The Affective Labour and Performance of Live Streaming on Twitch

Woodcock, J. and Johnson, M. R. (2019) 'The Affective Labour and Performance of Live Streaming on Twitch.tv', Television and New Media, 20(8): 813-823.

Introduction

“Live streaming” entails the live broadcast of video content to viewers over the internet. This practice has become dominated in most countries by the website Twitch.tv, which began broadcasting only video game content but has expanded to include a range of content, bringing with it a market value of approximately one billion US dollars. The platform boasts over one hundred million regular viewers, two million of whom broadcast their own material on the platform to viewer communities ranging in number from single-digits to millions. Those who are able to bring in the largest crowds now make a living on the platform through a range of monetization techniques. While most Twitch streamers are amateur producers, a significant number earn their living through streaming, while others are involved in associated activities like commentating on “Esports” (competitive gaming) competitions. A primary element of streamers’ labor is performance, much of it invisible and – until and unless a broadcaster becomes highly successful – unpaid.

Live streaming and game commentary are important for critical media studies for two reasons. Firstly, live streaming is a major global phenomenon and one only expanding in scope and reach. The scale of the audience is now impressive enough to rival many television channels and traditional sports broadcasts, whilst it is also becoming an increasingly central element of digital gaming culture more broadly. Secondly, streaming represents a career path that many young people – disproportionately impacted by the financial crisis – are pursuing, finding the apparent opportunity to play games for a living understandably compelling. These factors make Twitch an important element of contemporary youth employment dynamics, especially for those disaffected by, or unsuccessful in, traditional education or career paths. Therefore, this study of Twitch streamers highlights an emerging area within digital games that can be best understood through considering the labor that goes into the activities involved.

In this article, we examine the affective dimensions of video game streaming labor, which involves being compelling to watch and friendly to viewers, soliciting donations, building parasocial intimacy with spectators, and engaging audiences through humor. Our analysis pays particular attention to the extent to which streamers broadcast as “themselves”, or as a “character.” This disjunction is a tactic of self-branding (Hearn 2008; Marwick 2013) in an increasingly crowded marketplace. We also examine these kinds of labor within the dynamics of the platform, shedding light on the stresses and challenges for streamers. This inquiry is intended to examine the labor that Twitch streamers perform, and in doing so understand some aspects of the work that goes into this rapidly-growing, and increasingly visible, contemporary digital games labor practice.

Twitch.tv and Live Streaming

Although Twitch is an area of increasing scholarly concern, there is presently limited research on the platform. The rise of user generated content online has been a focus of academic interest (Banks and Humphreys 2008; Banks and Deuze 2009), with some viewing this as a progressive and participatory shift (Jenkins 2006; Burgess 2009), and others being more critical of the role of platforms in exploiting users (Hesmondhalgh 2010; Duffy 2015). Twitch, as a platform, can be compared with YouTube. In particular, Hector Postigo’s (2016) study of video game commentary on YouTube is an important reference for our analysis. Postigo (2016, 332) focuses on the technical and social affordances of the platform, examining how “the hybridity of play and production” provide the opportunity for transforming leisure into a form of work. Despite Twitch’s growth, “there is not much research studying this new merging and prosperous online community” (Churchill and Xu 2016). Kishonna L. Gray (2017) has explored questions of race and the experience of Black gamers streaming on Twitch, Max Sjöblom and Juho Hamari (2017) have looked at why people watch streamers, (Author removed for review) has examined disability and health on the platform (2018), and Emma Witkowski (2011) has examined the roles, positions, and struggles of women players in the live streaming domain. Others have studied the different kinds of cultures different streamers gather around them (Gandolfi 2016), whilst the authors have previously examined the lives and careers of Twitch streamers (Author removed for review). Nevertheless, the literature remains at a nascent stage. The intention of this paper is to contribute to the literature on digital games labor studies, an approach that has so far not been applied to Twitch.

Our research combines interview and ethnographic data. We draw on 100 semi-structured interviews conducted in 2016 and 2017 with professional and aspiring streamers and Esports commentators, of which approximately one-thirds identified as women and two-thirds as men. These interviews lasted between ten minutes and over one hour, and took place at gaming events in Poland, Germany, the United States, and the United Kingdom. The interviews focused on the themes of labor, performance, and work. Transcripts were coded and interviewees have been assigned pseudonyms. The authors also carried out ethnographic observation at these gaming events, which provided context for the interviews. This research was supplemented by observation of 200 unique live streams on Twitch from a diversity of streamers. The multi-sited ethnographic research helped to ground our analysis in not just the comments of live streamers as interview respondents, but also their lived experience at these important and highly visibleevents during which they must perform and be publicly engaging for long stretches of time. The combination of methods provides much needed context to Twitch streaming culture and enables us to access participants in an online phenomenon that are otherwise difficult to reach.

Affective Labor on Twitch

To make sense of one crucial part of generating the appeal of watching live streams – the interpersonal and social dimension – this article builds on arguments about “affective labour” (Hardt and Negri 2004). This starts with Arlie Hochschild’s (2012, 7) notion of emotional labor, which denotes a requirement to “induce or suppress feeling in order to sustain the outward countenance that produces the proper state of mind in others.” On Twitch, this requires the modulation of emotions and affect through performance for the streamer. Affective labor takes this further, emerging from Autonomist studies of work and refers to efforts designed to generate emotional responses, such as those of service workers who are expected to always be smiling, or the importance of “attitude” to one’s employability. This involves the communicative aspects of immaterial labor, referring to activities that do not resemble traditional work, but have been effectively commodified. This might include the production of user-generated content as in our discussion, many kinds of creative work, the performance of identity and legitimacy in one's domain of employment, and so on. These closely-related forms of labor are all interwoven in the activity of live streaming.

Twitch is a deeply social platform for content production and distribution. Beyond a simple broadcast relation, streamers and viewers communicate to one another through the platform’s chat system. The browser presents the user with three interactions: first, the live-streamed gameplay (mirroring what the streamer sees); second, an embedded live-stream of the streamer, showing them narrating and playing the game; and third, a chat window to the side that the audience can type messages to the streamer and each other. These aspects are crucial to the platform’s appeal, allowing multiple avenues for engaging viewers through the streamer’s performance. Beyond these affordances of the platform, streamers can choose how to perform in order to draw an audience. This is a challenge of this kind of creative labor, given the “difficulty of predicting public tastes and the impossibility of exactly duplicating a hit” (Prindle 1993, 5). What works for one streamer may not work for another, but the streamer attempts to produce “intangible feelings of ease, excitement, or passion” (Hardt and Negri 2000, 293).

For some streamers this means using humor. One streamer, Philip, explains, “I like to fancy myself somewhat of a comedian,” and indeed many streamers noticeably seek to be witty in their broadcasts. However, for others this means using more broad concepts, such as Marie’s comment of the central roles played by “my personality, my commentary, my reactions to things.” She drew on these as a “variety streamer,” or someone who plays a wide range of games, and as a streamer able to rapidly adjust to the demands of a particular game. What works for one streamer may not for another, and what is popular at one point can quickly become outdated. In Elizabeth Wissinger’s discussion of fashion model labor, we see the impacts that this variability in appeal can have: “fashion models are valued for their ability to unleash a wide range of responses, responses that might shift or be modulated faster than they can be subjectively recognized as emotions” (Wissinger 2007, 300). In a comparable way, Twitch streamers are expected to respond quickly to comments in chat, to in-game developments, to be emotional, to vary how they act according to game context, viewers, and time of day – and to modulate their performance to both current and potential viewers.

Despite commonalities with other forms of work like “vlogging” (Hou 2018), Twitch is developing this kind of online performance in new directions. Streamers broadcast a webcam video of themselves, which appears to viewers embedded within the game the streamer is playing. This picture-in-a-picture of the streamer playing the game illustrates how Twitch affords new ways to mediate affective performance, allowing the audience to connect with facial expressions, voice, and immediate surroundings. Streamers’ physical reactions are thus embedded visually within the virtual experiences generating those reactions – winning or losing, discovering something new in a game, and so forth. Equally, as John Short et al. (1976) demonstrate with social presence theory, the more acoustic, visual, or physical contact a medium allows, the greater the possibility for social impact. The recreation of the social setting of the living room through the direct facial contact with the player, the active conversation, and the relaxed, friendly manner in which streamers talk, all recreate the local multiplayer setting of video games, while simultaneously allowing people to view remotely from across the world. We also note that almost all successful streamers have set stream times, which allows a more personal connection to the content than an on-demand service like Netflix, being closer to a visit to a friend’s house than the acquisition of content that can be viewed any time. In this way, live-streaming reinforces the sense of spending time with a friend.

Alongside the pursuit of a stream that mirrors the comfort and intimacy of real-world interactions, we also note trends that seem to move in the opposite direction: specifically, in relation to the notion of streaming in character. This takes a range of manifestations, from mild to intense. On the mild end of the spectrum, we can look to the comments of Todd, who described how they are “definitely more animated on stream.” Similarly, Samuel, reflecting on their own practice, noted that streaming required “turning on your personality or turning on a character” and that “I think that I am like a, sort of, more animated version of myself (when I am streaming).” This animation is a form of “outward countenance,” (Hochschild 2012, 7), mediated through an ensemble of affects, deployed to make a connection with a distributed audience, all while playing a video game. The flow of affect that is mobilized within a stream is a real-time version of what Brian Massumi (2002, 36) describes as feeling “one’s own vitality, one’s sense of aliveness, of changeability.” Streamers have found that an effective way to keep viewers engaged in their stream is to be extremely active, even hyperactive, always giving viewers something new to look at, respond to, or comment on, and minimizing the “down time” that might lead to disinterest, and thus a viewer turning off.

Others, however, take the notion of streaming “in-character”, and streaming as containing an element of theatricality, further. As one streamer, Daniel, explained, the act of streaming should be understood fundamentally as “a performance,” including needing to emotionally prepare to ensure the performer is in the right “mood” beforehand. In contrast to recorded broadcast media content, “it’s all improvisational [because] Twitch is a live environment,” as Philip put it, who drew on their experience with improvisational comedy and the theatre. They continued, “I’m used to performance. I’ve done leading roles. I’ve done side roles. Character acting is my favourite thing to do. So, I like having the quirky, out there, has a, kind of, weird-tick-going-on character.” Another respondent, Rob, an Esports commentator, had previously acted in Shakespeare plays, but found that “with all that background behind me, I still realize how hard it is.” Many of our interviewees find that treating Twitch as an acting performance was helpful, albeit a performance designed to look like a comfortable, organic, relaxing time with friends. However, this still meant acknowledging that even experience with acting fails to fully prepare one for the intensity, and the volume, of labor required on the platform.

This more explicit acting many rely upon on Twitch was noted by numerous interviewees who had gone so far as to invent characters for their stream. Connor, for example, streamed in the guise of a pirate: “yeah, it’s a character, but it’s based on me as a person and the enhanced version of that.” This began because Connor “liked theatrics” and “so for my channel, I got a pirate ship on there.” This led to increasingly speaking in an archetypal pirate manner on Twitch, and branding his stream around the idea of piracy, until a distinctive character emerged. Another interviewee, William, who had “a Norse look,” said that “people kinda just put it up on me and I’m like, OK, I’ll be a Norse character.” Such explicit characterization accentuates the significance of performance, both for streamers who pretend to be, say, a pirate or Viking, and for streamers who put forward a seemingly “authentic” personality (Duffy 2017). This performance imperative introduces new pressures on streamers, which illustrate expectations that Lazzarato (1996, 135) associates with immaterial labor: “one has to express oneself, one has to speak, communicate, cooperate, and so forth.” It is no longer adequate to just be one’s self, as one must shape the kind of subject one appears to be on stream.

While motivations to undertake the affective labor of streaming requires further study, we present some initial hypotheses. First, we suggest streaming in character is a way to manage the stress and effort that live streaming demands. Although on one level acting “in character” might appear to be more demanding than the alternative, it might allow streamers to separate their off-stream and on-stream lives, and thereby maintain greater control over their emotional labor by only showing a single, character-based, facet of themselves on their broadcasts. Equally, it seems that streaming as a character also serves to mark out these streamers from others on the platform. Streaming as a character is therefore an element of self-branding (Duffy 2018) that becomes so deeply associated with streamers that few of these individuals ever stream out of character, except in rare cases when giving some emotional disclosure to their fans, or promising a “serious” conversation with their fanbase about the future of their stream, their health, and so on. Nevertheless, there is clearly much more work to be done here, especially through drawing more detailed conceptual and empirical comparisons with other kinds of live performance; although the live performance on Twitch is undoubtedly unique, it is still likely to share many characteristics.

What is the point of such efforts? Our research shows that it is primarily to generate feeling in viewers, and an attendant sense of closeness or association. Wissinger’s (2007, 253) account of the labor of fashion models is especially relevant here, particularly her consideration of how “the production and manipulation of affect or feeling can also be achieved through human contact that is virtual (such as the work of an actor appearing in a movie, or the work of a model appearing in a photograph).” Unlike the model, whose picture is circulated after the moment of capture, the streamer is able to, and indeed must, maintain a real-time connection with an online audience. Streaming is a form of cultural production, entailing a “series of activities that are not normally recognized as ‘work’ – in other words, the kinds of activities involved in defining and fixing cultural and artistic standards, fashions, tastes, consumer norms, and, more strategically, public opinion” (Lazzarato 1996, 132). As one streamer, John, explained: “if you’re a little bit of a social butterfly, I think streaming works really well”: whereas the traditional popular culture image of the “gamer” suggests introversion, it is the most gregarious, emotionally engaged, and outgoing individuals who find success in live streaming.

The value of Twitch streaming

Streaming on Twitch professionally – “playing” video games for a living – or aspiring to do so, is contingent on numerous forms of emotional, affective or immaterial labor, that transform that play into work. Such a blurring of work and play has been discussed at length in the context of older digital platforms, like Facebook and YouTube, and it is highly relevant also to Twitch. For example, Bruns (2008) has coined the portmanteau ‘produsage’, combining production and usage, while Kücklich (2005) opts for play and labor with ‘playbor’. In the literature on gaming specifically, Terranova’s (2000) concept of ‘free labor’ has been particularly influential, but is often shorn of its theoretical roots in Autonomist Marxism. This tradition is, unsurprisingly, critical of work and how it is organised under capitalism (cf. Virno and Hardt 1996; Negri 1989), often reading the emergence of video games as part of a subversive flight from work (Dyer-Witheford and de Peuter 2009; Kirkpatrick 2013). We can therefore see a broad and detailed awareness of these kinds of labor that blur traditional lines of work and leisure, and which allow us to more fully conceptualise this streaming labor.

A transition from playing video games as a form of leisure to doing so as a form of work is thus not straightforward. Postigo (2016, 341) uses a hybridized formulation to understand this transition: first, “gameplay,” to understand how this activity begins as a form of play (and one that many of us are familiar with); second, “making gameplay,” as this act of play becomes a form of cultural production; third, “making game pay,” as this is transformed into work as it becomes the activity through which the streamer makes a living. In the process, streaming develops from a ‘hobby’ into a ‘job,’ meaning that streamers, like the YouTube commentators that Postigo (2016, 343) studied, have to “produce content even if they do not want to.” This means a constant performance of their hobby just not in an individual, private sense – the enacting of gameplay in a digital environment – but in a public, closely-watched, and monitored context, where performance translates directly into income. As Philip explained, the work of the stream also meant re-watching recorded streams, “thinking back on any lessons on performance [...] for how to best connect, and best reach [viewers].” Streamers reflect on their own performances, which become less a question of watching one’s past leisure, and instead a reconstitution of that activity as work that needs to be improved, optimized, and turned into an even more convincing performance next time around.

For Twitch streamers there are two main ways that they can make an income. The first is through building an audience. Viewers on Twitch can be monetised through either paying a subscription to the streamer’s channel (which is split between the streamer and Twitch) or through making donations (either directly to the streamer or through Twitch’s in-app currency “bits”). This means that the larger a consistent audience a streamer can build, the higher an income they can make. The second method flows from this. Once a streamer has built an audience, it then becomes possible to negotiate with companies for advertising or sponsorship. This economic compulsion brings additional pressures and demands into the process. As professional streamer Ben Bowman (2017) notes, there are 168 hours in each week, and he is “expected to be live for as many of them as possible,” with only enough breaks between playing games to allow him to recover.

Elsewhere we have argued that Twitch streaming has been shaped by a “neoliberal subjectivity” that depends on the entrepreneurship of the self, equating success as the result of hard work via the volume of hours streamed per week (Author removed for review), which will generate an inevitable emotional toll. Bowman grew his channel by streaming “a mind-crushing” twelve to sixteen hours a day, every day of the week, for two years straight: “this,” he explained, “was the only way I could maintain growth.” Similarly, despite the glamour of contemporary Esports events, Vanessa explained that for a commentator, “a 15-hour day is not uncommon.” While getting to travel around the world for this work was exciting, the reality, she said, was: “I flew there, I saw the inside of the convention center, I flew back.” The need here for self-motivation in the face of monotony is captured by Bowman and others in a “common mantra,” which is to “Always Be Casting.” Streamers must be constantly “on,” friendly, or witty, and are required to remain in character for long periods, not only within a single day, but potentially for many consecutive months, years even. This takes a toll on even the most extroverted people, especially when maintaining intimacy is essential to having a career in an area with effectively zero job security, and it is clear from our research that live streamers feel these pressures of such labor.

Conclusion

In this article, we sought to elaborate on some of the particularities of the labor process involved when someone begins to stream on the Twitch platform, specifically the labor of being presentable, funny, engaging, and in-character. This is a central part of the complex relationships on Twitch between the streamer, the game, the platform, and the audience. In doing so, these forms of labor have been shown to be deeply creative and contingent on being socially active and emotionally responsive. What emerges on this platform is a performativity of micro-celebrity that is being mediated in novel ways: streamers seek to develop a unique manner of interaction with the audience, drawing on their own gameplay skills and personality, and, in some cases, on the theatricality of character acting. To viewers, streams might seem casual and not that far from the experience of watching friends play on a console in a living room; yet this belies the striking amount of affective labor that goes into a stream. We have presented several of the most important elements of streamers’ labor, but each of these is ripe for a more complete examination in its own right. In particular, the new pressures of being a professional streamer demand urgent attention. Given the rapid and ongoing growth of Twitch and the diverse forms of affective labor examined here, it is clear that understanding these forms of work on Twitch is central to theorizing digital games labor.

References

Banks, John. and Deuze, Mark. 2009. “Co-creative Labour.” International Journal of Cultural Studies 12: 419–431.

Banks, John. and Humphreys, Sal M. 2008. “The Labour of User Co-creators.” Convergence: The International Journal of Research into New Media Technologies 14 (4): 401–418.

Bowman, Ben. 2017. “How Twitch is turning “always be streaming” into a career with zero balance.” Polygon. Accessed July 3. http://www.polygon.com/2017/1/16/14240224/twitch-irl-risks-dangers.

Duffy, Brook Erin. 2017. (Not) getting paid to do what you love: Gender, social media, and aspirational work.New Haven, CT: Yale University Press.

Bruns, Axel. 2008. Blogs, Wikipedia, Second Life, and Beyond: From Production to Produsage. New York: Peter Lang.

Burgess, Jean. 2009. YouTube: Online Video and Participatory Culture. Cambridge: Polity Press.

Churchill, Benjamin. C. B. and Xu, Wen. 2016. The Modem Nation: A First Study on Twitch.TV Social Structure and Player/Game Relationships. Paper presented at the 2016 IEEE International Conferences on Big Data and Cloud Computing (BDCloud), Social Computing and Networking (SocialCom), Sustainable Computing and Communications (SustainCom). https://ieeexplore.ieee.org/document/7723697/citations

Duffy, Brooke. 2015. “Amateur, Autonomous, and Collaborative: Myths of Aspiring Female Cultural Producers in Web 2.0.” Critical Studies in Media Communication 32 (1): 48-64.

Dyer-Witheford, Nick. and Peuter, Greig de 2009. Games of Empire: Global Capitalism and Videogames. Minneapolis, MN: University of Minnesota Press.

Gandolfi, Enrico. 2016. “To watch or to play, it is in the game: The game culture on Twitch. tv among performers, plays and audiences.” Journal of Gaming & Virtual Worlds 8(1), 63-82.

Gray, K Kishonna L. 2017. “‘They’re just too urban’: Black gamers streaming on Twitch.” In Digital Sociologies, edited by Jessie Daniels, Karen Gregory, and Tressie McMillan Cottom, 355-367. Bristol: Policy Press.

Hardt, Michael, and Negri, Antonio. 2001. Empire. Cambridge, MA: Harvard University Press.

Hardt, Michael, and Negri, Antonio. 2004. Multitude: War and Democracy in the Age of Empire. New York: Penguin.

Hearn, Alison. 2008. “‘Meat, Mask, Burden’: Probing the contours of the branded ‘self.’” Journal of Consumer Culture 8 (2): 197-217.

Hesmondhalgh, Desmond. 2010. “User-generated content, free labour and the cultural industries.” Ephemera 10 (3–4): 267–384.

Hochschild, Arlie. R. 2012. The Managed Heart: The Commercialization of Human Feeling. London: The University of California Press.

Hou, Mingyi. 2018. Social media celebrity and the institutionalization of YouTube. Convergence. Online first.

Jenkins, Henry. 2006. Fans, Bloggers, and Gamers: Exploring Participatory Culture. New York: New York University Press.

Kirkpatrick, Graeme. 2013. Computer Games and the Social Imaginary. Cambridge: Polity.

Kücklich, Julian. 2005. “Precarious playbor: modders and the digital game industry.” The Fibreculture Journal 5.

Lazzarato, Maurizio. 1996. “Immaterial Labor.” In Radical Thought in Italy: A Potential Politics, edited by Paolo. Virno and Michael Hardt. 133-147. Minneapolis: University of Minnesota Press.

Marwick, Alice E. 2013. Status Update: Celebrity, Publicity, and Branding in the Social Media Age. New Haven: Yale University Press.

Massumi, Brian. 2002. Parables for the Virtual: Movement, Affect, Sensation. Durham, NC: Duke University Press.

Negri, Antonio. 1989. The Politics of Subversion: A Manifesto for the Twenty-first Century. Cambridge: Polity.

Postigo, Hector. 2016. “The socio-technical architecture of digital labor: Converting play into YouTube money.” New Media & Society 18 (2): 332-349.

Prindle, David. F. 1993. Risky Business: The Political Economy of Hollywood. Boulder: Westview.

Short, John. Williams, Ederyn. and Christie, Bruce. 1976. The social psychology of telecommunications. London: John Wiley & Sons.

Sjöblom, Max. and Hamari, Juho. 2017. “Why do people watch others play video games? An empirical study on the motivations of Twitch users.” Computers in Human Behavior 75: 985-996.

Terranova, Tiziana. 2000. “Free labor: producing culture for the digital economy.” Social Text 18(2): 33–58.

Virno, Paolo. and Hardt, Michael. 1996. Radical Thought in Italy: A Potential Politics. Minneapolis: University of Minnesota Press.

Wissinger, Elizabeth. 2007. “Modelling a Way of Life: Immaterial and Affective Labour in the Fashion Modelling Industry.” Ephemera 7 (1): 250-269.

Witkowski, E. 2011. “Following Ms_Fabulous: Women, live-streaming, and do-it-yourself visibility in e-sports.” Presented at the annual meeting for the Digital Games Research Association, Hilversum, The Netherlands, September. http://homes.lmc.gatech.edu/~cpearce3/DiGRA13/papers/paper_49.pdf

Author image
London Website