Quantifying Taste: Spotify's algorithmic effect on musicking

(1)

MSc New Media & Digital Culture

Media Studies

Master Thesis

QUANTIFYING TASTE

SPOTIFY’S ALGORTIHMIC EFFECT ON MUSICKING

by

Philip Creamer

11636254

June 2018 18 ECTS January – June 2018

Supervisor/Examiner: Assessor:

Dr. Bernhard Rieder Dr. Esther Weltevrede

(2)

Spotify’s increasing popularity as a music streaming service posits itself as an intermediator between the consumers and the musical realm, which encompasses every instance of musical engagement also know as musicking (Small, 2008). Spotify’s methods of content personalization are similar to platforms who retain and re-incorporate user information for product development, using methods such as habitual user feedback and social media integration. Data tracking methods implemented by Spotify are facilitated by the use of algorithms which affect traditional forms of musicking, such as musical discovery methods. This causes a paradigm shift in the ways music associates among itself and its encompassing ecosystem, much like its technological adoption of analogous to digitized forms of media, and through reintegrating methods such as user feedback, taste preferences and genre similarities, to name a few. Algorithmic elements pertain to numerical aggregation, which is utilized to compare and reinforce artist status popularity, followings and ranking, among the Spotify platform which therefore materializes into the public sphere. In addition to Spotify’s increasing use and popularity are the growing concerns of platform capitalism; Spotify’s use of personal data to be reintegrated back into its infrastructure for the return of increased use, attention, and consumption while acting as the leading example for corporatized digital music streaming models. It is argued that Spotify’s algorithmic influence plays an important role in forming and reinforcing notions of popularity, ranking and hierarchy, also factors that coordinate the corporatized music industry. Through the use of new media digital tools, it is, therefore, possible to observe the Spotify’s algorithmically formulated qualitative and quantified results, which are made viable through Spotify’s technical affordances.

(4)

1.1 Acknowledgements

I would like to thank my thesis adviser Dr. Bernhard Rieder, Associate Professor of the New Media and Digital Culture at the University of Amsterdam and a collaborator with the Digital Methods Initiative. The door to Prof. Rieder’s office was always open whenever I ran into a trouble spot or had a question about my research or writing. He consistently allowed this paper to be my own work, but steered me in the right direction whenever he thought I needed it.

Finally, I must express my very profound gratitude to my family, my friends and to my loving partner, Holly, for providing me with unfailing support and continuous encouragement throughout my year of studies and through the process of research and writing. This accomplishment would not have been possible without their love and undying support.

Philip Creamer

Amsterdam, 2018

(5)

2.0 Introduction

As mobile applications expand into our activities and habits, they have the ability to assist or enhance a user’s occupation by adding numerical, calculated and algorithmic resources, as Tarleton Gillespie comments on algorithms as being “a key logic governing the flows of information on which we depend” (Gillespie, 2014:1). These applications may directly effect traditional practices and customs, such as the technological and biopolitical leap from analogue technology, in favour of, digital technologies. The biopolitical, defined by Michel Foucault as the “management of life” (Foucault, 2003 as cited by Lazzaratto 2006: 9), is in reference to the increasing disappearance of consumable physical media, towards an increasingly digitized and stream-consumable society. In order to observe an aspect of society that is in favour of the digital transition, Spotify will serve as a case study in exemplifying how such a transition ultimately affects traditional behaviours such as the way a society interacts and engages with music.

Spotify is an online music streaming application which provides its members access to a large music library, currently hovering around 35 million (Spotify, 2018). Depending on the type of subscription model a user is subscribed to, different forms of access to its services are afforded, such as increased mobility through data usage or a decreased experience of application fluidity such as mandatory advertisements. Launched in October 2008, Spotify has seen an increase in popularity for the most part of a decade, learning from already established music streaming models such as Pandora and LastFM, as well as benefiting from their transparent business model that has instilled trust among its users as well as members of the musical industry, although some artist members may disagree, such as Braid’s guitarist Bob Nanna who worries that Spotify “is more interested in building a strong, lasting business than supporting artist’s careers and the industry” (Swanson, 2013: 212). Available on desktop and handheld devices, Spotify’s content is only accessible and playable through its proprietary application. Spotify has stated to have over 71 million paid subscribers as of December 31, 2017 (Spotify Technology S.A., 2018: 8) and is currently available in 61 countries, exemplifying its permeation of use and has increased its popularity among rival music streaming services such as Pandora, an alternative music streaming service associated with radio recommendations and the leading music service according to the Edison Research listening statistics (2017).

(6)

The permeation and the rising popularity of digital platforms are to be understood as drivers and co-ordinators of the social, essentially “reworking, mediating, mobilizing, materializing and intensifying social and other relations” (Ruppert, Law & Savage, 2013:3). Spotify’s position as an mediator requires an investigation of the activities occurring and surrounding its responsibilities and intermediary role as a gatekeeper of music among societies. One of Spotify’s platform characteristics is one of musical taste management. As a user engages with the platform, their tastes dictate the type of content they will interact with, whether the engagement is predetermined or an expected automated service of Spotify’s catalogue, either function is determined and led by taste and preference. According to Bourdieu, taste is also a result of class and status, an aspect that has historically organized and managed the way music is heard, engaged and consumed.

The intent of this paper will seek to observe to what extent does Spotify redefine or reconfigure traditions of musicking, such as personal tastes and preferences, enabled by musically digitized methods, such as quantification and algorithms, which will be established through a theoretical framework. Two methodologies will then be applied in order to observe how do these digital reconfigurations alter the musical ecosystem, observations such as the imposition of hierarchies, methods of artist association and a close-reading of Spotify’s implemented algorithms.

From the artist’s perspective, forms of engagement may lead to a hierarchical reinforcement of status and visibility within Spotify’s platform in addition to having a societal and cultural impact. This hierarchical structuring is of interest since Spotify implicates itself within the sphere of culturally musical expression, and assuming the role of an increasingly popular platform, it contributes to the quantifiable measurement of taste that is re-used to develop and further marketing strategies based on user preferences. This points to Spotify’s characteristic that is responsible for reshaping the way an individual or a society understand music, not simply how its made or produced, but also how we relate to it, the methods of musical discovery and the forms in which it is consumed. These qualities are described as a result of algorithmic elements that are implemented throughout Spotify’s infrastructure, algorithms that are responsible for analyzing both the music and the individual, in order to find the nearest matching qualities. As algorithms are employed towards its user base it creates a standardized method of the way music relates to

(7)

each other as well as an expectation of the way individuals supposedly engage with music. Through accumulated data provided by its users, Spotify is an intermediary responsible for the management of life “in the sense that it seeks to reproduce the conditions of existence of a population” (Lazzaratto 2006:9). This heterogeneous normalization of the way Spotify’s infrastructure is laid out, it is governed by Spotify’s social algorithms, essentially altering certain acts of musicking such as forms of engagement, forms of discovery and forms of identity.

As Spotify gains popularity through means of active users and paying members among the platform, it has been argued that such an accumulation of information, data from consumers and artists, positions Spotify as monopolizing the field of taste management among the digital music streaming landscape, which is also known as platform capital (Srnicek, 2017). Spotify uses the business model of music streaming, acquiring rights to large music catalogues and forming strong relationships with record labels, in combination with a transparent royalty distribution which gains the trust of their on-boarded artists, Spotify has aimed to combat piracy by legitimizing the trend of being in possession of a large library of music with relatively low payment installments. While other platforms offer similar fees as well as technical affordance with similar layouts, none have proven to be as successful as Spotify, which means that the platform bears much responsibility on the effect it has in the environment it engages directly and indirectly with. In this instance, musicking is also a factor that is altered; to question the means of consumption and production through digitized forms are required considerations to uphold as Spotify dominates and questionably monopolizes the field of digital music streaming.

API’s (Application programming interface) are programming interfaces that allow the reception of data from online services, which are publicly provided by Spotify rendering it possible to observe certain facets of its algorithms to a certain extent that limits areas of observation, rendering it not entirely possible to observe the entire process in the ways it conducts itself. In order to gain further insight into the ways Spotify’s musical management is conducted, a combination of available digital tools enables a reconstruction or an alternative method of visualizing the on-goings occurring within the platform. The Spotify Artist Network tool enables a visualization of the connection between related artists as well as provides a workable data sheet that is importable to other digital tools. Gephi allows further work and editing within the exported data sheet which

(8)

may be further visualized with the use of RAWGraphs, another digital tool that allows multiple forms of visualizing the input data. Together in combination, digital tools provide alternative forms and ways of describing the occurrences within a platform that is otherwise closed off from an internal investigation.

The results have shown an in-depth charting of artist relations through a combination of descriptions provided by both artists and users, essentially providing unintended associations and unlikely relevance as well as adding surplus information for the artist, while one artist may not associate themselves with one specific genre, the Spotify community as a whole may assign unforeseen descriptions allowing a deeper opportunity to relate itself to other artists. While it is not entirely possible to render conclusive results about the observation made, certain aspects of the algorithms implemented within Spotify are observable and mapped by visualizing the functionalities of the algorithm.

(9)

3.0 Musicking and Musical Taste

As digital technologies aim to provide accessible information to individuals, the information provided, in this case, musical, is likely to accompany the individual during certain activities. These activities may be the accompaniment of music through physical activities in order to motivate and stimulate, commutes to pass the time to assist and in a way escort the listener to their next destination. Actions, habits and routines are altered, reshaped and reflected by the accessibility and commodification of digital information that is made available (Morris & Powers, 2015, as cited by Fleischer 2017: 12), in the previously mentioned examples, music has a characteristic of being a motivational entity that adds substance to an already experienced existence, affected by musical unpredictability and its aim to promote reaction from the listeners. In the case of musical interactions such as performing, listening, musical preference, purchasing, licensing and any other method where the individual and music interact with music, traditional methods of interaction with music still exists but is altered and furthermore affected by the mediation of digital technology. Throughout this essay, any interaction or engagement with music will be termed musicking, which is defined by the interplay between individual and music, a term coined by Christopher Small (1998). In the case of musical taste, to have a preference and a disposition towards music is, in fact, a form of musicking, an action and reaction that engages and further promoting towards the culture of music.

(10)

Our interactions with music evolve and change throughout different cultures, in certain cultures music, rhythm and dance is synonymous to the way communication is made within society as Helena Wulff describes dance, as a feature of “socialization; the functional aspects of dance is that of transformation: the classical examples are rites of passage where people are moved from one stage in the life cycle to the next one, such as initiation rites, wedding celebrations, and funerals (Wulff, 2015: 1), in contrast to other cultures where traditional folk music has been introduced by the working class, as Philip Bohlman explores folk music stemming from oral traditions “comprises both musical and ethnographic concerns… also, a measure of a community’s sense of itself, its boundaries, and the shared values drawing it together ” (Bohlman, 1988: 14), where traditions are increasingly disappearing in favour of individualistic and technologically afforded forms of musicking, such as engaging with on-demand music streaming services. All these forms of engagement with any component that is in any form musical is considered to be an act of musicking (Small, 1998), which marks how music should be thought of, not as an object but as a practice, as an asset and a phenomenon that may act upon humans and vice versa (Leijonhufvud, 2018 :25). Small goes further in depth of the definition,

“Musicking is part of that iconic, gestural process of giving and receiving information about relationships which unites the living world, and it is in fact a ritual by means of which the participants not only learn about, but directly experience, their concepts of how they relate, and how they ought to relate, to other human beings and to the rest of the world” (Small, 1998:2)

An individual does not need to be a musical expert in order to perform the act of musicking; from the smallest gestures of turning on a car radio, hearing muffled forms of music through a follow commuter’s headphones, to a small monetary donation towards a musical street performer, Small further describes this deliberate and involuntary form action,

“we should certainly include dancing, should anyone be dancing-in many cultures if no-one is dancing then no music is happening-and we might even on

(11)

occasion stretch the meaning to include what the usher is doing who takes the tickets at the door, and the heft men who shift the piano around, and the cleaners who clean up afterwards, for what they do also affects the nature of the even which is a musical performance” (Smalls, 1998:5)

Musicking is every way, shape or form, in which an individual interacts with music and these methods are necessary to take into consideration as digital technologies continue to permeate societies, especially since analogous technologies offer entirely different forms of musicking from its digital counterpart, such as the disappearance of the physicality of musical mediums such as CD’s, cassettes, vinyls. The traditional methods of navigating music are in some sense performative, from the purchase of a physical album to the enabling of listening to it through various technological means to the gestural meaning of music ownership, which displays an individual’s performative relationship with music. One could also factor in mixtapes,

compilations and radio listening, any method or act put towards, where music is the central focus of attention and further promotes its cultural and historical existence. Small describes musicking to any call to social action, “which is to say performance, that is central to the experience of music. You do not actually need what Dalhaus calls a created form, which is to say a music work… Many of the world’s musical cultures get along very well without any such thing” (Small, 1998:4), meaning that musical mastery, linguistic understanding, is not required in order to promote or contribute to the act of musicking.

To further describe musicking would be to investigate the ways in which individuals and societies navigate and manage music, “to express the much broader idea of taking part in a musical performance… as a genuine tool for understanding the nature of the music act and its function in human life” (Small, 1998:5). The management and relationship we have with music and vice versa can be defined as the ecology of the biopolitical, which is further defined by Foucault as the “management of life” (Lazzaratto 2006:9) and in the case of this essay in relation to Spotify, it is referred to the increasing disappearance of consumable physical media, in favor of an increasingly digitized and stream-consuming society. This accumulation of interaction and events is further described by Smalls as “all these activities add up to a single event, whose nature is affected by the way in which each of them is carried out, and having a tool for exploring the nature and the

(12)

meanings of the even as a whole” (Smalls, 1998:5). By understanding the relationships created and formed by musicking, would enable a further understanding of how musical tastes are formed, how technologies change the way we, in various societies, perceive music and technology and how both subjects co-exist and influence each other. The act of musicking creates a foundation as to why and how personal preferences are created and how they can be unpredictable and uniquely personal.

(13)

3.2 Musical Taste

Musical taste is an attribute of human behaviour that has been observed to be unique to each individual, as Annukka Lindell & Julia Mueller describe the old adage “there’s no accounting for taste” (Lindell & Mueller, 2011: 1) when referring to aesthetic appreciation, in this case towards art. From a worldwide to smaller communities and more narrowly specific cultures, small groups may acquire a similar disposition to a particular artist, only to be discovered by show-going or expressed in public forums. But these assemblages or groups may not be reflected in the society as a whole or even close to encompassing the worldwide population.

“Like language, art is a defining human attribute. Yet where a long-standing and ever-expanding body of research has examining the way the brain processes and produces language, the neuronal underpinnings of art, from production to appreciation, have been comparatively ignored… [which] likely comes from the inherent belief that art, in terms of both production and appreciation, is not a rule-governed enterprise” (Lindell & Mueller, 2011: 2)

The importance of taste is that it renders works of art more valuable, contributing to its acceptance among society and attributes an acknowledgement of appreciation towards its existence. “the aesthetic response will be defined as the receptive process in which a viewer focuses attention on the aesthetic qualities of the artwork (e.g., beauty, balance) and reflects upon the evaluation of that artwork” (Lindell & Mueller, 2011:3). The human quality of taste may be a valuable asset to the organization of information, much like Spotify profiles their customer-base in order to provide material that is closely relevant.

New media environments engage users with the medium or content in forms that have not been possible a decade ago such as the interactive level between technology and user that associates the internal and behavioral human mechanisms, such as the “patterns of the length in which an individual spends listening to music” (Zhang et al. 2013:5), “the behavior of switching from desktop to mobile modes of accessing” (Zhang et al. 2013:5) regarding Spotify, as well as the

(14)

correlations between “the session length and downtime of successive user sessions on single devices” (Zhang et al. 2013: 5).

One behavioural facet is the human function of taste and distinction (Bourdieu, 1984), which is the ability to have a preference or to possess aesthetic disposition towards, for example, musical genre, artistic style or vocal tonality. According to Bourdieu, taste is inseparable from cultural competencies (1984:4) which are established through cultural practices (1984:3) and is the product of “upbringing and education” (1984:3). This, in turn, creates a hierarchy among the arts and “within each of them, genres, schools, and periods all correspond to a social hierarchy of consumers” (1984:3), the musical realm in this respect has numerous streams of categories, particularities and variations. All of which contributes to a niche or a popularized trend, fuelled by willing listeners.

Taste finds itself through many streams of human behaviour and the ability to have aesthetic judgment is a quality that differs among cultures and is a trait that is difficult to quantify, in this case, finding the median agreement of tastes across the entire make-up of cultural class is a difficult task. Since taste reflects greatly on human condition digital technologies have made this quality possible to categorize and commodify. An individual’s preference can be sold back through targeted qualities that have been analyzed and tested through algorithmic properties, which are put into place to focus on the most effective ways of promoting the inherent and individualistic quality of taste. Taziana Terranova describes this process as “affective labour” (2004) where types of cultural and technical labour, “have developed in relation to the expansion of the cultural industries and are a part of a process of economic experimentation with the creation of monetary value out of knowledge/culture/affect” (Terranova, 2004: 86), which can be seen as a Spotify user’s act of listening as a mode of production in order to produce quantity.

In music industry terms, tastes are most apparent and visible throughout record stores, which are less common among today’s society. According to the market research firm Almighty Institute of Music Retail, “one-quarter of the country’s 3,600 (U.S.) independent record stores had closed since 2003” (Commercial Observer, 2017) and according to ABC News “between 2000 and 2010, record store sales declined more than 76 percent (Commercial Observer, 2017). CD’s, cassettes and

(15)

vinyl’s correspond to an era’s technological situation which is currently evolving from analogue towards digitized mediums. The disappearance of the record store is due to the evolution towards a different form of materiality in favor of digitized forms of music. Musical genres were at full display, presented among isles and ranging from classical to rock to world, also showcasing music’s historical evolution in one singular place. While the era’s change and evolve, some genres stand the test of time, some niche’s fade away, or some develop and find some importance in another point and time.

Bourdieu points out that taste is subject to class, which entails the social and economic in reference to moviegoers and movies as an ‘art moyen’ or “middlebrow art” and according to Julien Duval, Bourdieu “demonstrates that the practice of cinema-going takes different forms which attract different classes and fractions of classes unequally there is a structural resonance or homology between the space of the cinema and the social space from which the audiences comes” (Duval, 2014:1)”. Garage punk may not garner much attention in a broader sense but for those it does interest, may be due to social up-bringing, a disassociation or opposition to normativity. Classical music may be regarded as an upper-class genre and associations with formal practices such as gatherings feature formal attire, or associations to a studious individual, where one would focus on the musically theoretical aspect of this genre or more broadly, developing a vocabulary for all that music encompasses.

(16)

3.3 Taste according to class

Digital technologies have evolved with a focus on reduction and efficiency and in turn, consumers have also come to expect the almost instantaneous delivery of cultural goods. This expectancy is a change in the traditional ways music is acquired and consumed, as previously mentioned, in regards to music, taste varies but correlate to class; Spotify challenges this by amalgamating a vast library of music and assorting and associating its content with the use of algorithmic methods. In addition to taste being affected by technological means, various classes and groups are exposed to music in different and unpredictable ways, re-shaped and re-associated within Spotify’s technological infrastructure.

As Bourdieu points to class as a symptom of differentiating tastes (1984: 2), Spotify, while maintaining class structures, is shaping and re-defining the ways that the world of music is informed, affected or governed by its techniques of quantifying music and its listeners (Passoth et al. 2014: 271). While Spotify’s services offer a vast library of music to its members, only a certain class may access its services through digital means, leaving those without access to technological means shut out. Furthermore, one’s country may not have access to Spotify’s services fueling Bourdieu’s link of taste and social origin.

The space that class occupies is distinguished in three parts, according to Bourdieu, the first, being the dominant class constituting a relatively autonomous space whose structure is defined by the distribution of economic and cultural capital among its members, “certain configurations of distributions corresponds a certain life-style, through the mediation of the habitus” (Bourdieu, 1984: 260), which in the second part, the capital among the fractions is “symmetrically and inversely structured”(1984: 260), and lastly, “the different inherited asset structures, together with social trajectory, command the habitus and the systematic choices it produces in all areas of practice-these structures should be found in the space of life-styles” (1984: 260).

Musical genres such as metal, hip-hop and classical, can all be indicators of social class and to mix genres may also indicate an individual’s educational level, but the outreach of new media

(17)

technologies and Spotify span across social classes and accrue information left by individuals to the benefit of the developers.

While album sales were an indication of an artist’s success and popularity, the specificity of the numbers (age, sex, location, etc.), are measurements that have often been overlooked. With digital technologies such as Spotify and the digital imprints left behind by its users, quantification techniques have been used to inform the company about the performance results of its product, effectively redefining the way music is listened to and also affecting the listeners engaging with the platform (Passoth et al. 2014: 274).

Tastes can be quantified through digital intervention and measurements can be taken through various methods: a like button, subscriptions and sharing, while the methods mentioned are means that engage with and accumulate numbers. As Passoth et. al. express, associating qualitative aspects and referring to numbers often results in the oversimplified definition of a normally diverse range of elements, “quantities are able to express novelties just as convincingly as they simplify complex topics” (Passoth et. al. 2014: 273), the numerical simplification affects instilled class structures and is then distributed across a variety of classes that perform among Spotify’s platform. While the act of engaging with art and what type of art is a measurement or result of class, these distinguishable features and divisions are increasingly melded, mixed and reorganized through algorithmic processes, processes in which Spotify employs in an attempt to reach the maximum engagement with their users. Classes in the past have been indicators of rigid forms of status, both economically and socially, Spotify and the introduction of on-demand access to content is vast and continually evolving (reaching backwards and forwards in time through the history of music), are dissolving classes of musical preference and a transition towards a more spread-out and diverse audience is being observed. Lazzarato refers to the plurality of publics as the touching of brains (2006), “The cooperation between brains and of inter-cerebral relations, publics delineate fluctuations and bifurcations which reformulate the rigid and univocal segmentations represented by classes and social groups (Lazzarato, 2006: 11).

(18)

While the technology itself is a subject of class and access is granted through economic power and means, a reflective 70 million paying users on Spotify indicate that the population that is accessing its services is significant and continually growing. Instead, the focus on class is the dissolving of the rigid class formed within the musical realm. Since musical choices less likely made through the purchase of individual albums, the ability to simply pay a standard fee for numerous albums break down the limited access to different musical forms. Individuals no longer have to make the decision to chose one album over another since economic barriers are dissolved through lower costs and materiality, i.e. the reduction in the cost of producing a physical album, to selling them through record stores, to the unpredictable price of each individual album.

Within the context of musicking, algorithms begin to take responsibility for music taste management, tastes and classes are being reshaped and transformed through modes of algorithmic musical discovery, bringing to light the effects of engagement with music and putting into question the amount of autonomy a user has in the act of discovering music. An individual’s practice of physical displacement for desired cultural goods may be stifled and localized in regards to either mobile or stationary access. Spotify acts as a musical waiter, presenting, recommending and suggesting music; while this in turn spreads and casts a wider net of similar artists and musical genres, traditional practices of musicking are lost therefor signifying a shift in musical engagement and further obfuscating the lineation of class.

(19)

3.4 The effect on discovery and musicking

Spotify has introduced a service for music where an individual can have access to over 35 million pieces of music (Spotify, 2018), resulting in a singular source of music as opposed to a physical space (music stores) that operates with the individual sale of albums. Methods of music acquisition are altered, classes of musical taste are merged into a seemingly endless catalogue musical library, ways of picking and choosing music are no longer decided by physical displacement, monetary restrictions, and genre discovery is increased to individuals who would otherwise not venture into different musical categories. Spotify is in effect a nexus for musicking entirely changing the ways and forms that consumable music is engaged with. Fleischer claims this singular musical access point where “singularity, is reached when everything is accessible, one click away, creating an abyss for the user instead of a great possibility of multiple choices” (Fleishcher, 2009 as cited by Leijonhufvud, 2018 :230).

While the access to music is greatly increased, as well as musical discovery, traditional methods of musical engagement disappear and cease to become normalized, inviting different and digital methodological forms of musicking. Traditionally speaking, discovering music has been done through radio, concert going, music store aisles, possibly through word of mouth made by close associations. Perhaps the artist discovered would be entirely irrelevant and perhaps it would broaden the scope of related artists situated within similar genres. Spotify centralizes its content and does much of the musical discovery automatically, developing a detailed user profile as the individual continually navigates, interacts and engages throughout the interface.

Spotify gathers detailed information provided by users such as the number of listens a song may have, the number of followers an artist may have accumulated and is fed back to the user through the use of curated playlists, reflect the user’s engagement with the application. The result is an accumulation of numbers and statistics, very comparable forms of information which enables comparative measurements of qualities that have previously been impossible to compare and associate, “television perceives its audience as statistical aggregates, primarily in the form of the notoriously famous ratings” (Passoth, Sutter & Wehner, 2014: 276). An artist’s profile page among Spotify contains their top songs according to the number of views

(20)

The ability to associate numerical attributes and quantify popularity to music is a method that Spotify employs, ultimately affecting the way musicking is performed, “To speak of individual ‘viewers’ or ‘the audience’ is partially possible because of the techniques of making media usage observable and understandable” (Meyen, 2004; Schorr, 2000; Schrage, 2001;2005 as cited by Passoth, Sutter & Wehner, 2014: 274), essentially associating numerical value to popularity. Of course, popularity is the economic driver of music intended to make a profit; this may not be the case for all artists.

Music publications, the authoritative voice for what should be heard, create album reviews with rating systems, as well as annual lists for the ‘best albums’, is reflective of what Spotify produces with the statistical data provided by their users although in contrast to music publications, the authoritive voice are the users casting a demographic decision as to what is being listened to the most and what should be listened to according to large numbers. Spotify’s platform determines the ‘popularity’ of a track through the use of algorithms, “and is based, in the most part, on the total number of plays the track has had and how recent those plays are. Generally speaking, songs that are being played a lot now will have a higher popularity than songs that were played a lot in the past” (Spotify, 2018). This can be understood as high and new activity determining an artist’s popularity, while an artist with an older release and retaining high engagement might be positioned lower compared to newer and more recent productions.

Another method of determining an artist’s reputation is through the use and numerical value of ‘followers’, which is an afforded “way to tighten the connection between artist and fan” (cf. Baym, 2013, Wikström, 2009 as cited by Leijonhufvud, 2018: 258), further described by Spotify as the “listeners who hit follow or ❤ on your artist profile” (Spotify, 2018). This method is not only democratically dictated, it determines an artist’s follower influence, in contrast to

popularity, follower numbers encompass an artist’s overall existence on the Spotify platform. Numbers and ratings have traditionally been determined by the number of album sales and backed up by music publications, numerous sources that have the influential capacity to keep popularizing a specific artist. While this may still be the case today, Spotify is exemplifying that their use of accumulated data performed among its services, self-perpetuates and determines

(21)

what should be listened and brought to the user’s attention.One area of concern, within the field of musicking, is popularity or reputation based on a single source of large numerical statistics or ‘big data’, determined and produced solely from Spotify’s internal algorithms. As Boyd & Crawford describe how the use of big data,

“may involve the traditional concept of a human subject as an individual, or it may affect a much wider distributed grouping or classification of people. It fundamentally changes our understanding of research data to be infinitely connectable, indefinitely repurposable, continuously updatable and easily removed from the context of collection” (Metcalf & Crawford, 2014: 2)

While some numerical values associated with Spotify’s artists may be democratically determined other areas, such as the popularity values which are influenced by algorithmic outputs, traditionally alters an artist’s visibility of being heard.

Spotify’s ability to quantify every aspect of its content also marks a departure from traditional assumptions of popularity, such as the number of album sales, and is able to make an in-depth analysis of every track and artist as users engage with the content made available.

“The platform has data extraction built into its DNA, as a model that enables other services and goods and technologies to be built on top of it, as a model that demands more users in order to gain network effects, and as a digitally based medium that makes recording and storage simple. All of these characteristics make platforms a central model for extracting data as raw material to be used in various ways. (Srnicek, 2016: 53-54)

(22)

3.5 Quantifiable musicking

While the ability to quantify allows for a greater introspection of the data which is made available, it is argued that it disassociates human agency, “what an audience is or what it is supposed to be… so is what it means to distribute and perceive music” (Passoth, Sutter & Wehner, 2014: 271), which may neglect qualitative and individualistic tastes, in favour of a musical hierarchy and promotional visibility based on numerical systems where techniques of calculation and measurement are boundless. A system created in which “[algorithmic musicking] seems to overcome the restrictions of all special qualitatively oriented languages and comparative methods and become more of a universal technique of observation and comparison” (Heintz, 2010; Manhard, 2008 as cited by Passoth, Sutter & Wehner, 2014: 274). Bettina Heintz, in this case, states that the use of numerical quantification among an audience and therefore promotes the acceptance of communication (Heintz, 2010: 162).

Similar to musical taste, which is a quality that is difficult to determine, musical similarities found associations through genres, performative styles, musical eras, geographical settings. The quantified element of Spotify’s musicking facilitates associations through its afforded technical infrastructure, such as the affordance of having a mass musical catalogue in tandem with accumulated and valuable user data, which allows the “quantifiable capability to relate and compare similar events and developments within one thematic field” (Siemes, 2009 as cited by Passoth, Sutter & Wehner, 2014: 273).

The ability to measure and analyze their audience is an asset to Spotify and their business model, used in order to promote audience retention through increased music relevancy, eliminating technical features that were found to be redundant or counter-intuitive, in this sense, any form of statistical representation of data feedback only seeks to benefit the business model, “statistics do not just inform about individual current situations and differences, but also present changes of the respective observed units and the ups and downs connected to them” (Passoth, Sutter & Wehner, 2014: 274).

(23)

The concern arises where the promotion of creative content is promoted or made visible somewhere between the natural progression (human) or the technically and purely statistical influenced forms (non-human), such as algorithmic technology. Metcalf and Crawford reference Twitter about the computational concerns of the increasingly quantified methods of promotion, “data science methods create an abstract relationship between researchers and subjects, where work is being done at a distant remove from the communities most concerned, and where consent often amounts to an unread terms-of-service or vague privacy policy… These shifts are hard to quantify and ameliorate” (Zwitter, 2014 as cited by Metcalf & Crawford, 2014: 2).

Musical creation has traditionally been a human act and the interaction with its audience has for the most part been based on promotion, word of mouth, or radio visibility, “there is real urgency to define what a “human subject” is in Big Data research and critically interrogate what is owed to “data subjects” (Metcalf & Crawford, 2014: 2), meaning the concerns lie in the musicking loss of human intervention in favor of a non-human assemblage, makeup or entity, made up of numerically converted human involvement and interaction.

“Part of the difficulty is that the precursor disciplines of data science-computer science, applied mathematics and statistics- have not historically considered themselves as conducting human-subjects research” (Metcalf & Crawford, 2014: 2), which reconstructs the perception of individual taste in favor of technical assemblages within a realm of a traditionally human subject. “Even though statistics do ultimately represent people, research into math, computational capacity and other numeric modes of analysis rarely exhibited the types of human subjects concerns that are baked into research ethics regulations designed to handle the types of harms found in biomedical research” (Metcalf & Crawford, 2014: 2)

(24)

4.0 Algorithmic governance

Algorithms play an important part among the interaction between user and technology as the algorithm are created for their decision-making powers (Beer, 2009:4) similarly as a form of management, often offering insights and making decisions for “organizations “to allocate different levels of service to different users on an increasingly automated basis” (Graham, 2004:224). This puts into question the element of human agency (Beer, 2009:4) and the algorithmic governance that is established and put into place as the user interacts with the technology.

As for Spotify, algorithms play an important role in delivering content attempts to be highly relevant to the user. Forms of musical discovery are managed through algorithms that may present itself through methods such as receiving location-specific playlist, reflective musical content through the monitoring of user behavior, collaborative efforts with digitized businesses such as Spotify and, Facebook, Songkick or Shazam, which enables the transmission of information throughout exterior platforms stemming from Spotify.

“We believe that a key differentiating factor between Spotify and other music content providers is our ability to predict music that our Users will enjoy. Our system for predicting User music preferences and selecting music tailored to our Users’ individual music tastes is based on advanced data analytics systems and our proprietary algorithms. We have invested, and will continue to invest, significant resources in refining these technologies; however, we cannot assure you that such investments will yield an attractive return or that such refinements will be effective. The effectiveness of our ability to predict User music preferences and select music tailored to our Users’ individual music tastes depends in part on our ability to gather and effectively analyze large amounts of User data” (Spotify: US securities and exchange Commission, 2018: 38).

The predictive efforts that Spotify employs are clearly stated as a business model effort which enables algorithmic elements to drive, promote and navigate among its users. The algorithmic authority oversees which forms of taste an individual is more disposed towards and enables

(25)

functions that are not possible through human intervention, as calculations and profile data become too complex and large to manage and analyze. Algorithmic governance in this instance takes into consideration the users seemingly autonomous behaviour among the app, while Spotify is, in fact, curbing their vision and their business mission onto their audience. While a particular sense of autonomy is only possible through a subscription to Spotify’s premium subscription model, meaning, Spotify creating an environment where the user feels unbound by the company’s presence, the mission statement clearly aims to predict, tailor and analysis through algorithmic systems. Certain modes of autonomy are granted through the paid subscription such as access to music while being offline, the elimination of randomizing advertisements, and the ability to play an album through its original track order (non-subscribers may only listen to a randomised album).

“In addition, our ability to offer Users songs that they have not previously heard and impart a sense of discovery depends on our ability to acquire and appropriately categorize additional songs that will appeal to our Users’ diverse and changing tastes. While we have a large catalog of songs available to stream, we must continuously identify and analyze additional songs that our Users will enjoy and we may not effectively do so. Our ability to predict and select music content that our Users enjoy is critical to the perceived value of our Service among Users and failure to make accurate predictions could materially adversely affect our ability to adequately attract and retain Users, increase Content Hours, and sell advertising to meet investor expectations for growth or to operate the business profitably” (Spotify: US securities and exchange Commission, 2018: 38).

The concern for musicking automation is put forward into question as Spotify exemplifies the importance of retaining user attention through algorithmic methods. Traditionally, artist success, popularity and attention has been attributed to album sales, therefore acquiring recognition numbers according per sale and not requiring the constant return and attention that Spotify requires in order to deem their product valuable to them and their users. Authoritative algorithms are put into place across Spotify with the goal of achieving retentiveness my means of technical analysis of music’s qualitative and conditional characteristics.

(26)

“We have the ability to personalize and curate the content we stream by measuring an individual User’s preferences against more than 40 different parameters. Our algorithms are designed to anticipate a User’s preferences using factors such as demographics and past listening behavior. Furthermore, we can combine situational context, such as time of day and location, to make better recommendations for appropriate content to an individual User based on his or her current activity. As we further collect and process data and understand our Users, we believe our technology will better understand and respond to our Users’ preferences, drive an even better overall User experience, and further differentiate our Service from our competitors” (Spotify: US securities and exchange Commission, 2018: 38).

Spotify is overt about the algorithmic methods that are deployed, which claims and seeks to benefit their user base. Methods of managing music content are examples of how musicking may be influenced by technical factors, altering traditional musical interactions in favor of an increased technical influence. What Spotify has introduced are algorithmic forms of musical discovery. While it is possible to reject recommendations, the user has the freedom to research and retain their own music, much of the applications infrastructure supports forms of automated listening (daily mixes, mood playlists, radio recommendations, concert suggestions based on interacted music). This form of algorithmic musicking “plays an important role in selecting what [music] is considered most relevant to us, which is a crucial feature of our participation in public life” (Gillespie, 2014: 1). The act of actively seeking and searching for music can now be an automated process, effectively eliminating manual methods in which music has traditionally been interacted with, an example is a shift from analogue to the digital. To assert autonomy upon the act of musicking, in this case any engagement with music, is to “[give up] the primacy to that of which the artist is master… i.e., form, manner, style, rather than the subject, the external referent, which involves subordination to functions – even if only the most elementary one, that of representing, signifying, saying something” (Bourdieu, 3:1984). Bourdieu is pointing to the primacy of the “master” or the individual who is musicking, to be unbound, freeform and away from institutional framing. To autonomize these acts would do away with this primacy resulting in a loss of human-related factors and considerations. To give up autonomy in favor of the automated according to

(27)

Bourdieu, is “the refusal to recognize any necessity other than that inscribed in the specific tradition of the artistic discipline in question: the shift from an art which imitates nature to an art which imitates art, deriving from its own history, the exclusive sources of its experiments and even of its breaks with traditions” (Bourdieu, 3:1984), the continual cycle of copying and re-copying eventually lose its originality, stemming further away from its traditional roots. Certain aspects of musicking, concerning the historical and the traditional, may be lost in the quest for automated forms of musical interactions.

(28)

4.1 Algorithmic presence

As algorithms begin to learn the user-end outputs an algorithmic entity begins to take form. A collection of human feedback in tandem with managerial formulas introduces an algorithmic presence that is continually engaging with the act of musicking. In this sense, algorithms produce the power to introduce truths in two ways : first, through material interventions, as ‘algorithms produce outcomes that become or reflect wider notions of truth’ (Beer, 2016 : 8) and second, algorithms are a notional presence in discourse, observing how the ‘term or notion is deployed to create or perpetuate certain truths about social orders or how certain truths are cultivated through discussions or evocations’ (Beer, 2016: 8), forming patterns of expected behavioural outputs therefore overriding traditional means in favor of more technically measured and tested methods. Further explaining the implementation of algorithmic behavioral elements, according to Scott Lash (2007b), "the ‘stuff’ that makes up the social and urban fabric has changed – it is no longer just about emergent properties that derive from a complex of social associations and interactions. These associations and interactions are now not only mediated by software and code they are becoming constituted by it” (Burrows, 2009 as cited by Beer 2009: 987).

It is possible to compare ‘the algorithm’ as a musical determining presence, bringing forth suggestive forms of musical selection, learning from both Spotify’s platform and its human correspondent, which the user, working in tandem forming a musical cyborg, which defined by Leijonhufvud as, “the interrelationship between man and machine… difficult to separate from each other, and further, impossible to determine who is prompting whom” (Leijonhufvud, 2018: 247), which further dictates the musical direction that takes place among its music catalogue. An algorithmic presence can be influential and therefore have the power dictate general tendencies. An important consideration is the preserved human agency that is a necessary component to complete an algorithmic objective, ultimately influencing an outcome. What the outcome represents is a collective participation of what content is being engaged with the most, which is taken into consideration and then re-distributed to be further engaged, in addition to experiencing waves of constant or varying degrees of popularity.

(29)

"We've made the recommendations a lot smarter. The big paradigm of sites within music discovery has been editorial versus the algorithms. But the more we thought about it, actually they're not mutually exclusive. You can marry them together." (Spotify, 2018)

Understanding the integrated algorithms implemented by Spotify is not a subject that is divulged, Pascal describes this lack of information provided by mediators as “black-boxing” (Beer, 2016:4). This black-boxing in effect, masks the “values and prerogatives, the encoded rules, [that] are hidden within black boxes” (2016:4), and makes an approach towards analyzing the intermediator and its algorithmic process difficult to observe, an issue that is concerning to Beer as this “opens up questions about the role of algorithms in the deployment or expression of power” (2016:4). With a catalogue of music of 35 million tracks, Spotify CEO Daniel Ek claims that Spotify is “in the discovery business”. A suggestion that algorithms that are shaped through its users will aid in the profiling and categorization of music in ways not previously conceived. The production and consumption of musicking have been altered.

(30)

4.2 Algorithmic social power

Algorithms facilitate communication and form unintended relationships, establishing an extensive mapping of the musical and artistic realm. A melding of various characteristics and categorizations are processed and managed accordingly, to the benefit of a number of artists under the umbrella of a record label, to a specific niche or trend that are linked by similarly associative qualities. As Lash defines the era of a ‘new new media ontology’ (Lash, 2007), Beer attributes the involvement with algorithms to creating “new forms of power in the context of apparent empowerment and democratization’ (Beer, 2009: 986) meaning there is benefit and new found possibilities of the social and technical collaboration forming likelier associations, ultimately empowering the user.

Artists are able to take advantage of Spotify’s technical affordances; such as connect with the various artist’s through their related artist infrastructure. Much like social networking and creating new webs of associations, Spotify offers connected similarities for artist’s that may be unintended and unexpected, power and formed through Spotify’s multiple algorithms. Artist’s are also granted an artist portal, which allows a detailed and technical analysis of their performance among Spotify, further granting the power to alter their behaviour among the platform accordingly. Beer sees these new forms of power among the Web-era of user-generated content (Post-Web 2.0), as an “apparent [form of] ‘empowerment’ and ‘democratization’ as ‘the people’ apparently reclaim the internet and exercise their ‘collective intelligence’.(Beer, 2009: 985-986), forms of power that allow artists to branch out into different or similar associations, ones made not only through agency but also through algorithmic components.

A user’s use of Spotify’s technical affordances is therefore reflected back as the engagement of musicking tastes and preferences facilitates or encourages what content will be engaged with. Spotify offers to their users the opportunity to create their own playlist, to ‘follow’ a desired artist (meaning they would continually be updated the moment displays any activity), to “❤” an artist (meaning the artist would show up in an algorithmically curated playlist), and the act of engaging with a specific artist ultimately determines the artist’s popularity which influences their visibility and popularity. Beer sees this integration of algorithmic participation as a point of consideration

(31)

“to the implications of software ‘sinking’ into and ‘sorting’ aspects of our everyday lives” (Beer, 2009:987), an increasingly ingrained normalization of an algorithmic participation of musicking. With all the engagement occurring among the platform, Spotify is constantly accumulating data from every interaction that takes place by each subscribed actor. The data that is amassed and reintegrated back to the participating parties are actions that draw considerable power and influence. Beer employs importance to the consideration of the “implications of software ‘sinking’ into and ‘sorting’ aspects of our everyday lives” (Beer, 2009:987). With this in consideration, Spotify’s accumulation of user data is an authoritative source of power, enabling the implementation of user preferences and actions, which are willing given, and redistributed throughout the algorithms and more broadly throughout the platform. This algorithmic entity is representative of a ‘new new media ontology’ (Lash, 2007 as cited by Beer, 2009: 998) referring to the scope of the Web 2.0 beyond just media but into knowledge, which is a shift towards forms of living in which information becomes active in shaping lifestyles and environments (Beer, 2009:988). While Spotify user’s express their authoritative ways of engaging with the platform, their actions which are reflected back through algorithmic processes work to reinforce and further dictate decisions of taste and dispositions towards a specific genre or artist for example.

By means of social and individualistic expression among Spotify, the algorithmic elements that reflect back onto the users mark a shift or according to Lash, “the collapse of ontology and epistemology” (Lash, 2006: 581), where the ontology of being and the epistemology of knowing is affected by the reconfiguration of the human agency, through technological means. This indicates that the lines or borders of human agency are being blurred or obfuscated by the unclear delineation of human agency and algorithmic influence, where information technologies ‘comprise’ or ‘constitute’ rather ‘mediate’ our lives (Beer, 2008: 988). Technological interactions are lost to “complex social associations and interactions… now not only mediated by software and code, they are becoming constituted by it (Lash, 2009 as cited by Beer, 2008: 987).

While a coordination of social movements to favour, alter or influence a specific artist among Spotify’s platform, it has been observed that Spotify’s popularity rating algorithm works in tandem with an artist’s most recent activity and the amount of user engagement it receives, indicating a

(32)

potential for social effect. While not being a platform of outright social expression, the social and individual interactions with Spotify grants societal and data-centric value to not only Spotify but also to the artist’s, the record labels and all the parties associated that may find some use for the data afforded by Spotify. It is understood that the social and the algorithmic are both at play, it should be further noted that the implemented algorithms facilitate and encourage a social engagement through a seemingly democratic form of engagement.

Platforms performing and acting as companions, continually servicing the user, may indicate how behaviours translate themselves, through thorough examination resulting in measurable quantifications of qualitative descriptions. Much like a continual line of circular communication and information serving to assist and enhance, acted upon by the user and afforded by technology. To a degree, the democratic self-perpetuation of data labour and the platforms’ capitalist intent to grow indefinitely, grants access to a great deal of public information as well as thrusts upon the platform creators a great deal of corporate responsibility. As Tizian Terranova draws on the concept of ‘collective intelligence’ (Terranova. 2004: 86), in this case the collective of Spotify user’s, who form a ‘general intellect’ as an assemblage of humans and machines at the heart of post-industrial production (2004: 87) and emphasizes how important free affective and cultural labour may be to the media industry old and new (2004: 88).

(33)

4.3 Musical Hierarchies

An assumption that can be made about Spotify platform influence is that it may enforce musical hierarchies within the music industry, meaning the upholding and favouring the popular, which promotes visibility and takes over digital space afforded by the platform. Artists are not only involved in the musical creation process, other parties are also responsible, engaged and have invested interests throughout creation process as well as after the ensuing results. Music labels and artist management are responsible for signing bands to contracts for the benefit of strengthening their label image and capital gains. Also involved are publishers, who are responsible for the exposure of an artist, their public image and public reach. Producers, who assist in the production of an album, influencing certain aspects of an artist’s sound. Promoters, responsible for booking artists and venues. Clearly, there are many individuals who are responsible for extending an artist’s public outreach, working behind the scenes and out of sight from the artist’s image. A signed artist’s existence goes through many ‘hands’ before it is released and heard by the public, including the ways an artist manages to be heard on Spotify.

Playlisting is a method of exposure that is beneficial to an artist on Spotify. While Spotify does offer algorithmically curated playlists, other actors are involved in the creation of playlists such as Spotify editors, artist publishers, artist’s themselves and Spotify users. The end result are playlists based on a hierarchical list of the most popular songs among specific playlists, popularizing certain artists through playlisting affordances offered by Spotify. The popularization creates a hierarchical system that can either be natural or influenced according to Kevin Breuner, meaning a Spotify user may have the autonomy to create their own playlist, follow an artist’s promoted playlist, follow Spotify’s in-house editorial curated playlist’s, follow the algorithmically generated Discover Weekly or follow a promoted playlist created by record labels who promote their own artists (Breuner, 2017).

Another hierarchy that can be found within Spotify is the visible number of plays, or a top play song, according to an artist. The highest performing or listened-to songs will generally be promoted on an artist’s main page causing a reaction by the user to base assumptions according to the numbers seen, be it in the thousands and even in the millions. Assumptions such as associating

(34)

numbers to the “best” song, or the most popular song, these are methods that primarily engage users into the further discovery of an artist. As Passoth and Wehner describe visual numbering cues and the acceptance of numbers, “[Numbers] somehow provide the impression of something preliminarily final. They offer a basis for connecting operations. […] Unfortunately, however, numbers only have two options: to increase or to decrease. Everything else is an ingredient, an interpretation” (Luhmann, 1981: 327). It is essentially and traditionally equivalent to an artist’s single, meaning an artist’s calling card song, the song which describes the artist’s style, aura or entity, although, in the instance of top songs among Spotify, the list is usually composed of five songs. “number systems enable us to observe and assess highly complex situations and to become involved in them with others-without having to be a participant or a person directly affected (Porter, 1995 as cited by Passoth, Sutter & Wehner, 2014: 274). An area of observation is that the numbers of plays generated by Spotify’s platform may be influenced and skewed, through previously mentioned methods of promoted playlisting, which enables visibility. Although a facet of producing music is for it to be heard, promoted and popularized, influenced numbers, in this case, enable the overshadowing of an artist over another. Tania Bucher discusses the friendship assemblage (2013) using Facebook as an example that challenges traditional forms of friendship and by highlighting the introduction of software assisted friendships, “While the traditional notion of friendship highlights the voluntary and durational aspect of becoming friends and becoming friends anew, the software, one may claim, encourages and functions as a suggestive force that “pushes” users to connect with the people he or she may already know according to the algorithm” (Bucher, 2013: 486). While traditional forms of musicking are always possible, applications such as Spotify’s software, has a digitizing effect, such as the involuntary introduction to an unheard artist or the suggestive artist recommendations based on a user’s specific qualitative or quantitative data.

What is not visibly seen are the hierarchical structures that are put into place through artist similarity search results, in reference to algorithmically influenced playlisting, songs are traditionally positioned as hierarchical, selected based on user preference. What is hidden is Spotify’s artist popularity formula and the emphasis being able or in this case, not being able to observe if an artist is referred to the user because either of their qualitative qualities or their numerical properties, which allows Spotify to be the ultimate promoting force of what should be

(35)

heard and experienced. As part of the methodology, observations will be made concerning the promotion of visibility based on hierarchy in hopes to answering if Spotify’s algorithmic elements reinforce or diversify its content.

Similar to the quantification of individual taste, the numerical values associated to hierarchies’ position artists in opposition and competition amongst each other as hierarchical structures promote visibility and numerical power. Passoth et. al. go into further detail, “whoever wants to know how successful the others are in their relevant field only has to look at their ratings, sales figures or circulation. Viewed in this light, audience statistics introduce a specific kind of competitive relationship between content providers by allowing a permanent comparison of their publicly accessible performance” (Passoth, Sutter & Whener, 2014: 277).

As Spotify’s community composed of users and artist’s continue to grow, the company becomes the gateway to an economy of attention and visibility, creating a landscape for comparison, competition and popular opinion, which in a domain that strives for attention, promotion and capital gain, Spotify is increasingly responsible for being the intermediator and the singular source of musical engagement and musicking, Passoth et. al. describe the potential and value of Spotify’s accumulated information as, “every actor involved is informed of its own as well as the other’s strengths and weaknesses and encouraged to self-optimize, which can lead to the potentially worldwide circulation of successful formats and content” (Passoth, Sutter & Whener, 2014: 277). The musical landscape created by Spotify is fertile grounds for new forms of musicking and the repositioning of artist hierarchies across digital forms of visibility and popularity, which may create instances of the reliance of information given by singular sources, marking the potential for platform dominance through value and information, among a specific field or sphere, in this case speaking of the platform capitalism of the digitized musical realm.

Quantifying Taste: Spotify's algorithmic effect on musicking

MSc New Media & Digital Culture

Media Studies

Master Thesis

QUANTIFYING TASTE

Philip Creamer

11636254

Supervisor/Examiner: Assessor:

Dr. Bernhard Rieder Dr. Esther Weltevrede

Table of contents

1.1 Acknowledgements

Philip Creamer

2.0 Introduction

3.0 Musicking and Musical Taste

3.2 Musical Taste

3.3 Taste according to class

3.4 The effect on discovery and musicking

3.5 Quantifiable musicking

4.0 Algorithmic governance

4.1 Algorithmic presence

4.2 Algorithmic social power

4.3 Musical Hierarchies