Thinking Culture in the Age of Global Finance: A review of Spotify Teardown

The authors of Spotify Teardown have written an important book. The first of what may soon become a cottage industry of texts about individual streaming services, Spotify Teardown shows what a multidisciplinary approach toward understanding Spotify can tell us about what drives the development of streaming services, and how they are changing media culture. Spotify Teardown is a collaborative effort from five Sweden-based scholars to understand ‘how people’s practices and approaches toward cultural forms such as songs, books, or films—practices including the production, expression, and exchange of those cultural forms—are transformed under the shift from commodity ownership to commodified experience.’¹ Building on Jeremy Wade Morris’s account of the passage from storage culture to streaming culture in Selling Digital Music, Formatting Culture, the authors show the ways that Spotify’s development registers transformations within streaming culture. Bringing together traditional anthropological methods, economic sociological analysis, and recently developed digital methods, Spotify Teardown historicizes the relation between Spotify’s front end and its back end, subjecting the former to cultural critique and the latter to digital experimentation. The authors’ ‘teardown’ of Spotify is therefore not a literal process of reverse engineering so much as a metaphor for ‘disassembl(ing) the way Spotify’s product is commonly conceptualized.’² Spotify Teardown poses a timely challenge to both prevailing conceptualizations of platforms and the digital methods developed for studying them. The authors’ approach is one that foregrounds processes of financialization, which enables them to identify fundamental features of Spotify’s history and organizational structure that are absent or are misrepresented in the dominant narratives that circulate around Spotify. Yet while the problematic of financialization facilitates important breakthroughs, financialization’s material effects simultaneously prevent the authors from incorporating musical content into their analysis. In other words, the absence of musicological inquiry in Spotify Teardown is not simply a shortcoming of the book, but a register of financialization as a larger historical situation. Paradoxically, if curation—an essential concept for the authors and matter of practice for Spotify—is the central aesthetic category this situation generates, it is also an index of a limit on our ability to think content into form. So we can read Spotify Teardown both for its powerful analysis of streaming culture, and also as an indicator of a threshold about how we are thinking culture in an age of global finance.

Historicizing Spotify, Conceptualizing Spotify

The dominant narrative about Spotify, circulated by its cofounder and CEO Daniel Ek and trade industry press alike, is that the Swedish company formed in 2006 to save the music industry. Resolving what once appeared to be an economic impossibility, Spotify promised to bring the music industry out of its early 21st century slump—too often narrowly attributed to mass file-sharing³ —while ‘giving users (legal) access to all the music in the world, for free.’⁴
Spotify Teardown tells a different story about the streaming service, one that has its origins in the practices and P2P (peer-to-peer) network infrastructure associated with piracy. The authors trace Spotify’s development from its beginnings as ‘de facto pirate service,’⁵ through its launch as a legal distributor of music, to its arrival as a ‘producer of unique music-related experiences.’⁶ Along the way, Spotify would walk back its supposed commitment to ‘free’ music and use its free ad-supported service to direct users toward becoming paid subscribers. Remaining attentive to the economic and online contexts that provide the backdrop for Spotify’s development, Spotify Teardown puts venture capital, rights holders, network operators, and policymakers at the center of how and why Spotify altered its strategies and goals over time.⁷ By ‘following the hype’—that is, by drawing upon how trade journals, news outlets, and data retrieved from old job listings, blog posts, and tweets have portrayed the company’s changing objectives—Eriksson et. al show how ‘narratives developed around Spotify resonate with its financial history.’⁸
In contrast to how we typically think about economic growth, with profits driving investment and the accumulation of capital, Spotify Teardown explains that venture capital’s primary interest in Spotify was not to make the company profitable, but to make it valuable: ‘Venture capital bets on return on investments at the time of an “exit,” that is, when Spotify is either acquired by a larger corporation or introduced at the stock exchange. Meanwhile, more and more venture capital is needed to cover the recurring losses and keep up the growth.’⁹ Thus, using ‘financialization as a structuring principle of media history,’¹⁰ Eriksson et. al organize Spotify’s development around seven rounds of funding primarily from global venture capital firms, which by the time of its initial public offering (IPO) on April 3, 2018 would total over $3 billion (with $1.6 billion in stocks and another $1.5 billion in the form of convertible debt).¹¹
The framework of financialization enables the authors of Spotify Teardown to conceptualize the ways that ‘coordination within Spotify’s production chain followed and appropriated models from the financial world.’¹² For example, they understand Spotify as taking advantage of arbitrage opportunities from undervalued songs, first illegally as free riders and then legally with its proportional revenue payout system.¹³ (‘Estimates… usually state that revenue per played track runs as low as $0.005 at Spotify.’¹⁴ They also portray Spotify as a broker, ‘making markets’ between the music industry and unauthorized file sharers.¹⁵ Yet the strength and the stakes of their problematic becomes clearest in their analysis of what is called ‘programmatic advertising.’ Programmatic, the industry shorthand, ‘is a mechanism for using personal data and algorithms to buy and sell ads’ that ‘includes the automation of online ad buys via interconnected online “trading desks” that allow the auctioning off of inventory within milliseconds.’¹⁶ While Spotify still sells most of its advertising inventory with human sales teams, programmatic exemplifies how financialization has come to shape the very structure of Spotify’s operations.
Given the ways that Spotify is embedded in different markets, the authors argue that the two-way and multisided market models that are popular within platform studies cannot adequately represent Spotify’s complex operations. For example, while the two-sided market model assumes that the content provided to users is ‘free,’ Spotify uses a pro rata revenue share approach, where ‘revenues are divided to the rights holders based on how many approved plays a certain track has in relation to all the other tracks played at the same time.’¹⁷ So too, while the model assumes that platforms’ ability to scale is based on advertising revenue, Spotify has never generated enough revenue from advertising to cover the costs of its free service.¹⁸ Because its scalability depended so heavily upon financial markets, Spotify requires a different conceptualization of its organizational structure and business model. Eriksson et. al write, ‘If we describe Spotify as a production chain, then songs, videos, audiences, and ads appear as auxiliary markets, despite the company’s promotional claims to the contrary when facing consumers.’¹⁹ For this reason, they conceptualize Spotify as a ‘stack—where trading sites are stacked into or on top of one another, in often opaque, unaccountable, and unsustainable ways.’²⁰ According to this model, losses in one market may generate revenues in others.
Spotify Teardown will prove both useful and provocative for a diverse readership. Scholars that study the relationship between software and culture should read the book for its consummate use of digital methods. Digital methods are those developed to study ‘natively digital’ objects that may or may not have nondigital antecedents or corollaries.²¹ To make conclusions intelligible to those outside the field, there is an emphasis on the visual representation of findings. Using an approach the authors of Spotify Teardown characterize as ‘following the data,’²² they map the vast data infrastructure that Spotify uses to ‘connect, prompt, and link together various distributed elements across large geographic distances.’²³ A diversity of methods are employed to accomplish this ambitious task, including but not limited to: creating a nonprofit record label to examine how data and metadata are generated, aggregated, and made interoperable between Spotify and other companies; performing network protocol analysis to detect packet transmissions across Spotify’s ‘data centers, backbone providers, CDNs [content delivery networks], and advertisement brokers;’²⁴ and running experiments designed to determine Spotify’s advertising supply-chain vendors and the structure of its ad-tech market.²⁵
Readers alternatively interested in how the case of Spotify challenges the frameworks of platform studies, the significance of Spotify for Swedish politics, or the ethics of performing research on digital media corporations averse to transparency will find much in Spotify Teardown. Though considering that the Spotify algorithm often stands in for Spotify itself in popular media discussions, I suspect that most people will turn to this text for an analysis of Spotify’s recommendation software, to which I direct the rest of this review.

Curation and Playlists

While Spotify Radio claims to provide both personalized and never-ending music, Spotify Teardown concludes that it does neither. Drawing on an experiment that used bot listeners to track streams, the authors visually represent how its algorithm produced music loops that ‘displayed a repeated pattern with only slight variations according to which artist a radio station was based on.’²⁶ Not only did these loops ‘look more or less the same independent of bot characteristics’ but ‘user feedback of “thumbs up” (like), “thumbs down” (dislike), or skip did not produce significant differences in the results.’²⁷ Their experiment confirms that Spotify has exaggerated its computational claims about its radio function while also corroborating user complaints about Spotify Radio’s repetitiveness and recommendation inaccuracy.²⁸
More provocatively, Spotify Teardown questions the ‘normative claim that the radio algorithm should produce apt recommendations.’²⁹ Recommendation inaccuracy may drive users away from its free service and toward a paid subscription.³⁰ It may also encourage users to stream Spotify’s Featured Playlists instead of its radio function. And finally, it might be a covert way to promote more established and financially well-supported major label artists at the expense of independent and ‘emerging musicians or neglected genres (with economic ramifications).’³¹
Independent of its accuracy, personalization is critical to Spotify’s marketing and business model. Spotify Teardown explains that the trope of personalization encourages users to develop a sense of intimacy with Spotify.³² For example, commercial playlists are marketed as ‘mixtapes,’ adding ‘a nostalgic and affective value’ to what are actually closer to compilation albums, wherein ‘business-related aspects are more important than the creative process.’³³ In addition, the assumption that more input in Spotify’s algorithm will produce increasingly accurate future recommendations compels users to do so. This means more revenue for Spotify because data produced from usage is used to lure advertisers.³⁴ While personalization remains significant for Spotify, it has since deemphasized its radio functionality in favor of in-house playlists and ‘new computational recommendation formats based on taste profiles, song identification, and digital fingerprints.’³⁵
The authors refer to this shift that began in 2012-3 as Spotify’s curatorial turn. The turn cemented the playlist’s status as Spotify’s primary object for streaming and marked a new emphasis on algorithmic and human recommendations.³⁶ They write, ‘A large number of these playlists are created by third-party services such as Filtr, Topsify, or Digster, owned by Sony, Warner, and Universal respectively—the three major record labels that, in turn, own stakes in Spotify.’³⁷ Prior to the turn, Spotify functioned as an on-demand service whose interface resembled iTunes, where users could search for music and assemble personal playlists. Spotify’s marketing corresponded to the design of its search-based interface, both foregrounding that consumers could choose from (then) over 20 million individual songs contained in its massive database.³⁸ In other words, Spotify emphasized consumer choice rather than recommendations, individual songs rather than playlists, and musical quantity over quality. Thus, Eriksson et. al argue that, ‘The user was effectively conceived of as a sovereign individual, who already knew exactly what he or she wanted to listen to and did not need help with music recommendations.’³⁹
Spotify’s move away from this individualist conception of users took different forms. In December 2012 Spotify announced it would introduce a Discover tab and a Follow tab. The former included ‘a new, personalized recommendation function’ based on users’ listening history.⁴⁰ The latter would provide recommendations from ‘artists, trendsetters, editors and experts’ that users could ‘follow.’⁴¹ They write, ‘This was a step away from the symmetrical sociality of Facebook (where friendship is a two-way relation) toward the asymmetrical following system that characterizes Twitter (where a small number of users tend to become highly influential).’⁴² Both tabs indicate that Spotify had reconceived of users as in need of guidance through its catalog.⁴³
Soon after emphasis would shift once more. Rather than foregrounding access to its database of songs, Spotify started emphasizing musical quality and playlists appropriate for different contexts. With the help of Tunigo—which Spotify would acquire in May 2013 and integrate into its client—Spotify increased the importance of its playlists based on activities and moods. Tunigo’s team of about twenty ‘music experts’ were joined by an expanding group of ‘music editors’ who were hired to curate local playlists for countries where Spotify was available.⁴⁴ The following year Spotify acquired the Echo Nest, which had run Spotify’s and its competitors’ algorithmic recommendations. The Echo Nest enabled Spotify to offer features that went beyond taste-centric personalization by incorporating users’ spatial and temporal data into its algorithms.⁴⁵ For example, the Spotify Running feature ‘used the smartphone’s sensors to detect the pace of a runner and play music at the same speed.’⁴⁶ In 2015 Spotify purchased Seed Scientific, another music intelligence company, and introduced its immensely popular Discover Weekly personalized playlist that updates every Monday. Discover Weekly anticipated other weekly playlist offerings like Release Radar and Fresh Finds. Thus the curatorial turn marks when ‘Spotify began to transform itself from being a simple distributor of music to the producer of a unique service.’⁴⁷
It also signaled a different conception of the Spotify user: ‘If the assumption had previously been that musical taste is a property of the individual, now Spotify seemed to implicitly accept the view that it is rather “an aggregate of the supra-and the subpersonal”.’⁴⁸ Spotify utilizes what it refers to as ‘taste profiles’ to inform its algorithmic recommendations, which use processes of collaborative filtering to identify users with similar listening patterns.⁴⁹ Following John Cheney-Lippold, they argue that algorithmic identities are kinds of measurable types produced out of observed data patterns that nevertheless ‘do not necessarily correspond with our nondatafied self-identifications or sociopolitical identities.’⁵⁰ Taste profiles are at once personal and social. On the one hand, they record a user’s past streams of artists and genres, assigning ‘scores that measure how heavily, actively and regularly those artists and genres are played and how much is streamed from an artist’s full catalog of music.’⁵¹ On the other hand, that data is mapped not only in relation to other users but to descriptions of artists and genres that are pulled from online sources including ‘blog posts, music reviews, tweets and social media discussions.’⁵²
Curiously, Spotify Teardown’s analysis of Spotify’s recommendation algorithm does not account for how the Echo Nest relates the two aforementioned sources of data to its own musical analysis software. As Robert Prey explains:
Unlike Pandora’s manual, labor-intensive method of aural classification, The Echo Nest utilizes acoustic analysis software to process and classify music according to multiple aural factors—from its pitch to its tempo to its danceability. ‘The system ingests and analyzes the mp3, working to understand every single event in the song, such as a note in a guitar solo or the way in which two notes are connected’, explained Brian Whitman, co-founder and CTO of The Echo Nest. ‘The average song has about 2000 of these “events” for the system to analyze. It then makes connections between that song and other song with similar progressions or structures’ (as cited in Darer, 2012).⁵³
Thus, the user’s taste profile fits into a larger musical map that connects sonic analysis of songs to semantic analysis of online discussions about artists and genres.⁵⁴
Taking up Anahid Kassabian’s concept of ubiquitous listening, Spotify Teardown argues that Spotify playlists exemplify and encourage a utilitarian approach to music, where music’s function is to accompany and enhance other activities rather than being an aesthetic experience unto itself.⁵⁵ For example, they argue that Spotify’s Featured Playlists, which update several times throughout the day and are accompanied by messages that subtly direct one’s activity, reproduce ‘chrono-normative prescriptions of “the good life” that instruct users to get out of bed, go to work (in an office), work out in the afternoon, and then socialize with friends, family, and lovers in the evening. Meanwhile, music is presented as a way of increasing productivity and performance in these time-bound activities.’⁵⁶ As I am writing this review, I am listening to a playlist entitled Intense Studying whose description reads, ‘Focus-enhancing piano for your study session.’ I cannot tell if it is working.
While the authors’ close readings of Spotify’s interface and recommendation software show that bourgeois ideology is embedded in the platform’s design, it is particularly acute in Spotify’s Mood playlists. Although Spotify claims to tailor playlists for our different moods, Eriksson et. al argue that the experience of listening to Spotify ‘evokes fantasies of one specific state of mind and the moral values that come with it: happiness.’⁵⁷ Mood playlists’ narrow idea of happiness is one that comes from the gendered discourse of popular positive psychology, where happiness is largely a result of a self-governing subject’s cognitive outlook, unrelated to social structures and power relations.⁵⁸ Thus, playlists like Happy Hits! whose description reads, ‘Hits to boost your mood and fill you with happiness,’ incite users to manipulate their emotions and ‘direct their desire for change inwards.’⁵⁹ Yet, rather than analyzing any music in these playlists, which would enable the authors to assess the relation between a playlist’s form and its content, the authors take a strong relativist position on music. They write, ‘Affective responses to and connotations of music are, of course, highly subjective. The same songs can be included in differently themed playlists and the same playlists can be found in different music categories.’⁶⁰ Playlist images, however, are deemed worthy of analysis.
Drawing on Lev Manovich’s concept of Instagramism, they observe how the formatted square-shaped images associated with Spotify’s Mood playlists lack contextualization, similar to commercial stock photos. Consistent with the aesthetic that emerged in the early 2010s, where Instagram users ‘stag[ed] unique moments, feelings, and states of being,’ Spotify playlist icons depict ‘relaxed and joyful atmospheres through which individuals are turned into cultural stereotypes of the young and happy middle class.’⁶¹ While the images tend to ‘foreground people presenting as women,’ the music contained in Spotify playlists is performed overwhelmingly by male artists. The authors of Spotify Teardown conclude that by representing women primarily as consumers and men primarily as artists, Spotify reproduces gendered inequality in the music industry rather than challenging it.⁶²

Conclusion

One might expect that a book about the world’s leading music streaming service would examine how Spotify has influenced musical trends or development, not only because music is the primary type of media that Spotify distributes and recommends, but because critical histories of earlier modes of for-profit music promotion and playlist construction—namely, commercial radio—require recourse to musical materials themselves. This has to do with the ways that music genres were incorporated into the practice of radio formatting since the postwar period.
The development and near immediate success of the Top 40 radio format in the early 1950s led independently-owned stations to move away from an airtime-for-sale model to an audience-for-sale model. With the former, revenue was generated out of ‘an agreement with the program provider to buy the station’s airtime to run the show and the provider’s commercials.’⁶³ Hour-by-hour segments of airtime were placed largely in the hands of advertisers, which resulted in inconsistent programming of variable quality. But under the audience-for-sale model, the goal became to ‘convince sponsors of the link between a mediated product and its never fully quantifiable audience.’⁶⁴ By segmenting audiences ‘explicitly by age and gender and tacitly by race and class,’ their consumption habits could be more competently predicted and thus advertising more persuasively sold.⁶⁵ Radio programmers constructed playlists based on perceptions about the demographic their station was targeting.
The results of the airtime-for-sale model were that programming became a station’s internal responsibility, all-day popular recorded music programming was regularized, and radio stations began to use the same format all hours of the day. Formats thus became identical with stations themselves, as well as their primary means of differentiating themselves from competitors. As Jody Berland explains:
Every format follows a complex set of rules for programming, including the style and range of music selections, size and origin of playlist, quotas for musical repetition, relative numbers of current and past hits and their usual sequence, conventional relationships between music and speech, and so forth. A major change in any one of these is inconceivable without a subsequent change in all of them and in the relationships amongst them.⁶⁶
Consider how Free Form rock radio developed as a response to Top 40. As radio became popular music’s most advantageous vehicle for promotion and hit-making, record producers scrambled to fit the expectations of formats. As radio historian Marc Fisher writes about Top 40’s impact on songwriting: ‘A hit song had to be instantly recognizable, like station jingles. The new songs were repetitive, with catchy hooks taking precedence over longer melodies. Hits came with instrumental lead-ins, so deejays could talk over the introductions’⁶⁷. So too, songs became progressively shorter because short songs had a better chance of making it on to tight playlists.⁶⁸ If during the first half of the twentieth century the technological limits of the shellac 78-rpm record required musical brevity, the Top 40 format reinforced concision at almost the very moment the LP had liberated popular song from the 10-inch disc’s temporal shackles. It was this influence—along with Top 40’s overwhelming commercialism—that Free Form FM radio would explicitly challenge in the mid-to-late 1960s by providing a platform for an emergent countercultural rock scene. Rock bands began to take advantage of the LP medium, recording complex, introspective, and often conspicuously political songs that sometimes went three or four times longer than the average Top 40 hit. Relatedly, Free Form DJs frequently refused to run spots from national corporations, particularly those understood to benefit financially from the Vietnam War.⁶⁹ Yet by the early 1970s, radio stations’ profitability requirements tended to undermine the capacity of DJs to turn down advertising dollars, while market research findings compelled programmers to take away their ability to choose their own songs. These forces would result in a form of playlist rationalization that anticipated the contemporary recommendation algorithm. While this is necessarily an abbreviated exposition, it is sufficient to exemplify how the history of commercial radio since the postwar period is simultaneously the history of radio formatting—that is, the linking of demographic groups to musical genres as well as advertisers in an all-encompassing style—and therefore, that historians must at least gesture toward an interpretation of how music fits within a format to understand the developmental dynamics of commercial radio.
By contrast, Spotify’s Featured Playlists are not a part of a larger formatted whole. Nor are they affiliated with individual stations with their own identities. (Let us bracket questions surrounding radio station consolidation, program syndication, and the formation of satellite radio for the sake of brevity and conceptual clarity.) Instead, a multitude of playlists are contained within a single platform. Yet these playlists also appear to stand on their own, as it were, recalling the sponsored hour-by-hour programming that once filled time on the radio one hundred years earlier.⁷⁰ Has this development made musical content inconsequential to understanding Spotify?
For all its insights, Spotify Teardown has very little to say about Spotify’s influence on the production of music today. Posed in theoretical terms, the question becomes: what is the relationship between Spotify and musical forms? It is hard to imagine that the failure to pose this question is a matter of the authors’ subjective oversight. So what does its absence tell us? Is it simply that Spotify does not have the symbiotic relationship with popular music that commercial radio once had? Has Spotify’s algorithm obliterated the kinds of constraints that formatting previously put on popular music production? Or has it only made them more obscure? Has the practice of micro-targeting users based on their taste profiles, moods, and contextual data not only undermined ‘the traditional order of knowledge that organizes music according to genre’ but genre formation and development itself?⁷¹ Is it possible that the transformations from ownership to access to context, and from storage culture to streaming culture, have not affected musical content?
As scholars and critics address these questions over the coming decade, Spotify Teardown will undoubtedly be an intellectual touchstone.⁷²
While the aforementioned questions clarify the limits of Spotify Teardown’s problematic, we must work our way back to identifying the problematic itself, and to its conditions of possibility. In his 2015 essay “The Aesthetics of Singularity” Fredric Jameson frames the turn towards curation as part of a larger historical situation, as the latest aesthetic practice generated by a financialized global capitalist system. By positing a homology between financialization, cultural production, and aesthetic experience, he provides an account of temporality—the “reduction (of time) to the present”—that appears to unify these disparate levels of the social totality.⁷³
Jameson’s analysis of curation, of course, does not refer to music recommendation software but to what is required once the system of fine arts has collapsed, where once distinct arts and media—painting, photography, performance, video, sculpture, etc.—are no longer able to be clearly differentiated. In this situation postmodern works appear as collages—‘one-time unrepeatable formal events (in their own pure present as it were)’—rather than the creation of new stable cultural forms.⁷⁴ ‘(P)aradigmatic of the postmodern artistic practice,’ he argues, are installations, which are not objects made for posterity or for the museum’s permanent collection, as were older forms of art.⁷⁵ They are events, organized around the moment of their exhibition. In their very form, installations are thus a ‘replication of the new museum in which it is housed… whose exhibits and cultural events are the equal of musicals or eagerly awaited films.’⁷⁶ As the contemporary museum has become ‘a popular and mass-cultural space’ and a siphon for global finance capital—where the paradigmatic derivative functions ‘more like a unique event than a contract’—it has undergone its own shift ‘from access to context,’ to use the parlance of Spotify Teardown, where our aesthetic experience of encountering artistic works takes on the character of the consumption of brand names.⁷⁷ Jameson argues that under these circumstances, ‘the form of the work has become the content; and that what we consume in such works is the form itself’—its temporality, its status as an event.⁷⁸
There is an analogy to be made here. Streaming—itself a metaphor—suggests a ‘continuous flow of music’ with a presentist temporality.⁷⁹ With streaming, data files are not stored on one’s hard drive to be retrieved for future use, but to be ‘played immediately after a small amount of audio data has been received.’⁸⁰ Copied from data centers, passed through a vast infrastructure in pieces as they make their way from servers to the Spotify client, streams are for the now. So too, Spotify’s contextual playlists are organized for the now, to guide and accompany events in one’s day. If we understand Spotify’s products as embodying presentist temporalities analogous to Jameson’s installations, if what we consume is the idea of a playlist—its form or its concept, be it a mood, an activity, etc.—rather than its sensory content—Spotify Teardown’s methodological formalism is brought into relief. It is the fact that listening to Spotify has its own structure of feeling—distinct from the formatted flows⁸¹ that characterized radio broadcasting, and where the process of curation eclipses what is curated—that makes formal analysis appear comprehensive. Thus, the omission of musicological analysis in Spotify Teardown must be understood not as an oversight, but as a consequence of Spotify’s design and effects.
It should suffice to conclude that the curatorial turn and the shift from access to context are not unique to Spotify, nor music streaming, nor online platforms. Indeed, if we follow Jameson on this point, the phenomenon might in fact be related to the infrastructural dynamics of financialization, which Spotify Teardown takes as its problematic, and of which it is also a symptom.
Jameson argues that in the museum the curator stands as the only human left in an institution no longer representable at a human scale, as the allegorical personification of the museum’s fusion with global finance capital.⁸² Headquartered in Sweden but with satellite offices around the world, and with a database of over 50 million songs, podcasts, and videos, Spotify operates at a far greater scale than any individual museum, no matter how globalized its collection or its financial partners. While a financialized global capitalist system is as much a condition of possibility for the contemporary museum as for Spotify, the latter points us beyond Jameson’s object, where even the figure of the curator has taken on the temporality of finance, replaced by that of the algorithm, the ‘black box’ that automates curation at the same computational speed Spotify sells micro-targeted audience segments to advertisers on its online trading desks.⁸³

Notes