Uses and gratification (U&G) theory argues that people use media for a certain motive out of cognitive and affective needs and to obtain a particular satisfaction (Katz et al., 1973; Ruggiero, 2000). The satisfaction audiences seek refers to their expectation from the media including information, fun, and other elements. It assumes that the audiences are rational and that they are expecting a certain function or use from the specific media (Wang et al., 2008).

U&G has been used in a wide range of different social media contexts. Ruggiero (2000) contends that the rise of computer-mediated communication has brought uses and gratifications back into sharp focus and can be used to predict the future of mass communication theory. Video game streaming is a completely suitable field for the U&G perspective to be employed. In a global context, Hamari and Sjöblom (2017) led a study to test the foundational motivations of spectatorship by quantifiably measuring gratification from potential aspects including social interaction, learning, and entertainment as shown in former studies (Kaytoue et al., 2012; Hamilton et al., 2014). Hamari showed the impact of five classifications including “Cognitive, Affective, Personal Integrative, Social Integrative, and Tension Release” summarized by West et al. (2010).

The initial studies disclose motivations for the Twitch viewing experience. U&G can be employed to work on questions about viewers and live game streaming as it captures the essence of the relationship between viewers and platforms.

With changing communication paradigms and diverse modes of interaction, U&G’s categorization is slightly crude. The path of addressing social networking site usage like Facebook shows requirement of refining. Smock et al. (2011) find that Facebook can be viewed as a toolkit that provides various functions. The usage patterns among different individuals are not uniform. Similarly, the approach of this research to Huya reflects this understanding. The live streaming platform is considered to contribute to content democratization (Pires and Simon, 2015), rather than one-dimensional content delivery. Therefore, when considering Huya, it seems very important to connect specific types of content with use motivations. In real media environments, users will choose to subscribe and follow streamers that can satisfy their expectations most; likewise, Facebook users select functions when they wish to perform different activities and YouTube viewers watch different channels for entertainment or getting knowledge. The question is what attributes make the streamers special and why? However, among all the research on the motivations, most of them start from the point of the streaming platforms. For practical content creators and distributors, there is a lack of understanding of how differences in streamers can influence viewers’ engagement.

More attention should be given to granular aspects of media practice to form a sophisticated understanding. Studies have now found the influence of preferences in more detailed studies, which also provided the basis for the analyses in this study. The social attributes of live gaming platforms are the first to be noticed. When focusing on the community of the platform, friendliness and other attitudes of streamers are mentioned to maintain a healthy relationship between viewers and the platform (Hamilton et al., 2014). Streamers come into the research field as they connect the users and platforms. Building U&G on a socio-based context, Hilvert-Bruce broadened the scope of motivations, adding community-related factors such as social interactions, social support, and a sense of community (Hilvert-Bruce et al., 2018). By capturing psychological and behavioral engagements on four indicators including the amount of time and money spent, their work investigated the effect of social motivations on user engagement. Moreover, in order to analyze the characteristics of streamers, the features and outcomes they represent need to be added before the attributes of streamers are linked to abstract motivations and values. Shen (2021) applied laddering interviews and MEC in her study to construct spectators’ value structures, identifying humor, streamers’ skills, physical appearance, and other attributes to connect U&G values. These also became an important source of theoretical framework for this study. From the early stage of U&G application in live gaming to the inclusion of social relationships in the study, many factors have been mentioned and refined.

Therefore, this research originally proposes a viewer motivation-streamer attribute analysis model (shown in Figure 1) where some common attributes of streamers are modified and matched with the motivations found in previous U&G as a theoretical framework to investigate the relationship between users, streamers, and platforms.

The research methods used in the study of audience engagement behavior in the field of live broadcasting are summarized in Table 1. From past experience, research methods are mainly divided into quantitative (empirical analysis) and qualitative (empirical materials). Online questionnaires and data crawling can respond to the research topic by summarizing broad trends from macro data. Interviews, on the other hand, can complement single-source data to develop a deeper understanding of the audiences’ personal experiences.

The discussion in this paper focuses on the data-based analysis of subscription behavior and the specific dimensions discussed subsequently will follow the framework mentioned above. Hamari’s and Shen’s methodological uses also inform the methodology of this study (individual live-streaming participation behavior), such as the use of scales to measure the relationship between motivation and behavior and the use of the MEC chain to decompose the specific performance of participation in live streaming. Also, the open-ended questions allowed participants to give responses other than the alternatives, reducing the influence of the researcher. The advantages of open-ended questions include the possibility of discovering spontaneous responses given by individuals, and biases from previous research can be corrected (Reja et al., 2003), which is in line with the new perspective of women who are the subject of this study. The addition of an open-ended question section provided guidance to refine the design.

Empirical research has been preferably employed in research on streaming audiences (Sjöblom and Hamari, 2017; Cabeza-Ramírez et al., 2020). Although the Chinese e-sport market is expected to reach 32.11 billion dollars in 2022 according to Chinese consultancy iResearch (Statista, 2022), domestic empirical data is still lacking. Questionnaires have been on the rise since 2017 as an important survey method for game streaming (Harpstead et al., 2019). With a five-point Likert scale, the importance and possibility of attributes affecting their subscription choice will be graded (where 1 = strongly disagree and 5 = strongly agree). Moreover, qualitative data in brief open text format provided an approach to develop knowledge of the detailed conditions and the reason for their grading.

Thus, quantitative and qualitative data were collected. The target population was Chinese female Huya users watching Honor of Kings (HOK) live stream frequently.

To avoid invalid responses and to find a pool of sufficient participants, the link was distributed through Weibo as well as a few forums and online communities dedicated to games. As a participatory incentive, we provided an opportunity to win a redemption voucher for a hero skin worth 88 yuan or 11 pounds. A lottery was held among all eligible members.

The target population is women who watch Honor of Kings (HOK) streaming on the Huya app frequently. For the purpose of enhancing the generalizability of future research, all volunteers are chosen at random. Replies from respondents who misreported their check question, “What is your gender?,” are filtered out as invalid. Additionally, an age limitation was set in the questionnaire. According to the Notice on Further Strictly Regulating and Effectively Preventing Online Video Gaming Addiction in Minors, which was published by The General Administration of Press and Publication of the People’s Republic of China (2021), teenagers are allowed to play video games for 1 h per day on holidays. That means people under 18 have limited experience with gaming and streaming. The responses for Question 1 equal to “under 18 years” are omitted. The final number of valid responses obtained in total is 202. Apart from the five gender-related and four age-related invalid data, the other missing responses reported are non-HOK streamers in Question 3 in watching habit questions.

Participants under 25 comprised 68.93% of all the valid data and only six respondents reported their age to be over 30. This is consistent with the Huya Annual Big Data Report in 2019 (Huya Inc., 2020), which shows that over 80% of the users were born after 1990. The groups working on or having obtained a bachelor’s degree are the majority (n = 127, 61.6%), indicating a relatively high education level. It is noteworthy that the ratio of income under 4000CNY per month (n = 97, 47%) is relatively high. This can be inferred that students make up approximately half of the sample. Meanwhile, the majority of the respondents spend over 7 h a week watching streams and have a habit of donating.

Table 2 displays the grading results of streamers’ attributes when female audiences decide whether to subscribe to their channels. As mentioned above, the attributes are based on existing research. The top three characteristics of streamers were rated for importance by the 202 respondents: speech style (Average = 4.51), quality of live broadcast (Average = 4.50), and humor (Average = 4.51). Table 3 presents the results of the derivative questions of speech style; of the respondents, 57.77% preferred a peaceful and rational style of speech instead of drama, while 64.08% could not accept the streamer swearing during the broadcast. As Hamilton et al. (2014) found in his research, streamers’ attitudes can influence the atmosphere in the community and their qualities may have an effect on the platform’s openness.

The characteristics of the streamers that ranked low in importance were excellent gaming skills (Average = 4.02), voice (Average = 3.97), and physical appearance (Average = 3.81). Table 4 shows the game skills that participants view as suitable. The response rate and popularity of two items, national heroes and all-around anchors, were significantly higher. This result is largely consistent with the ranking of gaming levels.

Basically, Table 5 illustrates that audiences usually watch streams at night, which is the same as most streamers’ working time.

Responding to pop-up comments, satisfying fans’ requests and lottery are favorite interactions (see Table 6). These kinds of engagement behaviors are commonly observed in streaming platforms and the function of pop-up comments and gift-giving has been identified in prior research (Chen and Lin, 2018).

A variance analysis was conducted to investigate how women with different demographic characteristics had different attribute preferences when subscribing to HOK streamers.

Table 7 illustrates that the difference in preference for the appearance of the streamers when subscribing to an e-sports anchor varied by age (F = 10.351, p = 0.000). The standard deviation analysis results are shown below. The mean scores of the groups with more significant differences in this factor were “26–30 years old >18–25 years old; 26–30 years old >31 years old and above,” which means that women aged 26–30 are more interested in the physical appearance of the hosts than women aged 18–25 and 31 and above.

Table 8 shows that there was no statistically significant difference in the preferences of women with different education levels (p > 0.05).

Our findings not only reveal the particular types of streamers that win popularity but also provide a deeper insight into the gratifications female audiences seek in terms of streamers’ personal performance and technological facilities when engaging. With regards to the female group, the attributes they voted for and the comparison with the male choices are also clearly presented in the results. The following is a detailed discussion of the theoretical issues.

Based on the empirical results and theoretical implications presented in the final section, the author would like to present a discussion and suggestions related to the practical implications of the preference of female viewers for broadcasters. Since women are experiencing growing participation in streaming viewing, this may enable streamers to further drive monetization of the platform while maintaining their own competitiveness and influence in the media market. The following recommendations will be developed from three perspectives: platforms, streamers, and female users.

From the platform’s perspective, as the female user base grows, it is important to have a reasonable balance of preferences for different genders. Johnson and Woodcock (2019) emphasized that “liveness” is a crucial feature of streaming platforms such as Twitch. Diversity of users can help platforms grow healthily and enhance their vitality. As we are in the era of surveillance capitalism, any user actions can be recorded as user-generated data, which will be recorded to make economic income for platforms (Saura et al., 2021). What should be emphasized is that the use of user data is moral and platforms should protect user privacy. As a streamer, understanding the preferences of female viewers can help them create an image that viewers love and create popular content. The following suggestions are given combined with the most important attributes found in the findings. In order to gain audience satisfaction, streamers should practice using appropriate language and giving meaningful opinions. They should be able to illustrate that they are at ease with the terminology, culture, and concepts associated with gaming to ensure that they can present themselves as genuine broadcasters and players (Duffy, 2017). Streamers are “super-individuals” who have been empowered in an unprecedented manner like stars. Streamers should behave themselves and have a positive influence on their fanbases, which can benefit their future branding activities. Having good looks and a good voice is not essential, but it is recommended that the streamers present a natural and clean appearance in front of the camera.

When female audiences repeatedly engage with their favorite streamers or subscribe and develop a periodic viewing habit, they are likely to generate a one-sided, hypothetical intimate relationship with media celebrities, which is defined as a parasocial relationship (PSR) (Rubin and McHugh, 1987). Scheduling meetings, communicating with streamers through comments, and receiving thanks for donations can all contribute to a unique interactive experience. The study’s rank correlation analysis confirms that the amount of time spent watching and donations made are positively correlated with interaction (see Table 10).

Based on the former discussion, it is recommended that female users should prevent themselves from over-indulging in and irrationally spending money on this kind of online intimacy.