The promise of personalisation: Exploring how music streaming platforms are shaping the performance of class identities and distinction

Abstract

Not only do music streaming platforms offer on-demand access to vast catalogues of licensed music, they are actively shaping what and how it finds us through personalisation. While existing literature has highlighted how personalisation has the potential to transform the part that music taste and consumption play in the performance of class identities and distinction, little is empirically known about its sociological consequences. Drawing on 42 semi-structured interviews with a combination of key informants and Spotify users, this article demonstrates that personalisation is undermining opportunities to achieve social distinction by taking over the labour of music curation and compressing the time needed to appreciate music for its own sake. It demonstrates that those with cultural capital at stake – in the case of this study, young, (primarily) male cultural omnivores – experience personalisation as a threat, highlighting how particular claims to social distinction are being contested in the platform age.

Keywords

Bourdieu curation datafication distinction music consumption music streaming personalisation platformisation social class taste

Introduction

From social networking sites to on-demand streaming services, digital platforms have disrupted marketplaces globally and entrenched themselves as powerful economic, political and social actors. They are contributing to the rapid restructuring of marketplaces, social relationships and forms of value creation and capture (Srnicek, 2017). Streaming platforms, such as Netflix, Spotify and Amazon Prime, have made available vast catalogues of film, television, music, podcasts and other digital media, at little or no cost, while sharing platforms, such as Airbnb, Uber and TaskRabbit, have created marketplaces for ‘ordinary’ people to make money by exploiting their property, vehicles and physical labour.

The dramatic rise of digital platforms has been met with a growing body of academic research about what platforms are and how they are shaping social and cultural life (Duffy et al., 2019; Langley and Leyshon, 2017; Meier and Manzerolle, 2019; Nieborg and Poell, 2018; Poell et al., 2019; Prey, 2020; van Dijck, 2013; van Dijck et al., 2019). Digital platforms are multinational businesses that provide the sociotechnical means (i.e. the software, interfaces, licensing deals) for different actors to interact. Through governance and infrastructural ownership, digital platforms manipulate what and how information, people and cultural goods circulate to ensure interactions are optimised for the platform’s desired ends, such as creating platform dependency or selling people’s attention to third-party advertisers (Poell et al., 2019; van Dijck et al., 2019).

Personalisation – the large-scale manipulation of what and how information, people, goods and services are selected, arranged and presented to individuals – is an important characteristic of platformisation (Fuez et al., 2011; Pariser, 2012; Srnicek, 2017; van Dijck, 2013). From Google’s search engine results to Facebook’s news feed, personalisation is used to create engaging user experiences and power digital platforms’ advertising businesses. Personalisation relies on the creation of immense troves of digital data about the identities and interactions of platform users and the application of big data analytics to make predictions about future behaviours, needs and desires (Cheney-Lippold, 2011; Drott, 2018; Prey, 2016). As digital platforms continue to capture people’s attention and accumulate immense – and increasingly heterogeneous – troves of digital data, their capacity to achieve personalisation in new and intimate ways grows.

Music streaming platforms, such as Spotify and Pandora, are at the forefront of innovations in personalisation (Drott, 2018; Hracs and Webster, 2020; Prey, 2016; Prior, 2018; Wikström, 2020). These platforms have invested heavily in personalisation and leverage it as a means for attracting and engaging users, outcomes integral to the commercial success of their ad-supported and subscription-based business models (Hracs and Webster, 2020; Wikström, 2020). Combining the expertise of playlist editors and the latest advancements in music recommendation technologies, music streaming platforms have developed suites of editorially curated and personalised playlists and single-item recommendations, all of which are selected, arranged and presented to users on an individualised basis across the microspatialities of their platforms. To illustrate the immense scale at which music streaming platforms personalise, Spotify produces over 320 million ‘Discovery Weekly’ playlists, a uniquely personalised playlist of new music, every week, and Apple Music produces a similar ‘New Music Mix’ for its 60 million plus subscribers also on a weekly basis.

Personalisation raises fundamental questions about the role that music taste and consumption play in the performance of class identities and pursuit of social distinction. According to the influential sociologist Pierre Bourdieu (1984), cultural taste is a product and producer of class divisions in society. Taste is one of the assets – what Bourdieu (1984, 1986) calls ‘cultural capital’ – underpinning class privilege. Cultural capital refers to knowledge, skills, tastes and mannerisms acquired formally through education or informally through friends, family and the consumption of culture, and it sits alongside economic (e.g. financial wealth) and social (personal and professional networks) capital in Bourdieu’s (1986) conceptual framework. The accumulation and conversion of capital is central to Bourdieu’s theories about how class divisions are reproduced. Being able to consume the ‘right’ culture in the ‘right’ way (i.e. what is socially accepted as legitimate) has the potential to be converted into different forms of economic and social advantage, such as helping people to network and gain employment (Bourdieu, 1984, 1986; Friedman et al., 2015; Savage et al., 2013).

While Bourdieu (1984) identified tastes for highbrow culture (e.g. classical music) as a marker of class privilege, accounts have changed in recent years. In Performing Rites (1996), for example, Frith highlights that popular music knowledge and discriminatory skill also have the potential to represent cultural capital, producing the same hierarchical effect on what is counted as ‘legitimate’ popular taste, thereby (re)producing social divisions between people with the time and money to invest in the accumulation of such knowledge and those without. More recently, sociologists have demonstrated that ‘cultural omnivorousness’ (Peterson and Kern, 1996) – the consumption of diverse cultural forms, spanning highbrow and popular culture – is central to how social distinction is achieved through taste and consumption.

The original claim of the omnivore thesis has since been questioned. For example, Savage (2006) observes the diminishing effect of class on patterns in music preferences with increasing diversity across the board, also highlighting that age and gender may be stronger differentiators of musical tastes than class background (Rimmer, 2012; Savage and Gayo, 2011). As such, omnivorousness may not be an exclusive trait of privileged individuals and groups, which Wright (2011) argues is indicative of the increasing availability and affordability of cultural goods in the 21st century. While these findings may suggest a weakening relationship between taste, class and consumption, attention to differences in how people consume culture has highlighted the ongoing importance of class identities to taste and consumption. Qualitative research has demonstrated that cultural omnivorouness is better thought of as a ‘mode of consumption’ (Jarness, 2015), characterised by an openness to and command over diverse cultural forms (Bennett et al., 2009; Prieur and Savage, 2013; Rimmer, 2012; Savage and Gayo, 2011; Warde et al., 2008). It is a claim to social distinction performed by capital-rich consumers through consumption practices and attitudes, acting to legitimise social differences whether consciously and deliberately, or not.

Against a backdrop of increasing platformisation and personalisation, sociologists have considered how the relationship between taste, class and consumption might again be changing (Barna, 2020; Beer, 2013; Prey, 2016; Prior, 2018; Webster, 2019; Wright, 2015). Whereas for Bourdieu (1984) social background plays an important part in the formation of taste, with our families, friends and schools exposing us to different ideas, values and cultural norms, music streaming platforms are employing big data analytics to make judgements about what music is relevant to individuals, tastes and lifestyles (Cheney-Lippold, 2011). As such, Beer (2013) suggests that personalisation has the potential to detach the cultivation of taste from class-related socialisation processes, such as friendship networks and the consumption of the right type of broadcast media, while others raise important questions about who (or what) has the authority to make judgements about what counts as ‘good’ or ‘bad’ taste, influencing how hierarchies of taste are formed and related to class identities (Barna, 2020; Bonini and Gandini, 2019; Morris, 2015a; Webster et al., 2016).

Other debates allude to Bourdieu’s (1984) homology thesis, which claims that individuals with similar class backgrounds typically have tastes for similar music due to shared class experiences (Prey, 2016; Prior, 2018; Wright, 2015). The reliance on digital data about past user behaviour to generate future recommendations has the potential to reproduce class divisions in music taste in automated and pervasive ways through the creation of ‘filter bubbles’ (Pariser, 2012). In commercial contexts, music recommender systems are optimised for the goal of driving user engagement (Jannach et al., 2016; Seaver, 2018), which can result in giving people more of what they are familiar with, rather than introducing music which is beyond the horizons of a person’s recorded interests. This has the potential to culminate in a ‘technologically-managed form of homology’ (Wright, 2015), meaning the structural similarities and differences in the tastes of different social classes have the potential to be algorithmically reinforced, rather than promoting change. Indeed, Fuez et al.’s (2011) research on Google’s search engine has demonstrated that personalisation is not very subtle. They highlight how it is limited in its ability to make the long tail of content available, suggesting that music streaming platforms may not be able to diversify the listening of individuals and groups.

While this literature usefully sets an agenda for future research and helps us to think about the structural implications of platformisation and personalisation, little is known about if and how people are incorporating or resisting the affordances of personalisation in everyday life. The everyday use of music streaming platforms has become a subject of scholarly attention (Anderson, 2015; Fuentes et al., 2019; Hagen, 2015; Hanrahan, 2018; Johansson et al., 2018; Kjus, 2016; McCourt and Zuberi, 2016; Sinclair and Tinson, 2017). Fuentes et al. (2019) explore how people have incorporated abundant access to music into their everyday lives. They highlight how music streaming platforms have enabled people to extensively ‘soundtrack’ their everyday practices to ‘. . . help accomplish the normative goals of those practices’ (p. 499), what Kassabian (2013) describes as ‘ubiquitous listening’. Sinclair and Tinson (2017) and Hagen (2015) explore the strategies adopted by individuals to manage the ‘overwhelming’ amount of music made available by music streaming platforms. Sinclair and Tinson (2017) demonstrate that playlist curation (the creation of ordered or unordered lists of tracks, a functionality provided by music streaming platforms) is a way to cultivate a sense of psychological ownership in a context of immaterial and abundant access to music, while Hagen (2015) similarly highlights how user control is an underlying motivation for the different ways playlists are created and enjoyed, serving as a representation of what is unique and individual.

Yet, these accounts pay limited attention to the personalisation performed by music streaming platforms, the everyday performance of class identities and the pursuit of social distinction. Not only do music streaming platforms offer up the possibility for people to consume any style of music, they have the potential to shape how people consume it. Music streaming platforms open up possibilities for individuals to outsource the labour of music discovery and the ‘soundtracking’ of activities and emotions to the human and technical actors involved in personalisation (Beer, 2013; Burkhart, 2008; Morris, 2015a; Prior, 2018; Webster, 2019). Given the ongoing importance of culture to the formation and reproduction of class divisions in society (Friedman et al., 2015; Savage et al., 2013), it is important to consider if and how the affordances of the digital platforms through which millions of people now access and engage with recorded music are incorporated into how people reproduce class differences through what and how they consume music.

Drawing on 42 semi-structured interviews with a combination of music industry key informants and Spotify users from across class backgrounds, this article begins to address these gaps. It demonstrates that the promise of personalisation – the opportunity for people to outsource the labour of music discovery, collection and curation – is challenging how some individuals seek to achieve social distinction through the performance of accumulated knowledge and discriminatory skill associated with cultural omnivorousness. This is because personalisation is both contesting the need for musical expertise and undermining opportunities to take time to appreciate music for its own sake. However, personalisation is not experienced in the same way by everyone, with perceptions shaped by generational and gendered differences in music consumption practices. This article demonstrates that those who have cultural capital at stake – in the case of this study, young, (primarily) male and middle-class cultural omnivores – perceive personalisation as a challenge because they have something to lose. Their musical expertise and discriminatory skill are perceived as being replaced by the personalisation performed by music streaming platforms.

This article makes several important contributions. First, by examining why and how music streaming platforms personalise, it contributes to existing accounts about what digital platforms are and how they intervene in social and cultural life (Duffy et al., 2019; Langley and Leyshon, 2017; Meier and Manzerolle, 2019; Nieborg and Poell, 2018; Poell et al., 2019; Prey, 2020; Srnicek, 2017; van Dijck, 2013; van Dijck et al., 2019). Second, by exploring how people incorporate personalisation into their everyday practices, it contributes to the burgeoning literature on the everyday use of music streaming, which to date has largely focussed on individual responses to abundant access to music, rather than personalisation (Anderson, 2015; Fuentes et al., 2019; Hagen, 2015; Hanrahan, 2018; Johansson et al., 2018; Kjus, 2016; Sinclair and Tinson, 2017). Third, this article contributes to debates about class, culture and consumption in the 21st century, including the qualitative research about cultural omnivorousness as a mode of consumption and claim to social distinction (Bellavance, 2008; Bennett et al., 2009; Friedman et al., 2015; Jarness, 2015; Rimmer, 2012; Savage et al., 2013; Savage and Gayo, 2011; Warde et al., 2008). It provides a much-needed empirical and qualitative account of how changes in the availability of recorded music are shaping how class identities and distinction are performed through music taste and consumption (Barna, 2020; Beer, 2013; Prey, 2016; Prior, 2018; Wright, 2015).

After reviewing the methodology, this article provides an empirically informed account of why and how music streaming platforms achieve personalisation, the foundation upon which the remainder of the article builds. This is followed by three empirical sections which demonstrate that the promise of personalisation is in contention with how people seek to achieve distinction by (1) demonstrating their musical knowledge, (2) taking time to appreciate music for its own sake, while (3) highlighting that personalisation is perceived differently depending on who has cultural capital at stake. The conclusion highlights the need for ongoing research to continue to refine our understanding of how platformisation and personalisation are shaping the social dynamics of music consumption.

Methodology

The empirical material presented in this article draws on data collected from 2016 to 2018.

First to investigate why and how music streaming platforms invest in personalisation, 22 semi-structured interviews, lasting between 45 and 60 minutes, were conducted with key informants. These individuals work as playlist editors and data scientists at leading music streaming firms, executives at major record labels, and journalists covering the music industry. Participants were selected purposively based on their expertise about and professional relationship with music streaming platforms. Participants were identified by reviewing company websites and LinkedIn pages and recruited through prospective contacting via email. Snowball sampling was subsequently used to facilitate the recruitment of additional key informants. All key informants are anonymised.

Although interviews generated detailed data and allowed respondents to express experiences and opinions in their own words, why and how personalisation works is a closely guarded secret and respondents were reluctant to answer some questions. Therefore, interviews were supplemented with the collection and analysis of 120 documents, including promotional and marketing material, technical publications, and media articles about leading streaming firms, their strategies and the marketplace more broadly. Documents were sampled purposively and collected from the Web using the search engine, Google. Combinations of key terms, such as ‘music streaming platforms’, ‘Apple Music’, ‘Spotify’, ‘curation’ and ‘consumption’, were used to seek out relevant documents.

Second, to explore how personalisation is shaping the social dynamics of music consumption, a further 20 semi-structured interviews, lasting between 45 and 60 minutes, were conducted with Spotify users. I used occupation and education as a proxy for class background and targeted different occupations found at a British university (e.g. managerial professions and manual labourers) to recruit a mixed sample. My sample consisted of ages ranging from early 20s to late 40s, 13 women and 7 men, 15 university-educated and occupations ranging from customer service advisors to IT engineers. My sample did not aim to achieve statistical representativeness. Rather, in line with my culturalist approach to class analysis (Savage, 2000), I explored the individual experiences of a mixed group of people to point to potential ways in which music streaming platforms are shaping class and distinction, helping me to set an agenda for further research. Building on the qualitative literature about cultural omnivorousness as a mode of consumption and claim to social distinction (Bennett et al., 2009; Jarness, 2015; Prieur and Savage, 2013; Rimmer, 2012; Savage and Gayo, 2011; Warde et al., 2008), my data collection focussed on social differences in how people consume music, exploring how consumption practices have been shaped by the affordances of platformisation and personalisation. All participants are pseudonymised.

To probe deeper into the experiences of platformisation and personalisation, I conducted app ‘walk alongs’ (Light et al., 2018). Participants were invited to bring along their own music device (e.g. smartphone) and individuals were encouraged to discuss their feelings about different aspects of their chosen service, such as the platform’s attempts to personalise the user experience through individually tailored home screens, playlists and recommendations. This method facilitated triangulation – comparing what participants said about their experiences using Spotify with what was observed.

Data analysis involved a systematic process of coding and re-coding. Each transcript/document was analysed phrase by phrase, while thematic codes, annotations and reflective notes were added. After this ‘open coding’, the data were organised into categories which corresponded to the themes and questions from the interview guides, literature, annotations and reflective notes. A process of axial coding followed through which connections between and within categories and subcategories were made. At this stage, some codes and subcategories ‘broke down’ while others emerged as more pervasive or poignant across the sample. I then moved towards identifying preliminary theories and collapsing categories into overarching themes through an iterative process of moving back and forth between the data and the research questions, interview guides and literature.

In the following section, I explore the affordances of personalisation, why music streaming platforms have invested in it and how personalisation is achieved. As I go on to demonstrate, there are specific aspects of how music streaming platforms personalise, such as the optimisation criteria used to evaluate personalisation, which have implications for how my participants engage with recorded music.

The promise of personalisation

Why do music streaming platforms personalise?

The business models of commercial radio and music streaming are both built on the premise that music is ubiquitous and amorphous, allowing it to be channelled, monetised and monopolised according to particular interests (Sinnreich, 2015). They do not rely on scarcity to generate commercial value; rather, they accept that recorded music has been ‘unbundled’, and they can no longer control its flow. Instead, they seek to persuade people to continue tuning in or paying for a subscription through quality, convenience, branding and curation (Anderson, 2014; Hracs and Webster, 2020; Sinnreich, 2015; Wikström, 2020), creating what Morris and Powers (2015) describe as ‘branded music experiences’. In turn, Arditi (2018) argues that streaming firms seek to engage individuals in an ‘unending’ cycle of consumption, where a platform’s control over the sociotechnical means for accessing music has the potential to trap individuals into perpetually paying for music. By promising to make it easier to find and engage with relevant music, personalisation is one competitive strategy adopted by music streaming platforms to attract and engage users (Anderson, 2014; Hracs and Webster, 2020; Wikström, 2020). As one key informant explained,

Their ultimate goal is to keep you using Spotify, keep you using Apple Music, that, they need you to be there listening to their free tier or paying for a subscription. And the more that they can show you that they understand what you are looking for, or the more that they can find new music that you really enjoy, the more you are going to be tied into their service and enjoy using it. (Interview, Director, Recorded Music Trade Body)

Personalisation promises to help individuals to find music to complement different activities, moods and intents, while helping creators and rights holders reach relevant audiences (Anderson, 2015; Hracs and Webster, 2020; McCourt and Zuberi, 2016; Morris and Powers, 2015; Wikström, 2020). It seeks to help listeners overcome the effort and anxiety introduced by the ever-increasing amounts of music made available by platforms. For example, both Spotify and Apple Music have created discovery-orientated personalised playlists, ‘Discover Weekly’ and ‘My New Music Mix’, respectively, to take the effort out of exploring new music. Reminiscent of radio programmers who blended quantification, institution and commercial interests when selecting what and when to broadcast music (Razlogova, 2013), music streaming platforms have assembled teams of playlist editors to create and manage suites of branded playlists, organised around activities and emotions, complementing the ‘ubiquitous listening’ that characterises contemporary music consumption practices (Kassabian, 2013). Both the editorial and personalised playlists and recommendations are distributed to consumers on a personalised basis. For example, Spotify’s ‘Home’ and Apple Music’s ‘For You’ are personalised pages presenting a plethora of editorial and personalised content based on patterns in past listening and contextual factors, such as the time of the day. In doing so, Eriksson et al. (2019) argue that listeners are encouraged by music streaming platforms to form an affective and intimate relationship with the service that will make their life better, relying on it to provide the right music for every mood and moment. Or as one key informant put it,

Spotify’s recommendation engine is great at understanding Rebecca and it has spent the last 12 months understanding Rebecca’s manifest musical DNA. And then on a weekly, monthly, quarterly, yearly basis it will serve Rebecca recommendations based on Rebeccaness, which from a user experience point of view kind of feels good and that speaks to the DNA of a 21st-century service because services make Rebecca’s life better. (Interview, Senior Director, Major Record Label)

Personalisation is also important to the advertising businesses of music streaming platforms (Anderson, 2015; Drott, 2018; Eriksson, 2020; Hracs and Webster, 2020; Prey, 2016). Several music streaming platforms, such as Spotify and Deezer, provide a free, ad-supported tier that is used to cultivate brand loyalty and in turn convert ad-supported users into more lucrative paying subscribers. Editorially curated and personalised playlists transform listeners into a resource whose attention can be sold to advertisers (Anderson, 2015; Drott, 2018; Eriksson, 2020; Wikström, 2020). Several ‘calculative operations’ are triggered within and around a playlist (Eriksson, 2020), allowing these firms to infer information about the social, spatial and emotional contexts of consumers at particular points in time, creating audience segments to be sold to third-party brands. For example, moods and moments playlists (e.g. motivational playlists, heartbreak playlists) are crucial to the provision of programmatic advertising, allowing advertisers to reach audiences in specific spatial and emotional contexts (Anderson, 2015; Drott, 2018; Eriksson, 2020; Wikström, 2020). Music streaming platforms seek to persuade consumers that lives are decomposable into a series of moments, each of which can be paired with music and, crucially, advertisements (Drott, 2018).

How do music streaming platforms personalise?

Music streaming platforms seek to engage listeners in both ‘unending consumption’ and a ‘data feedback loop’ (Prey, 2016). Sustained engagement is not only important to the business models of music streaming platforms, as it increases the likelihood of subscription renewal and creates more opportunities to sell advertisements, it is also crucial for data creation. Listeners are continually working on their future recommendations – knowingly or not – by streaming more music and providing the implicit and explicit signals – a data feedback loop – that enables algorithms to make predictions about future wants, needs and desires (Anderson, 2014; Eriksson et al., 2019; Morris, 2015b; Prey, 2016; Wikström, 2020)

Music streaming platforms create data about a variety of interactions, from what people search to the tracks, artists and albums they stream, skip and save to their libraries. This is combined with demographic information (e.g. age, location) that is collected when a user registers their account. With the expansion of smart devices (e.g. smart speakers and smart cars) and their integration with music streaming platforms, the types of data these platforms create has further expanded, allowing for the creation of voice and dynamic locational data (Hracs and Webster, 2020).

Digital data feed into the computational systems used to achieve personalisation. Data enable these systems to create quantitative representations of people’s music taste – ‘taste profiles’, as Prey (2016) describes them – and use these representations as a basis for identifying relevant music to recommend using techniques such as item-to-item collaborative filtering (Jannach et al., 2016). For example, for the creation of Spotify’s ‘Discovery Weekly’, collaborative filtering is used to analyse the co-occurrence of pairs of tracks in people’s playlists, and then it uses co-occurrence to select related but unheard items for a given user to compile into a playlist (Pasick, 2015).

Meanwhile, these data also inform the decisions made by the playlist editors involved in the creation and management of editorial playlists. According to Bonini and Gandini (2019), the creation of editorial playlists is underpinned by ‘algotorial’ forms of expertise. Digital data are used to help experts identify tracks that are performing well and relegate or remove tracks that are underperforming in a given playlist, as defined by metrics such as the number of streams and skip rates (Barna, 2017; Eriksson, 2020; Webster, 2019). While algotorial forms of expertise are not unique to the streaming age, as radio programmers have an established history of programming tracks by combining their own expertise with programming software and audience research (Anderson, 2014; Razlogova, 2013; Taylor, 2005), streaming playlist editors are uniquely positioned to evaluate the impact of their decisions at scale (i.e. with tens or hundreds of millions of listeners) and in near real-time, allowing for continuous optimisation.

The data feedback loop enables the human and technical actors involved in personalisation to continuously optimise decisions to ensure that the content being produced and distributed is having a measurable impact on user engagement (Hracs and Webster, 2020; Jannach et al., 2016; Seaver, 2018). Playlist editors evaluate what content is having a positive or negative impact on user engagement and adapt accordingly (Barna, 2017; Bonini and Gandini, 2019; Eriksson, 2020), while the recommender systems used to generate personalised recommendations and playlists are optimised using what Seaver (2018) describes as ‘captivation metrics’. Rather than focussing simply on predictive accuracy, the personalisation performed by music streaming platforms is optimised to capture user attention, measured through captivation metrics, such as ‘time spent listening’. Adopting this approach seeks to increase the likelihood that content drives user engagement, an outcome integral to the commercial success of music streaming platforms. The comments of a Data Scientist working at a leading music streaming platform illustrate captalogical thinking at work:

We are obviously optimising for user engagement, right, so we want to provide them (users) with content which we would actually guarantee us or would increase the probability of users staying on [name of platform] for so much longer. (Interview, Data Scientist, Music Streaming Platform)

Personalisation is an integrated part of the experience of using music streaming platforms. What music people encounter when they use music streaming platforms is no coincidence. Combining data about past listening behaviour with the latest advancements in recommendation technologies, music streaming platforms are making predictions about what a person might want to listen to next, serving up these predictions as personalised playlists and recommendations in automated and pervasive ways across the microspatialities of their platforms.

Listeners – or end users, as they are better understood in this new service-driven economy (Anderson, 2014; Arditi, 2018; Burkhart, 2008) – are at the heart of the personalisation performed by music streaming platforms. They are an essential part of the sociotechnical systems that collect and analyse data about music consumption practices and make predictions about future needs, wants and desires. End users’ interactions with digital platforms produce these data and users’ reactions to changes in the contents of an editorially curated playlist or the design of an algorithm are the basis against which the success of recommendations – measured through captivation metrics (Seaver, 2018) – is evaluated and optimised. Morris (2015b) and Anderson (2014) argue that end users should also be understood as labourers, whose leisure practices are commodified by creating value – in the form of personalised recommendations and advertisements – from the digital data traces these practices leave behind. Indeed as Prey (2016) makes clear with the concept of ‘data feedback loop’, digital platforms are explicitly designed for capturing user interactions and feeding them back into human and algorithmic decision-making processes dedicated to optimising user experiences (Anderson, 2014; Morris, 2015b; Wikström, 2020). As Eriksson et al. (2019: 137) summarise, the promise of personalisation ‘. . . encourages users to work on their future recommendations by streaming more music and inputting more data into the system’.

Yet while important progress has been made in theorising the role that people – as end users and labourers – play in the design and architecture of digital platforms and the personalisation they perform, less is empirically known about how people experience personalisation and the affordances it presents in everyday life. Focussing on the part that music taste plays in the performance of class identities, the following sections demonstrate that personalisation’s promise to help individuals efficiently and effortlessly find and engage with relevant music is a challenge to some people’s claims to class distinction. This is because it perceived to be undermining the need for musical expertise and compressing the time needed to appreciate music for its own sake, important parts of how some people seek to achieve distinction through the performance of cultural omnivorousness.

Displaying musical expertise in the platform age

Existing literature about the everyday use of music streaming platforms has closely examined playlist curation, a practice distinctive but not unique to the platform age (Fuentes et al., 2019; Hagen, 2015; Johansson et al., 2018; Sinclair and Tinson, 2017; Webster, 2019). Playlist curation is identified as a way to both exert control over the vast amounts of music made available by music streaming platforms and imbue psychological ownership over what would otherwise be a vast and impersonal space. This research has demonstrated how individuals have extended the collector fetish – to covet, collect, stockpile and enjoy cultural artefacts – associated with analogue modes of consumption (e.g. vinyl record collecting) to the virtual realm through playlist curation, using it as a way to display knowledge and expertise about music (Hagen, 2015; Webster, 2019).

Indeed, a new set of occupations has emerged around the practice of playlist curation (Barna, 2017; Bonini and Gandini, 2019; Webster, 2019). Music streaming platforms have assembled global teams of playlist editors, who create and manage large suites of branded music playlists. These teams are often composed of individuals who used to work at record labels, radio stations and in the music press, hired for their knowledge of trends in recorded music, contacts in the industry, and their ability to analyse and interpret data generated about user interactions with playlists (Bonini and Gandini, 2019). The emergence of these occupations has enabled privileged individuals to profit from playlist curation, converting their cultural and social capital into economic opportunities (Webster, 2019).

While previous research has demonstrated how music streaming platforms have created opportunities to display musical expertise and perform identities through playlist curation, this article highlights that the promise of personalisation is at the same time undermining these same opportunities.

The problematic promise of personalisation

Delivering on its promise to do the ‘work of discovering music’ (McCourt and Zuberi, 2016: 123), personalisation has seemingly made it more efficient to explore and extract value from the abundance of music made available by music streaming platforms. For example, Christian (mid-20s, Arts Co-ordinator, university-educated) discusses how interacting with personalisation has encouraged him to be more ‘inquisitive’ and ‘thirsty to discover music’. He describes how using Spotify’s ‘Discover Weekly’ has enabled him to learn more about independent artists who would otherwise have gone under the radar. As he explains, ‘It’s always easy to catch wind of what big bands are doing, but having “Discover Weekly,” I’m discovering artists that are actually quite small, so it’s really nice to be able discover new bands, up-and-coming music’.

Yet, the promise of personalisation is, at the same time, creating conflict for those whose claims to social distinction rest on the performance of cultural omnivorousness. The concept of cultural omnivorousness is used to describe a ‘knowing’ and expert engagement with cultural forms (Jarness, 2015; Prieur and Savage, 2013; Savage and Gayo, 2011). It is a way of pursuing social distinction by demonstrating to others one’s expertise and command over diverse cultural forms. For a subset of my participants, specifically Jamie (mid-20s, Digital Marketing, university-educated), Ben (early 30s, Policy Advisor, university-educated), Christian (mid-20s, Arts Co-ordinator, university-educated), Joel (mid-20s, Administrator, college-educated), Greg (late 20s, Customer Service Advisor, university-educated) and Elizabeth (mid-20s, Administrator, university-educated), deploying accumulated knowledge and discriminatory skills is important to how they seek to achieve distinction through music consumption. They share particular ideals about what and how to listen to music, distancing themselves from the most popular of popular music, and privileging attentive listening and knowledge about how to ‘match’ music with particular moments. For example, Elizabeth is in her mid 20s and grew up in a middle-class family where value was placed in the arts and humanities. Her father is a bluegrass musician, her mother is a classically trained singer and her older brother is an academic working in the field of musicology. Elizabeth was introduced to the cello at a young age and she went on to study Drama at university in London. Elizabeth enjoys ‘hunting around’ for new music and giving recommendations to her friends. Elizabeth seeks to define her tastes in opposition to the mainstream, while confidently asserting what she likes, as her comments illustrate:

I guess this is quite judgmental, but I think the layperson’s taste in music is quite bad (laughs). So, if I was to listen to Radio 1 or the top 40, like at one point Ed Sheeran’s entire album was the top 40, I don’t like Ed Sheeran, but everyone else seems to. I don’t really pay much heed to what everyone else likes. I’m more interested in what I like.

As I will demonstrate in this section and the next, these claims to distinction and ideals about how to consume music are challenged by the possibility to outsource the labour of music discovery afforded by personalisation. As for the rest of my participants, Tracy, Catherine, Marie, Rebecca, Sarah, Deborah, Steve, Josephine, Sian, Andrew, Johanne, Phil, Claire and Michelle, I will demonstrate that personalisation is not perceived as a threat in the same way because their cultural capital and claims to distinction are not at stake.

Undermining opportunities to mobilise musical expertise

Jamie (mid 20s, Digital Marketing, university-educated) is a cultural omnivore, displaying assured handling of a range of musical styles, citing specific stylistic features and sub-genres when describing his tastes. Akin with what other research has observed (Hagen, 2015; Webster, 2019), creating playlists on Spotify is a way for Jamie to display to others his musical expertise and familiarity with diverse musical styles. As he puts it, ‘I take a lot of pride in finding the right situation for a playlist and then putting it on and seeing how people react to it’.

However, when asked about Spotify’s personalised playlists and recommendations, Jamie describes the act of discovery as being ‘consigned to the scrapheap’. Personalisation is a challenge to Jamie’s claims to social distinction because it is perceived as taking over the labour of music curation, which for Kjus’ (2016) research participants has generated distrust in the favour of one’s own musical discretion. In contrast to existing literature highlighting that the practice of playlist curation has created opportunities for individual consumers to reassert control over abundance and imbue a sense of psychological ownership (Hagen, 2015; Johansson et al., 2018; Sinclair and Tinson, 2017), Morris (2015b) argues that the delegation of discovery and curation to digital services is eroding these same notions of ownership and control. Indeed, Burkhart (2008) suggests that the loss of user control over choices about playing music has potential to challenge the ‘music collector’s ego’, reflected in the experiences of Jamie. Rather than relying on Spotify’s recommendation systems to make choices for him, Jamie actively tries to ensure that he is continuing to expand his knowledge beyond what the algorithms find for him and engage with music omnivorously. Jamie’s response echoes the critique made by Hanrahan (2018: 300), who suggests that ‘. . . personalisation is a trap rather than a fulfilment, a foreclosing of rather than opening to the possibilities of aesthetic experience’. Or as Jamie puts it,

I do feel a certain amount of pressure to keep my listening habits wide enough that I am not getting six months down the line and realising that, oh, all I’m ever listening to is this now because this is all I ever discover.

Personalisation is a double-edged sword. While existing literature has highlighted how music streaming platforms have created opportunities to display expertise through playlist curation (Hagen, 2015; Webster, 2019), a practice enabled and encouraged by these platforms, these findings highlight that these same opportunities are in contention with the personalisation they perform. Through the use of techniques such as collaborative filtering, personalisation increases the scale at which new music can be discovered, but it is at the same time undermining opportunities to perform the cherished labour of music curation for some. Frith (1996) highlights that a key point of contention in debates about popular cultural judgements is ‘who’ has the authority to make them. Music streaming platforms have invested in personalisation to help people find the right music for the moment, which is generating anxieties for those, such as Jamie, who feel their expertise is being replaced by these computational processes.

In the following section, I discuss how personalisation is further challenging the part that cultural omnivorousness plays in the pursuit of social distinction by compressing the time available to appreciate music for its own sake.

Taking time to appreciate music in the platform age

Access to time is one the foundations of privilege in society. It is an unevenly distributed resource that allows privileged individuals and groups to display to others that they can afford to not work and do not have to attend to the immediate needs of the body (Veblen, 1912). Indeed, time is crucial to how social distinction is achieved through taste and consumption (Bourdieu, 1984). The expert and ‘knowing’ mode of consumption associated with cultural omnivorousness emphasises the importance of taking time to appreciate cultural goods for their own sake, and it is a disposition that takes time and sustained engagement to acquire (Bennett et al., 2009; Jarness, 2015; Savage et al., 2013; Savage and Gayo, 2011).

In recent years, sociologists have returned to the question of how time is experienced and how access to it is unevenly distributed (Wajcman, 2016). Time-saving technologies, such as smartphones and smartwatches, have introduced a ‘time paradox’, where technologies designed to make people’s lives simpler and easier have had the paradoxical effect of making people feel busier, as time becomes more visible and time spent more measurable. This section demonstrates how music streaming platforms are contributing to this paradox by compressing the time available to appreciate music for its own sake. This is because the continuous selection and presentation of new music is making the experience of consuming music ephemeral by design.

The accelerated availability of recorded music

The imperatives to attract, engage and retain users that underpin personalisation are accelerating the rate at which recorded music is selected, arranged and presented to individuals. The contents of editorially curated and personalised playlists are updated on a regular basis to ensure content stays ‘fresh’ and ‘up-to-date’ (Barna, 2017; Bonini and Gandini, 2019). For example, Spotify’s ‘New Music Friday’ playlist is updated every Friday and on Monday the tracks are re-arranged, and Spotify’s homepage is dynamically updated across the course of a day, changing according to recent listening, time and other contextual factors, such as the weather. As such, media commentators have described the playlists created by music streaming platforms as ‘. . . factory outputs that push the industrial production of commercial musical assemblages to new levels’ (Ugwu, 2016).

These continuous changes to what, when and how music is made available to individual users are enabled by the ‘data feedback loop’ (Prey, 2016) created by music streaming platforms (Barna, 2017; Bonini and Gandini, 2019; Eriksson, 2020; Seaver, 2018). The editors responsible for creating and managing playlists quantitatively evaluate the performance of tracks, making decisions about what to include, exclude, demote or promote using metrics, such as how often a track is skipped or added to a library (Barna, 2017; Bonini and Gandini, 2019; Eriksson, 2020). The ‘taste profiles’ (Prey, 2016) used to generate personalised recommendations and playlists are updated on a regular basis to ensure changes in user preferences are reflected in future recommendations, as one informant (Interview, Computer Scientist 1) puts it: ‘. . . you collect the data continuously and you continuously learn’. And the labour of the human and technical actors involved in personalisation are continually evaluated and tweaked against ‘captivation metrics’ to arrive at the output that has the most desirable impact on user engagement (Seaver, 2018).

Compressing the time to appreciate music for its own sake

The accelerated availability of recorded music is undermining the claims to distinction of my group of omnivorous music consumers (Jamie, Ben, Christian, Joel, Greg, Elizabeth) by compressing the time needed to appreciate music for its own sake. The seemingly endless supply of music being selected, arranged and presented by music streaming platforms is generating anxieties for these individuals because they feel unable to give music the attention they believe it ‘deserves’. For example, Ben’s (early 30s, Policy Advisor, university-educated) taste in music is omnivorous, characterised by a confident handling of a diverse range of musical styles, ranging from trip-hop to classical music, cultivated through his musical education. In our interview, Ben talked in a reflexive way about how he feels that he is not paying ‘enough’ attention to the music made available to him by Spotify. Because Spotify is continuously presenting him with new and relevant music, Ben feels increasingly unaware of the artists behind the music. ‘Old’ music is routinely replaced by something ‘new’, creating an inability to be satisfied in what Hanrahan (2018) describes as a ‘condition of limitlessness’. This is a source of resentment for Ben and he finds himself over-compensating to ensure he is engaging with music in the ‘right’ way. As he explains,

I might pay attention to a track that I like, if it’s good, I might start wondering who it is, but often it’ll pass you by and without looking at the screen you won’t know who it is. And just because you’ve looked at the screen once, doesn’t mean you’ll remember their name. This is why I’m trying to get better at ‘liking’ tracks (adding songs to his personal library) when I hear them.

Ben’s story highlights the emergence of a time paradox (Wajcman, 2016). Personalisation – a process designed to make it easier to navigate abundance and discover new music – is in fact making the experience of consuming music feel more overwhelming for some. While music streaming firms may claim that personalisation helps individuals to manage the immense amount of music made available, it is making it harder to engage with the music being recommended because it is affording people less time and space to do so. This is undermining opportunities to take the time to appreciate music for its own sake, an important part of how taste and consumption are historically used to pursue social distinction (Bennett et al., 2009; Bourdieu, 1984; Savage et al., 2013; Veblen, 1912).

Who has cultural capital at stake?

This article has highlighted some ways in which music streaming platforms are shaping the everyday performance of class identities and the pursuit of social distinction. The promise of personalisation is perceived to be undermining opportunities to display musical expertise and take time to appreciate music for its own sake. However, music streaming platforms are not experienced in the same way by everyone. Rather, this section demonstrates that perceptions of personalisation and platformisation are shaped by who has cultural capital at stake, which is shaped by generational and gendered differences in music consumption practices.

Seeing personalisation as a ‘friend’ or ‘foe’

Musical expertise or ‘connoisseurship’ is traditionally associated with masculine identities and is encapsulated in the figure of the male record collector (Shuker, 2004; Straw, 1997). Musical recordings are described by Straw (1997) as the material goods around which homosocial interactions take place. They are both public displays of knowledge and power and ways of seeking refuge from the sexual and social world in spaces such as the record store and the home. The knowledge and social connections that can be gained from musical connoisseurship has the potential to reproduce privilege in a way that excludes female participation. Indeed, Frith and McRobbie (2007) argue that rock culture in general is marked by a gendered dichotomy, where males understand themselves as collective and active audiences who relate to musical performances by identifying with it, while females understand themselves as individual and passive audiences.

These gendered differences in music consumption practices shaped how my participants perceive personalisation. For example, playlist curation is conceived of differently by several of my female participants, specifically Catherine, Maria, Rebecca, Michelle and Deborah. For example, music is valued by Catherine (early 30s, Public Engagement Manager, university-educated) based on its ability to facilitate particular moods, such as happiness, nostalgia and relaxation, rather than as a display of expertise, as it is for Jamie. How Catherine uses music to ‘soundtrack’ (Fuentes et al., 2019) her everyday life extends to how she incorporates personalisation into her music consumption practices. Both editorially curated and personalised playlists and recommendations help her to identify music that fits with the mood or moment she is in. As she explains,

Again, it depends on the mood, or what I’m, the activity I’m doing. So, if it’s more work-based, maybe I’m looking for something that’s quite chilled, I don’t know, ‘Focussed’, I know they’ve got those playlists as well, you know, like ‘Focus’. ‘Monday Motivation’, as well. If I want to go to the gym, it’ll be stuff that will have be quite focussed, and quite motivated, in quite an upbeat way.

Marie (early 30s, International Development Manager, university-educated) presents a similar story. She feels limited in her ability to dedicate time to discovering and curating music, due to the demands of childcare and her full-time job, reflecting Cunningham et al.’s (2007) findings that discovery is closely tied to the availability of leisure time. Instead, Marie chooses to outsource choice to Spotify, which is able to find music that is familiar and relevant to her interests, without requiring her to spend precious time curating it herself. As Marie does not place such importance in the possession of musical expertise, personalisation is seen as a ‘friend’ who makes her life easier. As she explains,

So that’s what I like about it (Spotify), that it almost makes that choice (picking what song to listen to) for me, so it’s one less thing I have to worry about. I have so many decisions to make in my day, with my children, and my job, juggling twenty different things, that actually it’s quite nice that something else makes that decision on my behalf.

Second, generational differences in music consumption practices also shape how music streaming platform’s affordances are seen and understood. Across my sample of participants, age shapes attitudes towards the divisiveness of music taste and the value placed in accumulating and deploying musical knowledge and discriminatory skill. Coupled with the affordances of abundant access to music provided by Spotify, a laissez-faire attitude has allowed my older participants, particularly Tracy (late 40s), Rebecca (early 40s), Phil (early 30s), Deborah (early 40s) and Steve (early 40s), to readily traverse musical boundaries that they would previously have avoided for fear of being seen as ‘uncool’ or ‘out of place’. The affordances of music streaming platforms are not perceived as a threat because their claims to distinction are not at stake, creating the conditions for them to be more liberal about what they consume and how. For example, Steve’s comments highlights that under different circumstances and without the affordances of Spotify, namely, the private nature of music streaming, he wouldn’t be able to listen to Katy Perry, an artist he recognises as not being appropriate for a 40-year-old man:

Nobody’s judging you for your music tastes here (on Spotify); you don’t have to go and justify yourself to the guy at the counter why you’re buying a Katy Perry album at 40, you know, that’s quite a serious thing, it does help you, I think, because you can just listen to anything on here

Both gendered and generational differences in music consumption practices shape how music streaming platforms are implicated in the pursuit of social distinction. In particular, gender and generational differences challenge the legitimacy of musical expertise as a source of distinction, which in turn shapes how music streaming platforms and the personalisation they perform are seen and understood. For some of my male participants, such as Jamie and Ben, whose identities draw on masculine notions of musical connoisseurship, personalisation is seen as a challenge to their claims to social distinction. While for Catherine, Marie and Steve, the stakes are different. Musical expertise does not hold the same social currency for them and, as such, the affordances of music streaming platforms are understood differently. They value the convenience personalisation brings and embrace the opportunities to traverse musical boundaries afforded to them by Spotify.

Conclusion

Music streaming platforms have transformed the circulation and consumption of recorded music. In order to attract, engage and retain valuable users, they have invested in personalisation to make it easier to find and engage with relevant music. As a consequence, what and how people encounter and experience recorded music is being shaped by the decisions made by a new and powerful set of human and computational actors. While existing literature has identified ways in which platformisation and personalisation might be shaping the part that music taste and consumption play in the cultural reproduction of class, important gaps persist. In particular, little empirical consideration has been given to how individuals incorporate or resist the affordances of personalisation and with what consequence for the everyday performance of class identities and the pursuit of social distinction.

To address these gaps and nuance our understanding of the social and cultural consequences of platformisation and personalisation, this article examined how music streaming platforms are shaping the social dynamics of music consumption. It demonstrated that the incentive to outsource the labour of music discovery and ‘soundtracking’ offered by personalisation is undermining opportunities to display musical expertise, while the rate at which new music is recommended is making it harder for people to take the time to appreciate music for its own sake – two ways in which cultural omnivorousness is performed in the pursuit of social distinction. However, personalisation is not experienced as divisively by everyone. Rather, this article demonstrated that gendered, generational and class differences in music consumption practices shape how the affordances of personalisation are experienced and understood. It demonstrated that for those who have cultural capital at stake, personalisation is seen as a challenge to their claims to social distinction because their musical expertise and discriminatory skill is perceived to be under threat. This highlights how claims to social distinction and forms of cultural capital are contested by the changing material conditions in which recorded music is accessed, curated and consumed in the platform age.

These findings make several important contributions. This article contributes to debates about platformisation (Duffy et al., 2019; Langley and Leyshon, 2017; Meier and Manzerolle, 2019; Nieborg and Poell, 2018; Poell et al., 2019; Prey, 2020; Srnicek, 2017; van Dijck, 2013; van Dijck et al., 2019), developing our understanding of what digital platforms are and why and how they intervene in the circulation and consumption of cultural goods. It contributes to our understanding of datafication (Drott, 2018; Eriksson, 2020; Prey, 2016; Seaver, 2018), illustrating how digital data are used to generate and evaluate editorially curated and personalised playlists and recommendations. It contributes much-needed empirical insight to debates about how music streaming platforms are disrupting the relationship between taste, class and consumption (Barna, 2020; Beer, 2013; Prey, 2016; Prior, 2018; Wright, 2015), focussing on everyday performance of class identities and the pursuit of social distinction (Bennett et al., 2009; Friedman et al., 2015; Jarness, 2015; Savage, 2000; Savage et al., 2013). In so doing, it also contributes to a growing body of literature about the everyday use of music streaming platforms (Anderson, 2015; Fuentes et al., 2019; Hagen, 2015; Hanrahan, 2018; Johansson et al., 2018; Kjus, 2016; Sinclair and Tinson, 2017), which to date has paid limited attention to both personalisation and class dynamics. Finally, by exploring how gendered, generational and class differences in music consumption practices shape experiences of personalisation, this article contributes to our understanding of how multiple social structures intersect with technological practice (Ziellen and Hargittai, 2009).

Looking ahead, I would like to highlight three important areas for further research.

First, while this article contributes to our understanding of the mobilisation of cultural capital, additional research is needed to consider how the accumulation and deployment of economic and social capital – in line with Bourdieu’s conceptual and theoretical framework – intersects with the use of music streaming platforms. For instance, the layers of technologies required to use music streaming platforms, such as broadband, 3G + mobile Internet, smartphones and headphones, may shape who and how people use music streaming platforms, while research has highlighted that the collection of analogue recordings, such as vinyl LPs and cassettes, is a way to deploy economic capital in resistance to the popularisation of music streaming platforms (Hayes, 2006; Magaudda, 2011; Webster, 2019). Meanwhile, complementary forms of musical participation, such as concerts and music festivals, may create opportunities for people to convert economic capital in arenas that may be more socially profitable than the consumption of recorded music.

Second, future research should consider if and how the curation performed by music streaming platforms is shaping the processes of hierarchisation that underpin what and how particular cultural forms (e.g. genres of music) become implicated in the pursuit of social distinction. Not only do music streaming platforms draw on the human expertise of playlist editors to provide curation, this article demonstrated how they also leverage computational techniques to make judgements about what music is right for individuals. This raises important questions about who (or what) has cultural authority over what counts as ‘good’ and ‘bad’ music taste in the 21st century, and the changing logics around which cultural judgements are made (Anderson, 2015; Morris, 2015a; Prey, 2020; Webster et al., 2016; Wikström, 2020). While genre classifications have traditionally been an important – and contentious – organising scheme for the production, marketing, retailing and consumption of recorded music (Frith, 1996), this article has demonstrated how personalisation selects music according to individual preferences and frames it in individualised and genre-agnostic terms (e.g. Spotify’s ‘Your Discover Weekly’). There is a need for future research to consider what implications this has for the formation of the hierarchies around which class differences in music taste are organised and sustained.

Third, this study explored the consumption practices of a group of individuals in the United Kingdom who use Spotify as their primary music streaming service. It is important to recognise that the United Kingdom, like any other nation, has a distinctive social class system and local music cultures, shaping how class manifests through consumption practices as well as the contents of cultural hierarchies. Therefore, the ways in which music streaming platforms are shaping the performance of class identities and distinction is likely to be different in other parts of the world, requiring dedicated and geographically sensitive research to complement the British perspective presented here. Indeed, the popularity of particular music streaming platforms is geographically differentiated. For example, Spotify is absent in China, where Tencent Music Entertainment Group, which owns QQ Music and Kuwo Music, is the market leader, with different pricing strategies and features to those offered by Spotify and its competitors in Europe and North America, further highlighting the need to consider geographical differences in platformisation and music consumption.

The rise of music streaming platforms has brought with it radical changes to the way recorded music is circulated and consumed. This article has demonstrated that how music taste and consumption is used by people in the performance of class identities and distinction is being challenged by the affordances of personalisation, highlighting a need for ongoing research.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Engineering and Physical Sciences Research Council, UK: [Grant Number EP/LO16117/1].

ORCID iD

Jack Webster

Author biography

Jack Webster is a graduate of the Web Science Centre for Doctoral Training (CDT) at the University of Southampton, UK. He is interested in how platformisation is shaping the production and consumption of cultural goods.

References

Anderson

(2015) Neo-muzak and the business of mood. Critical Inquiry 41(4): 811–840.

Anderson

(2014) Popular Music in a Digital Music Economy: Problems and Practices for an Emerging Service Industry. New York: Routledge.

Arditi

(2018) Digital subscriptions: the unending consumption of music in the digital era. Popular Music and Society 41(3): 302–318.

Barna

(2017) The perfect guide in a crowded musical landscape: online music platforms and curatorship. First Monday 22(4). Available at: www.firstmonday.org/ojs/index.php/fm/article/view/6914/6086 (accessed 21 July 2020)

Barna

(2020) The relentless rise of the poptimist omnivore: taste, symbolic power, and the digitization of the music industries. In: Tofalvy

Barna

(eds) Popular Music, Technology, and the Changing Media Ecosystem. Cham: Palgrave Macmillan, pp. 79–95.

Beer

(2013) Popular Culture and New Media. Basingstoke: Palgrave Macmillan.

Bellavance

(2008) Where’s high? Who’s low? What’s new? Classification and stratification inside cultural ‘repertoires.’ Poetics 36(2–3): 189–216.

Bennett

Savage

Silva

, et al. (2009) Culture, Class, Distinction. Abingdon: Routledge.

Bonini

Gandini

(2019) ‘First week is editorial, second week is algorithmic’: platform gatekeepers and the platformization of music curation. Social Media + Society 5(4): 1–11.

10.

Bourdieu

(1984) Distinction: A Social Critique of the Judgement of Taste. Oxford: Routledge.

11.

Bourdieu

(1986) The forms of capital. In: Richardson

(ed.) Handbook of Theory and Research for the Sociology of Education. New York: Greenwood, pp. 241–258.

12.

Burkhart

(2008) Trends in digital music archiving. The Information Society 24(4): 245–250.

13.

Cheney-Lippold

(2011) A new algorithmic identity: soft biopolitics and the modulation of control’. Theory, Culture & Society 28(6): 164–181.

14.

Cunningham

Bainbridge

McKay

(2007) Finding new music: a diary study of everyday encounters with novel songs. In: Proceedings of the 8th international conference on music information retrieval, Vienna, Austria, 23–27 September, pp. 83–88. Vienna, Austria: Austrian Computing Society.

15.

Drott

(2018) Why the next song matters: streaming, recommendation, scarcity. Twentieth-Century Music 15(3): 325–357.

16.

Duffy

Poell

Nieborg

(2019) Platform practices in the cultural industries: creativity, labor, and citizenship. Social Media + Society 5(4): 1–8.

17.

Eriksson

(2020) The editorial playlist as container technology: on Spotify and the logistical role of digital music packages. Journal of Cultural Economy 13(4): 415–427.

18.

Eriksson

Fleischer

Johansson

, et al. (2019) Spotify Teardown: Inside the Black Box of Streaming Music. Cambridge, MA: MIT Press.

19.

Friedman

Savage

Hanquinet

, et al. (2015) Cultural sociology and new forms of distinction. Poetics 53: 1–8.

20.

Frith

(1996) Performing Rites: On the Value of Popular Music. Cambridge, MA: Harvard University Press.

21.

Frith

McRobbie

(2007) Rock and sexuality. In: Frith

(ed.) Taking Popular Music Seriously. Aldershot: Ashgate, pp. 41–58.

22.

Fuentes

Hagberg

Kjellberg

(2019) Soundtracking: music listening practices in the digital age. European Journal of Marketing 53(3): 483–503.

23.

Fuez

Fuller

Stalder

(2011) Personal web searching in the age of semantic capitalism: diagnosing the mechanisms of personalisation. First Monday 16(2). Available at: https://firstmonday.org/ojs/index.php/fm/article/download/3344/2766

24.

Hagen

(2015) The playlist experience: personal playlists in music streaming services. Popular Music and Society 38(5): 625–645.

25.

Hanrahan

(2018) Hearing the contradictions: aesthetic experience, music and digitization. Cultural Sociology 12(3): 289–302.

26.

Hayes

(2006) ‘Take those old records off the shelf’: youth and music consumption in the postmodern age. Popular Music and Society 29(1): 51–68.

27.

Hracs

Webster

(2019) From selling songs to engineering experiences: the competitive strategies of music streaming platforms. Journal of Cultural Economy. Epub ahead of print 26 September. DOI: 10.1080/17530350.2020.1819374.

28.

Jannach

Resnick

Tuzhilin

, et al. (2016) Recommender systems – beyond matrix completion. Communications of the ACM 59(11): 94–102.

29.

Jarness

(2015) Modes of consumption: from ‘what’ to ‘how’ in cultural stratification research. Poetics 53: 65–79.

30.

Johansson

Werner

Åker

, et al. (2018) Streaming Music: Practices, Media, Cultures. Abingdon: Routledge.

31.

Kassabian

(2013) Ubiquitous Listening: Affect, Attention, and Distributed Subjectivity. Berkeley, CA: University of California Press.

32.

Kjus

(2016) Musical exploration via streaming services: the Norwegian experience. Popular Communication 14(3): 127–136.

33.

Langley

Leyshon

(2017) Platform capitalism: the intermediation and capitalisation of digital economic circulation. Finance and Society 3(1): 11–31.

34.

Light

Burgess

Duguay

(2018) The walkthrough method: an approach to the study of apps. New Media & Society 20(3): 881–900.

35.

McCourt

Zuberi

(2016) Music and discovery. Popular Communication 14(3): 123–126.

36.

Magaudda

(2011) When materiality ‘bites back’: digital music consumption practices in the age of dematerialization. Journal of Consumer Culture 11(1): 15–36.

37.

Meier

Manzerolle

(2019) Rising tides? Data capture, platform accumulation, and new monopolies in the digital music economy. New Media & Society 21(3): 543–561.

38.

Morris

(2015a) Curation by code: infomediaries and the data mining of taste. European Journal of Cultural Studies 18(45): 446–463.

39.

Morris

(2015b) Selling Digital Music, Formatting Culture. Oakland, CA: California University Press.

40.

Morris

Powers

(2015) Control, curation and musical experience in streaming music services. Creative Industries Journal 8(2): 106–122.

41.

Nieborg

Poell

(2018) The Platformization of cultural production: theorizing the contingent cultural commodity. New Media & Society 20(11): 4275–4292.

42.

Pariser

(2012) The Filter Bubble: What the Internet Is Hiding from You. London: Penguin.

43.

Pasick

(2015) The magic that makes Spotify’s Discover Weekly playlists so damn good. Quartz, 21 December. Available at: http://qz.com/571007/the-magic-that-makes-spotifys-discover-weekly-playlists-so-damn-good/ (accessed 21 July 2020)

44.

Peterson

Kern

(1996) Changing highbrow taste: from snob to omnivore. American Sociological Review 61(5): 900–907.

45.

Poell

Nieborg

van Dijck

(2019) Platformisation. Internet Policy Review 8(4): 1–13.

46.

Prey

(2016) Musica analytica: the datafication of listening In: Nowak

Whelan

(eds) Networked Music Cultures. London: Palgrave Macmillan, pp. 31–48.

47.

Prey

(2020) Locating power in platformization: music streaming playlists and curatorial power. Social Media + Society 6(2): 1–11.

48.

Prieur

Savage

(2013) Emerging forms of cultural capital. European Societies 15(2): 246–267.

49.

Prior

(2018) Popular Music, Digital Technology and Society. London: SAGE.

50.

Razlogova

(2013) The past and future of music listening: between freeform DJs and recommendation algorithms. In: Loviglio

Hilmes

(eds) Radio’s New Wave: Global Sound in the Digital Era. New York: Routledge, pp. 62–76.

51.

Rimmer

(2012) Beyond omnivores and univores: the promise of a concept of musical habitus. Cultural Sociology 6(3): 299–318.

52.

Savage

(2000) Class Analysis and Social Transformation. Buckingham: Open University Press.

53.

Savage

(2006) The musical field. Cultural Trends 15(2–3): 159–174.

54.

Savage

Gayo

(2011) Unravelling the omnivore: a field analysis of contemporary musical taste in the United Kingdom. Poetics 39(5): 337–357.

55.

Savage

Devine

Cunningham

, et al. (2013) A new model of social class? Findings from the BBC’s Great British Class Survey experiment. Sociology 47(2): 219–250.

56.

Seaver

(2018) Captivating algorithms: recommender systems as traps. Journal of Material Culture 24(4): 421–436.

57.

Shuker

(2004) Beyond the ‘High Fidelity’ stereotype: defining the (contemporary) record collector. Popular Music 23(3): 311–330.

58.

Sinclair

Tinson

(2017) Psychological ownership and music streaming consumption. Journal of Business Research 71: 1–9.

59.

Sinnreich

(2015) Music cartels and the dematerialization of power. In: Bennett

Waksman

(eds) The SAGE Handbook of Popular Music. London: SAGE, pp. 613–628.

60.

Srnicek

(2017) Platform Capitalism. Cambridge: Polity Press.

61.

Straw

(1997) Sizing up record collections: gender and connoisseurship in rock music culture. In: Whiteley

(ed) Sexing the Groove: Popular Music and Gender. Abingdon: Routledge, pp. 3–16.

62.

Taylor

(2005) ‘I am what I play’: the radio DJ as cultural arbiter and negotiator. In: Beck

(ed.) Cultural Work: Understanding the Cultural Industries. London: Routledge, pp. 73–100.

63.

Ugwu

(2016) Inside the playlist factory. Buzzfeed, 12 July. Available at: https://www.buzzfeed.com/reggieugwu/the-unsung-heroes-of-the-music-streaming-boom?utm_term=.flR3Xyplq#.emjev5ZMz (accessed 21 July 2020).

64.

van Dijck

(2013) The Culture of Connectivity: A Critical History of Social Media. Oxford: Oxford University Press.

65.

van Dijck

Nieborg

Poell

(2019) Reframing platform power. Internet Policy Review 8(2): 1–18.

66.

Veblen

(1912) The Theory of the Leisure Class. New York: A.M. Kelley.

67.

Wajcman

(2016) Pressed for Time: The Acceleration of Life in Digital Capitalism. Chicago, IL: University of Chicago Press.

68.

Warde

Wright

Gayo-Cal

(2008) The omnivorous orientation in the UK. Poetics 36(2): 148–165.

69.

Webster

(2019) Taste in the platform age: music streaming services and new forms of class distinction. Information, Communication & Society. Epub ahead of print 28 May. DOI: 10.1080/1369118X.2019.1622763.

70.

Webster

Gibbins

Halford

, et al. (2016) Towards a theoretical approach for analysing music recommender systems as sociotechnical cultural intermediaries. In: Proceedings of the 8th ACM conference on web science, Hannover, DE, 23–25 May, pp. 137–145. New York: ACM.

71.

Wikström

(2020) The Music Industry: Music in the Cloud. 3rd edn. Cambridge: Polity Press.

72.

Wright

(2011) Making tastes for everything: omnivorousness and cultural abundance. Journal of Cultural Research 15(4): 355–371.

73.

Wright

(2015) Understanding Cultural Taste: Sensation, Skill and Sensibility. Basingstoke: Palgrave Macmillan.

74.

Ziellen

Hargittai

(2009) Digital distinction: status-specific types of Internet usage. Social Science Quarterly 90(2): 274–291.