Popular Music Reception, Data, and Digital Technologies (continued)
by Craig Hamilton
Part two of two. For part one, click here.
A starting point is Webster et al’s (2016) argument that a key function of automated recommendation in digital music interfaces is the leveraging of a competitive advantage in a crowded and undifferentiated marketplace. When all services offer the same (or largely the same) catalogues of music, at similar price points and in similar ways, one of the only competitive spaces that remains is the quality of listening experience delivered. As Vanderbilt (2016) shows, in the construction and iterative rationalisation of automated recommender systems, implicit feedback – which can be understood as data gathered about which songs are played, skipped, shared, or added to playlists – is often viewed as a more useful ‘raw’ material than the explicit feedback volunteered by users in the form of star ratings, purchases or reviews. It is here where complex and inter-related acts of cultural translation occur: from the reduction of an individual’s experience with a song to a data point, through the algorithmic processing of that data at scale, to the foregrounding of one type of music over another to publics via dynamic interfaces – whereupon the cycle repeats.
Tania Bucher’s concept of “the algorithmic imaginary” (2016) is a useful way of thinking through this. It allows us to understand both how data-processing impacts upon experience, but also how experience impacts upon the design, function and use of algorithms. The algorithmic imaginary can be observed in action through a consideration of automated music recommendation services in particular: data is gathered on listener activity from which abstracted inferences of taste are derived; leading to recommendations that can positively or negatively influence choice; which in turn creates data about listener tastes; and the process repeats. Interestingly the decision of a user to ignore a recommendation is equally important here because it too creates data that is used to tweak recommendation algorithms. It is important also because although ultimately listeners can choose whether or not to follow recommendations, they cannot chose whether or not their activities are recorded and subsequently used in the creation of recommendations. Our relationship with such systems is thus not entirely top-down, but rather one of co-production that is based on an unequal relationship. The conditions under which the user operates are not entirely known and are inescapable as long as the user continues to use Spotify, ultimately helping produce consequences in the form of recommendations and dynamic changes to the user interfaces that foreground or not particular content. It is in around these points where debates of the potential outcomes of relationships between cultural practices and digital monitoring find their foundation.
Tufekci (2015), for instance, highlights the negative reactions to Facebook’s emotional contagion study (Kramer et al., 2014), which measured the emotional effects on users of different types of content, and argues that questions around the potential harms/benefits of the algorithmic production of experience have moved “beyond hypotheticals” now that “algorithms act as de facto gatekeepers of consequence” (2015: 206). Claims of this kind have gained considerably more traction in light of recent developments around Facebook and Cambridge Analytica. In terms of popular music, however, this is not limited to the delivery of recorded music via digital interfaces. Bucher’s algorithmic imaginary is also at play in the foregrounding of media content and advertising through social media and news media platforms, and increasingly in the promotional and A&R activities of record companies (Thompson, 2014). These are all areas where algorithms, fuelled by consumer activity data and cultural content metadata, are deployed as subjective decision makers.
The irony amidst all this is that the systems and practices causing the concern are facilitated partly by users’ engagement with digital interfaces. As such, the users themselves are can also be viewed as the ‘agents’ facilitating the translation of a cultural form from one context to another. Spotify would not be able to offer Discover Weekly in its present form if people did not use their system to create playlists. Likewise, the recent debates around Facebook, Cambridge Analytica and the implications for elections, would not exist without the users and their everyday use of the platforms concerned. Van Dijck (2014:1) argues that user data has become “a regular currency for citizens to pay for their communication services and security – a trade-off that has nestled into the comfort zone of most people.” This uneasy covenant has been described by Barnes (2006) as a “privacy paradox”, where people are uneasy about the information collected about them, which they know will be packaged and monetised, but nevertheless accept this as a condition of using certain services. As such, listeners, users, and publics have their roles, agencies and choices and are by degrees similarly implicated in the concerns raised by Bucher, van Dijck, Tufecki and others. It is in the tensions that exist between these relationships that I find the location for my own research into popular music reception, data and digital technologies.
One of the key research findings from my doctoral analysis of the thousands of text-based reflections gathered by the Harkive Project indicated that respondents are developing intriguing new cultural practices based around their engagement with a variety of digital, online and data technologies. From this I have speculated on plans for post-doctoral work that will involve the creation of tools and interfaces that could enable us arrive at more meaningful, critical and reflexive relationships with the data technologies that are now so central to our everyday lives – in other words, for users to better understand the forms, processes, roles and agencies that are in play when cultural translation occurs during their everyday acts of music reception. My hope is that work in this area may contribute to wider debates around data literacy and critical data/algorithm studies, and I look forward to seeing what Harkive 2018 will tell me about this on Tuesday 17th July.
Craig Hamilton is a Research Fellow in the School of Media at Birmingham City University. His research explores contemporary popular music reception practices and the role of digital, data and Internet technologies on the business and cultural environments of music consumption. This research is built around the development of The Harkive Project (www.harkive.org), an online, crowd-sourced method of generating data from music consumers about their everyday relationships with music and technology. Craig is also the co-Managing Editor of Riffs: Experimental Research on Popular Music (www.riffsjournal.org)
Useful links
The Harkive Project website: www.harkive.org
Harkive on Twitter: https://twitter.com/harkive
Harkive on Facebook: https://www.facebook.com/harkive/
Bibliography
Adorno, T.W., Simpson, G., 1942. On popular music. Institute of Social Research.
Barnes, S.B., 2006. A privacy paradox: Social networking in the United States. First Monday 11.
Bourdieu, P., 1984. Distinction: A social critique of the judgement of taste. Harvard University Press.
Bucher, T., 2016. The algorithmic imaginary: Exploring the ordinary affects of Facebook algorithms. Inf. Commun. Soc. 1–15.
Cohen, L., 2004. A consumers’ republic: The politics of mass consumption in postwar America. J. Consum. Res. 31, 236–239.
Kramer, A.D., Guillory, J.E., Hancock, J.T., 2014. Experimental evidence of massive-scale emotional contagion through social networks. Proc. Natl. Acad. Sci. 111, 8788–8790.
Lears, J., 1995. Fables of abundance: A cultural history of advertising in America. Basic Books.
McCourt, T., Rothenbuhler, E.W., 2004. Burnishing the brand: Todd Storz and the total station sound. Radio J. Int. Stud. Broadcast Audio Media 2, 3–14. https://doi.org/10.1386/rajo.2.1.3/0
Tufekci, Z., 2015. Algorithmic harms beyond Facebook and Google: Emergent challenges of computational agency. J Telecomm High Tech L 13, 203.
Vanderbilt, T., 2016. You May Also Like: Taste in an Age of Endless Choice. Knopf.
Van Dijck, J., 2014. Datafication, dataism and dataveillance: Big Data between scientific paradigm and ideology. Surveill. Soc. 12, 197.
Webster, J., Gibbins, N., Halford, S., Hracs, B.J., 2016. Towards a theoretical approach for analysing music recommender systems as sociotechnical cultural intermediaries, in: Proceedings of the 8th ACM Conference on Web Science. ACM, pp. 137–145.