Digital Labor and Gender Gaps in Livestream E-commerce

Part of this article is based on my observation and work experience as an account assistant in a Chinese digital marketing and technology company. During my internship in this company, I was responsible for analyzing data about livestreaming performances extracted from Alibaba's data bank and communicating with the client brands who outsourced their digital marketing campaigns to our company. While I collaborated with Multi-Channel Network (MCN) agencies to optimize livestreaming, a part of the digital marketing campaigns, I also monitored live streams in real time to ensure its quality. Having closely witnessed and experienced the governance of algorithms in the livestreaming e-commerce industry, I realized the cruel fact that it is the labors, especially the live streamers that are ultimately governed and exploited by algorithms.

The exploitative nature of algorithmic governance of live streamers

The governing logic of algorithm is very close to the logic of the stimuli-organism-response (SOR) model. (Mehrabian and Russell 1974,  in Hu and Chaudhry 2019). The SOR model, therefore, can well account for the rationality of the exploitative nature of algorithms and also explains why brands, MCN agencies, and digital marketing agencies are forced to exploit streamers. The SOR model describes that various elements in the environment perform as stimuli to influence individuals' emotional states, namely the organism, which in turn leads to their behavioral responses (ibid.). Hu and Chaudhry (2019:1019-1041) applied the SOR model to livestreaming e-commerce, discovering that relational bonds by the streamers (S) influence viewer's affective commitment (O), which in turn affect viewer's engagement (R) and ultimately their behavior: consumption and social sharing. Consumers with strong affective commitment trust and are emotionally attached to the streamer (Evanschitzky, Iyer, Plassmann, Niessing & Meffert, 2006:1208). And viewers may extend the time and increase the frequency of engagement while watching a live stream, if they perceive the streamer to be attractive to them (Sjöblom and Hamari 2017:985-996). The attractiveness of streamer may come from their appearance, humor, and knowledge of products. In short, viewer's emotional state, engagement, and behavior influenced by streamer's performance are greatly valued because they are directly linked to financial profits. Therefore, streamer's attractiveness, relational bonds, and information quality together play a critical role in influencing the viewer's emotional states which have a direct effect on consumption and social sharing (Xu, Wu & Li, 2020:145-148).

Thus, when a streamer is live broadcasting, they are required to comply with various rules and are under multiple surveillance by both non-human and human stakeholders. Algorithms evaluate streamer's performance through data collection and processing to decide traffic flows to content visibility. The brand's marketing director, the MCN agency, and the digital marketing agency together acted as the human surveillants. The digital marketing agency that I worked for placed me to monitor the streamer's behavior in real time to ensure the quality of their performance. During livestreaming, streamers are supposed to wear make-up, not allowed to leave more than five minutes, and required to broadcast in an enthusiastic, vibrant, and friendly way. In addition to product introduction, streamers also have to respond to every comment of viewers and answer their queries, which demands a lot of extra energy. If incompliance with these requirements were found, I was responsible to communicate immediately with the MCN agency in our WeChat work group. Then, the MCN agency would remind the streamers to adjust their behaviors such as tone, gesture, and facial expression. Any sign of exhaustion, inappropriate attitude and language, or undesirable behavior of streamers is intolerable despite the prolonged time they have worked. In the livestreaming room, streamer's emotion is no longer a personal feeling but "an object that is managed, monitored, and manipulated" (Hochschild 1983:68-95).

Co-governed by algorithms and other stakeholders, streamers not only work under enormous pressure, but also work in unregular working time. Data extracted from Alibaba's data bank clearly shows that the longer the livestreaming duration, the more viewers and the higher transactions. Therefore, the brands, the digital marketing agency, and the MCN agency filled the schedule of streamers, especially during major shopping festivals such as 618, the mid-year shopping festival and Double Eleven, the Single's Day in China. A four-hour duration of livestreaming from 7pm to 11pm is common in working days. And the livestreaming duration could expand from 2pm to 11pm on weekends. Beyond the livestreaming time period, streamers are also required to spend time reflecting their performances and learning knowledges about products to optimize the livestreaming. This is reflective of Marxist theory of value based on labor time argues: the more time streamers spend live broadcasting, the more consumers they attract, and the more transactions they generate, "allowing for the accumulation of capital" (Zou. 2018:810). Few streamers can endure such pressure and working rhythm, so there are many a "new streamers appearing and old streamers disappearing from the platform every day" (Wang & Li, 2020:6).

The reinforced gender inequality and objectification of women

In addition to exploiting streamers, the algorithmic governance also reinforces gender inequality and helps to materialize and objectify women. Although Statista's data shows that around 67 percent of Taobao's live streaming users were female in March 2020, the small proportion of male users also play a significant role that indirectly contributes to the existing gender inequality. The product that outsourced digital marketing to our company was smart sound box and its main target customers are males. One male and two female streamers were selected to live broadcast the sound box for the brand. However, only the two female streamers were required to wear make-up and use the beautifying function of camera during livestreaming, even though they already look beautiful. One of the two ladies has a tanned skin which she herself is very satisfied with. However, the brand and my supervisor required her to turn on a high level of beautifying function because her skin looks swarthy and her face shape is not oval. The image algorithms could magically give the female streamers fair skin, perfect face shape, and good complexion. Both the human and non-human actors assumed that the female streamer's standardized beauty face is helpful to attract and retain the male viewers to enter and stay in the livestreaming room. Although no data has been tested to examine the authenticity of this assumption, it indeed treats women as an object to please man.

Another interesting observation is that most viewers would consult the male streamer more professional questions. On the other hand, some male fans frequently showed up to chat and flirt with the two female streamers without any purchasing behavior. Some offensive questions in regard with private life of the female streamers would occasionally appear in the live chat box. When encountering with such situation, the female streamers only could respond smartly and avoided disappointing the male viewers. Even if the brand's marketing director and my supervisor are females too and noticed this phenomenon in the livestreaming room, they chose to ignore such offensive behavior toward the two female streamers. Because more viewers could lead to higher visibility in the livestreaming webpage on Taobao, which will then attract more viewers and higher transaction according to the logic of algorithms. Wang and Li's study (2020:1-12) found out that female streamers have greater advantages than male streamers. However, this advantage may stem from the objectification of women. In the livestreaming room, women's face and body are nothing but functional objects (Zou. 2018:814). This objectification of women reinforced by algorithm will further devalue women's equal competence in fields traditionally dominated by man, making the course of gender equality even harder. 

References

Evanschitzky, H., Iyer, G.R., Plassmann, H., Niessing, J. & Meffert, H. 2006. The relative strength of affective commitment in securing loyalty in service relationships. Journal of Business Research. 59(12), pp.1207-1213.

Hochschild, Arlie. R. 1983. The Managed Heart: The Commercialization of Human Feeling. Berkeley, CA: University of California Press.

Hu, M.Y. & Chaudhry, S.S. 2019. Enhancing consumer engagement in e-commerce live streaming via relational bonds. Internet Research. 30(3), pp.1019-1041.

Sjöblom & Hamari. 2017. Why do people watch others play video games? An empirical study on the motivations of Twitch users. Computers in Human Behavior. 75(2017), pp. 985-996.

Wang M, Li D (2020) What motivates audience comments on live streaming platforms? PLoS ONE. 15(4), pp.1-12.

 Xu, Xiaoyu., Wu, Jen-Her. & Li, Qi. 2020. What drives consumer shopping behavior in live streaming commerce? Journal of Electronic Commerce Research. 21(3), pp. 144-167

Zou, S. 2018. Producing value out of the invaluable: A critical/cultural perspective on the live streaming industry in China. TripleC. 16(2), PP. 805-819.