-Ávila V. (2019), Adolescentes, smartphones y consumo de audio digital en la era de Spotify, Comunicar; Comunicar 60: Cómo llegar a ser un genio. Aprendizaje personalizado y altas capacidades en la sociedad conectada; 103-112.
• Pichl, Martin & Zangerle, Eva & Specht, Guenther. (2015): Combining Spotify and Twitter Data for Generating a Recent and Public Dataset for Music Recommendation. 1313.
• Prey R. (2018): Nothing personal: algorithmic individuation on music streaming platforms. Media, Culture & Society. 2018;40(7):1086-1100.
• Rogers, Richard (2013): Digital Methods, Cambridge, Mass.: MIT Press.
• Schneider, Lior (2017): “Big Data As a Marketing Strategy – SPOTIFY’s 2016 Campaign”. Overdrive Interactive Digital Marketing Blog. March 7, 2017.
• Sinclair, Gary & Green, Todd. (2015): Download or stream? Steal or buy? Developing a typology of today’s music consumer. Journal of Consumer Behaviour. 15. 10.1002/cb.1526.
• Singer, Natasha (2014): Listen to Pandora, and It Listens Back. New York Times. Jan. 4, 2014.
• Van Dijck, J (2007): Mediated Memories in the Digital Age. Stanford: Stanford University Press.
• Zuboff, Shoshana (2021): The Coup We Are Not Talking About. New York Times. Jan. 29, 2021.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s