CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

You Watch, You Give, and You Engage:

A Study of Live Streaming Practices in China
Zhicong Lu, Haijun Xia, Seongkook Heo, and Daniel Wigdor
University of Toronto, Ontario, Canada
{luzhc | haijunxia | seongkook | daniel}@dgp.toronto.edu

ABSTRACT popularity of Twitch.tv, live streaming in North America

Despite gaining traction in North America, live streaming has only recently begun to enter into mainstream culture,
has not reached the popularity it has in China, where live- through the advent of Facebook Live, YouTube Live, and
streaming has a tremendous impact on the social behaviors Periscope [7,18,33]. Although a few research projects have
of users. To better understand this socio-technological focused on understanding the live streaming phenomenon,
phenomenon, we conducted a mixed methods study of live they have been largely confined to understanding North
streaming practices in China. We present the results of an American usage [7,8,18,26,33]. The present work, however,
online survey of 527 live streaming users, focusing on their casts a lens on a country and culture that has almost
broadcasting or viewing practices and the experiences they ubiquitous live streaming usage: China.
find most engaging. We also interviewed 14 active users to
explore their motivations and experiences. Our data The continuous growth of live streaming in China has
revealed the different categories of content that was resulted in more than 200 million viewers watching
broadcasted and how varying aspects of this content streamers perform live each night on more than 200 live
engaged viewers. We also gained insight into the role streaming platforms, creating an estimated 5-billion-dollar
reward systems and fan group-chat play in engaging users, industry in 2017 [23]. Chinese live streams differ greatly in
while also finding evidence that both viewers and streamers content, style, and form compared to those in North
desire deeper channels and mechanisms for interaction in America and Europe. Previous studies of US and Canadian
addition to the commenting, gifting, and fan groups that are live streaming have found that live streaming was almost
available today. exclusively for live events [7,33] or sharing among close
friends [18]. In contrast, Chinese users utilize live
Author Keywords streaming for a wide array of uses, such as pan-
Live streaming; social media; user engagement; social entertainment (i.e. so-called “showroom performances” of
network. singing, dancing, music instruments, and talk shows hosted
ACM Classification Keywords and performed by individual streamers), e-commerce,
H.5.m. Information interfaces and presentation (e.g., HCI): personal knowledge sharing, and personal experience
Miscellaneous. sharing [42]. Although differences in live streaming
activities have been identified, there is little understanding
INTRODUCTION as to why these differences exist, what makes live
In recent years, the proliferation of mobile devices streaming activities so engaging and popular in China, and
equipped with high definition cameras and high-speed what we as the designers of live streaming platforms can
internet has led to a surge of individuals making live learn from their use.
streams. These live streams are supported by software that
enables anyone to share their experiences at live events, To better understand this socio-technological phenomenon,
hold weekly talk shows, and more. Twitch.tv, the popular we conducted a mixed methods exploration that included an
gameplay-based live streaming platform has more than 2.2 online survey (N = 527) and interviews (N = 14) with
million active streamers per month [8,35]. Despite the regular users in China. The survey queried the motivations
behind engaging with live streams and live streamers, in
Permission to make digital or hard copies of all or part of this work for personal addition to the ‘rewards’ viewers provide to streamers, and
or classroom use is granted without fee provided that copies are not made or other methods of interaction they have with streamers off
distributed for profit or commercial advantage and that copies bear this notice
and the full citation on the first page. Copyrights for components of this work the live streaming platforms. Our results revealed that
owned by others than the author(s) must be honored. Abstracting with credit is Chinese viewers are more interested and engaged in live
permitted. To copy otherwise, or republish, to post on servers or to redistribute streams about the personal experiences of strangers than
to lists, requires prior specific permission and/or a fee. Request permissions
from [email protected]. they were in the experiences of their friends, live events or
CHI 2018, April 21–26, 2018, Montreal, QC, Canada civic content, contrasted prior results on North American
© 2018 Copyright is held by the owner/author(s). Publication rights licensed to live streaming usage. We also found that reward-based
ACM.
ACM 978-1-4503-5620-6/18/04...$15.00. systems and fan groups on instant messaging apps further
DOI: https://doi.org/10.1145/3173574.3174040 afford social interactions between viewers and streamers

Paper 466 Page 1

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

and facilitate community building. We provide an overview platform. Sjoblom et al. [31] also studied how video game
of the opportunities and challenges of live streaming genres and content types influenced viewer gratification in
exposed by this research, to inform the design of future live streaming, and showed that archetypal structure of the
platforms and services that support social live streaming. stream is more important than genre of games being
streamed. Although relevant, the focus on gaming as a
BACKGROUND AND RELATED WORK
We first provide background of live streaming in China, shared experience makes this sort of live streaming a
then present related work regarding vloggers on YouTube, unique community. We focused our research efforts on the
video game live streaming, and general live streaming. entire set of live streaming uses.
General Live Streaming
Live Streaming in China
In China, live streaming began when the Internet users Live streaming has attracted increased attention in the HCI
started repurposing the public video chat room services of community due to the popularity of mobile live streaming
YY [44] for performative live streaming [2]. Most live applications. Juhlin et al. [12] explored early generations of
streaming platforms gradually began to provide mobile mobile live streaming apps Bambuser and Qik, finding that
versions of their live streaming services. As of April 2017, users could not find appropriate topics to stream and
there have been more than 200 live streaming platforms in streams suffered from poor image quality. Dougherty [4]
China [42], with 324 million live streaming users [5]. The explored the use of Qik, showing that 11% of videos had
top 9 popular platforms all have over 2 million daily active civic importance, such as journalistic and activist value, e.g.,
users [11]. Due to the popularity of live streaming, activist meetings. Tang et al. [33] conducted a mixed
‘streamer’ has become a new profession in China. methods study of early users of live streaming on Meerkat
Corporations such as Tencent provide facility support and and Periscope, finding that the broadcasted live streams
training for full time streamers who have the potential to consisted a diverse range of activities, and the motivations
become popular [37]. These professional streamers go live of the streamers were mostly for branding purposes.
almost every day and live streaming acts as a primary Other work explored the demographics of users on live
source of income for them. streaming services and what makes such stream engaging.
YouTube and Vloggers Scheibe et al. [28] analyzed information production and
YouTube allows for the sharing of user-generated video reception behavior on YouNow, revealing that YouNow
content, and provides a road to fame for many vloggers streams were mostly made by adolescents for adolescents.
who use YouTube to share their life and experiences. Lottridge et al. [18] studied the live streaming practices and
Wesch [38] analyzed the self-awareness generated by motivations of teens, highlighting that teens were streaming
vlogging, and provided evidence of context collapse [20] on like taking a long-form selfie, showing one’s self and life,
YouTube. The commenting system on YouTube was to engage with small group of friends. Work by Hu et al.
analyzed by Siersdorfer [30] and it was found that most [10] did attempt to understand why users kept watching on
comments were concentrated on a small portion of the most live streaming platforms, and found that self-identification
popular videos. Some research in HCI has also explored with the broadcaster and viewing group increased the
how to craft a conversational experience and promote social motivation to keep watching. Haimson et al. [7] studied
interactions through YouTube [9]. Although relevant, the what makes live events engaging by comparing Facebook
present work thus seeks to understand how real-time, Live and Snapchat Live Stories. They found that immersion,
synchronized interaction during large-scale live streaming immediacy, interaction, and sociality were important to the
influences viewer engagement and the social interactions engagement of watching live events.
between viewers and streamers. This existing research explored the different contexts where
Video Game Live Streaming (on Twitch.tv) live streaming is used, streamers’ behavior, and user
Hamilton et al. [8] found that live streaming communities engagement in North America and Europe, but little
form around shared identities from streams’ content and research has explored the practices of live streaming in
viewers’ shared experiences. Pellicone et al. [21] conducted China, where there are more daily active users [5], and
a qualitative study of an online forum about game more professional and regular streamers [11]. Work by
streaming and found that a key attribute of streaming is the Zhou [42] attempted to understand live streaming behaviors
development of a unique attitude and persona as a gamer. in China, revealing information about users’ demographics,
Sjoblom et al. [32] investigated why viewers watch others usage, and perceptions of live streaming, however, he did
play video games, and showed that tension release, social not explore what content was being viewed, how viewers
integration, affective motivations, and information seeking interact with streamers, and what factors contributed to
impacted the hours spent watching, subscription behavior, viewers’ engagement. To better inform the design of future
and the number of streamers watched. Similarly, Gross et al. live streaming interfaces, we seek to dig deeper into the
[6] investigated how motivations influence user’s phenomenon of lives streaming in China and understand
gratifications on watching live streaming on Twitch.tv, how the practices, engagement, and interactions in North
finding that time spent influences the money spent on the America differ from those in China.

Paper 466 Page 2

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

METHODS active for 2 weeks in July 2017. We received 902

Inspired by the prevalence and popularity of live streaming completed surveys, which we reviewed using manual
in China, we explored the following research questions: inspection and SoJump’s data cleaning tools. In total, 375
RQ1: What are the practices and motivations for watching responses were removed from the data set for one or more
and conducting live streams in China? How do they differ of the following reasons: not being a live streaming user
from those in North America? (257), failing our “trap” questions (12), time of completion
(55), patterned responses (36), or for providing fake or
RQ2: What type of interactions between viewers and gibberish text responses (15); this cleaning method
streamers take place? How are these interactions different followed [7]. This left 527 completed responses for
from those in North America? analysis. Each respondent who finished the questionnaire
received a cash payment. Our analysis of results was
RQ3: What mechanisms exist to reward streamers or show
conducted in Chinese, and translated to enable broader
one’s appreciation for their content? How do such
collaboration with the team and reporting in this paper.
mechanisms affect engagement with a streamer?
Data Collection: Interviews
RQ4: What are the factors contributing to a viewer’s
To understand live streaming practices and some findings
engagement [24] while watching live streams? How do they
from the survey data more deeply, we interviewed 14
differ from previous work? participants who participated in the survey and agreed to
Data Collection: Survey participate in a follow-up interview (Table 1). The
To explore these questions, we adapted the methods of interviews were conducted remotely using video calls
Rader et al. [27] and Baumer et al. [1], who used an online during August 2017. Each interview lasted about 50
survey-based methodology to collect stories about issues minutes and participants were provided with a 50 CNY
related to computer security and Facebook non-use, honorarium for their time. The interviews were semi-
respectively. We developed an online questionnaire in a structured, with questions about participants’ current
similar manner, using two types of questions. The experiences regarding live streaming applications.
questionnaire was first developed in English with the Interviews were conducted in Mandarin, audio-taped, and
research team, then translated into Chinese by one of the transcribed by the author who conducted the interviews.
native Chinese authors and validated by another. Years Has Frequency
ID Gender Age of Streamed of Use in Location
The first line of questioning probed the live streaming Use Themselves times/week
experience of viewers. These questions included mostly P1 F 21 3 Yes 7+ Southwest
yes/no, multiple choice, or 5 point Likert-style questions P2 M 29 0.5 No 3 Midwest
and focused on understanding whether the respondent is a P3 M 34 1 Yes 3 Southwest
viewer, a streamer, or both, how long s/he watches a live P4 F 27 2 Yes 2 Mideast
stream, which features of live streaming platforms s/he uses P5 M 33 0.5 No 4 Midwest
most often, what factors of streamers attract him/her to keep P6 M 28 1 No 3 Northwest
on watching them, and other similar questions. The items P7 M 39 2 No 2 Mideast
were adapted from existing survey study on live streaming P8 F 31 1.5 No 7+ Southeast
in China [42], as well as previous research on user P9 M 35 0.5+ Yes 3 Northeast
engagement [7] (e.g., entertainment, sociality, information P10 M 23 1 No 5 Southeast
and interactivity). English and Chinese versions of the P11 M 26 1.5 Yes 4 Mideast
questionnaire, along with complete anonymized response P12 M 32 1 No 3 Northwest
data, can be found in an appendix to this submission. P13 F 22 2 Yes 7+ Southwest
P14 F 24 1 No 4 Northeast
The second set of questions were open-ended and probed
live streaming experiences. Similar to Rader et al. [27] and Table 1. Summary of interview participants’ demographic
Baumer et al. [1], all respondents were asked to tell a story information and location (within a 9-sector grid of China).
about a live stream they watched in the past three weeks Analysis
that they thought was the most engaging, creative, or Responses to the open-ended questions were analyzed using
enjoyable. They were also encouraged to provide a link or a an open coding method [3]. The two native Chinese authors
screenshot of the stream they mentioned. We also asked coded the first 10% of responses and met to gain consensus
them to briefly describe interactions or features they wish on their codes. One author coded the remaining responses
were provided in current live streaming platforms. and met with the second coder to reach agreement. All the
Recruitment codes were then translated into English and were discussed
To avoid our results being biased by just one platform or by the broader research team using affinity diagramming to
certain groups of users, we used services from SoJump.com group and find emerging themes. The transcription of the
to recruit respondents, which is a firm specializing in interviews was analyzed in the same way as the open-ended
recruiting study participants in China. The survey was questions in the survey.

Paper 466 Page 3

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

We employed factor analysis to cluster sets of questions apps were all well represented in our respondents
about engagement, in order to find significant differences in responses, and most respondents (93%) used mobile
the patterns of the data. Factor loadings were derived versions of the live streaming services. Although most
through principal component decomposition [15] and then platforms have a variety of genres of streams, viewers
rotated using the varimax (orthogonal) rotation, following a chose to use different platforms for different content, for
similar approach as in [13,18]. example, “Douyu is more for games and performance while
Kuai Shou is for dramatic performance” (P7).
RESULTS
The results have been organized based on the themes that App Name # of Users % App Name # of Users %
emerged from the affinity diagramming exercise. We first Douyu.tv 287 54.46% Panda.tv 131 24.86%
present a demographic profile of the respondents and then YY Live 263 49.91% Yi Zhi Bo 104 19.73%
detail users’ motivations and general practices while using Momo 210 39.85% Long Zhu 88 16.70%
Inke 200 37.95% Qi Xiu (x.pps.tv) 81 15.37%
live streaming apps, what the reward system afforded, and Kuai Shou 198 37.57% Lai Feng 74 14.04%
the importance of fan groups on instant messaging Hua Jiao 188 35.67% Ha Ni (ihani.tv) 49 9.30%
applications, the characteristics engaging streams had in Hu Ya 157 29.79% Xian Dan 44 8.35%
common, and the factors contributing to engagement. We Table 2. Respondents’ usage of live streaming services.
present a fusion of the open-ended questions in the survey
and the interviews to explain the findings in greater detail. Respondents reported that they found live streams of
interest via content shared by friends (60%), via browsing
Profile of Respondents featured streams (59%), or via the leaderboard of streamers
The questionnaire took on average 20 minutes to complete (57%). This indicates that views could be skewed to the
(M=20.30, SD = 15.49), and was completed by 243 males most popular fraction of streamers who were featured on
and 282 females (Figure 1). Respondents reported being the platform or had lots of fans, similar to findings of [33].
full-time students, managers, developers, researchers,
teachers, professionals, and so on. The age and education The average time spent watching a single live stream was
level distributions align with previous market report [2] and 62.0 minutes (M=62.0, SD=33.6), compared to the average
research [42] about live streaming users in China. video length on YouTube (4 minutes and 20 seconds) [22].
Respondents also spent on average 7.2 hours (M=7.2,
SD=5.0) per week watching live streams, demonstrating
that most respondents were regular live streaming users.
Motivations for Watching
More than half (69%) of respondents reported that they
watch live streaming to relax, while 55% reported being
attracted by certain streamer. Killing time, making more
friends, communicating with others and sharing life,
sharing their point of view, finding a community, or
desiring to gain new knowledge were also reported as
important motivations for watching live streams.
The comments from interviewees also provided other
motivations. For example, some reported watching live
streaming because they felt lonely and wanted to seek
comfort and the feeling of company. Others sought out
experiences that were different from their own social life,
“We seldom have parties in China. Sometimes I really want
to escape from my close-tie or the circle of my fellows, and
live streaming provides a perfect place for that” (P1). Some
Figure 1. Distribution of age, highest level of education,
location, and experience with live streaming.
also mentioned motivations for seeking specific information,
“Before I went to Hong Kong for a business trip, I
Of all the respondents, 363 only watched others’ streams intentionally watched a stream from a local streamer for
but did not stream themselves, 14 streamed themselves but some advice and her own recommendations for food” (P14).
seldom watched others, 45 spend more time streaming
Activities and Topics
themselves than watching others, and 105 spend more time
With regards to the content that was most favored by
watching others than streaming themselves.
respondents, there were a variety of activities that viewers
Respondents reported using multiple live streaming had a high interest in watching (Figure 2). We now
platforms, with an average of 4 platforms per respondent elaborate our understanding by further analyzing the stories
(M=3.9, SD=3.1; Table 1). We note that similar to previous from the survey, and mainly focus on topics and activities
reports on live streaming in China [2,42], the most popular not explored in previous literature.

Paper 466 Page 4

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

inspire deeper conversations – much unlike the live and

Interesting life experience 63%
unpredictable nature of most North American streams.
Travelling 49%
Singing, dancing/instruments 45%
There were some conversational streams mentioned that
Gaming 31% were focused on streaming itself, i.e., the streamer shared
Science and technology 27% his/her own life experience, how s/he became a popular
Movies/TV shows 26%
Health and fitness 25% streamer, and maintained and got more fans. Viewers
Talking about social skills 21% appreciated the inspiration and positive attitudes conveyed
Career and future planning 20% by such chatting (S343), as it gave them opportunities to
Video-chatting with a guest 19%
Finance and investment 16% know more about streamer because they did not appear to
Chinese culture 14% be into streaming to gain popularity or fame, but enjoyed
Education and knowledge 13% the activity itself and were willing to share this joy and their
Variety shows 12%
Life abroad 6% process with others. In many cases, this stimulated one’s
Western culture 4% willingness to follow the streamer, as mentioned by P13.
Live events 2%
Others 2% Performance
Watching amateur or semi-professional spontaneous
Figure 2. Responses to “What kinds of live streams do you performances was also a very popular streaming activity to
enjoy watching the most (Choose up to 5)”. watch (i.e., 125 responses in total, with 85 “singing”
Sharing Personal Experiences responses, 19 “talk show” responses, 13 “dancing”
Unlike many North American streams, streams that were responses, and others included magic tricks, playing music
conversational in nature, whereby the conversation was instruments, and performing unique skills like xiangsheng
between a streamer and a group of strangers, were very (i.e., Chinese comedic crosstalk) and ventriloquism).
popular with respondents (131 responses in total). Sixty- Most streams in this category took place indoors, and were
eight of the responses were about streams where the performed by beautiful women or men who were amateurs.
streamer was a non-celebrity and shared their real-life story They had other professions, even high education level, but
or personal experiences. In other streams, conversations really liked performing for fun, to attract fans, or to
centered around performing or playing video games, but the “realizing early dreams in life” (S121). Respondents
emphasis was on engaging in the conversation with others reported appreciating the personality, taste, temper, and the
who were not one’s own friends. This desire to talk with humor of streamers, and enjoyed chatting with the streamer
others who are not part of one’s traditional social circle on topics about his/her personal live or current affairs.
differs from previous live streaming research [18]. “Although she is not a professional singer, her performance
Most streams that were told about focused on meaningful, is of high quality and makes me happy and relieved” (P4).
serious topics, such as dealing with the pressure from work Viewers also appreciated the live and unedited nature of
or family, work-life balance, career planning and these performances, and enjoyed the interactions with the
development, dealing with relationships, marriage, raising performer and other viewers, e.g., asking the streamer to
the next generation, or supporting one’s parents. Given the sing a specific song and getting a response in real time, or
pressure that most Chinese adults from 26 to 35 years old listening to the stories related to the songs and commenting
face from their career and their family’s expectations for on them. The improvisational nature of such streams has
marriage and descendants, live streaming seems to provide resulted in a new form of improvisational music, Han Mai
a medium through which such individuals can have a social (a form of rapping) [17]. “He once rewarded some of his
support system without disclosing too much information or fans by writing lyrics of a Han Mai with stories of his fans
involving those close to them. “It’s really beneficial to have and performing it, which moved a lot of his fans” (S127).
interactions with streamer and viewers. I can release the
pressure from my daily life, my work, my family, children, Knowledge Sharing
parents, and high pace of life of modern society, through Another popular set of topics for live streams were those
discussion with the streamer and viewers during the stream. relating to acquiring new knowledge (92 responses, among
Peers around me cannot fully understand my pain” (S354). which “Cooking” had 33 responses). Different from Ask
Me Anything style streams [33], where viewers ask
With streams that were more serious in nature, some questions about certain topic and the streamer answers
streamers announced the topic to be covered during the them, these streams were more like lectures or a training
stream in advance. As noted by S163, “The streamer sessions, where the streamer talked about the topic,
notified all the fans in advance that the topic would be showing the procedure or essential information using slides
marriage, which attracted a lot of fans. The streamer told or other media, and answered questions from viewers. Most
about experiences of her friends, with unique and thorough streamers who do this also sought to provide a series
perspectives. Feeling her emotion and the realness of the weekly of streams on certain topics, which made the
story, I was inspired by the stream greatly.” This gave streams more like a series of live, online courses.
viewers more time to think about the topic and potentially

Paper 466 Page 5

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

Such topics often included, formal knowledge (e.g., Unlike a number of North American streams, many
language learning, graduate school entrance exam respondents reported that game streamers sometimes also
preparation, college-level mathematics, psychology, finance stream other activities, such as singing performances or
and investment, Chinese traditional culture or techniques, outdoor activities, thereby raising their engagement with the
etc.) or informal knowledge (e.g., cooking, interview skills, stream. “[The streamer] regularly changes her content
pick up skills, antique evaluation, skin-care, personal since it would be boring for some fans if she always streams
health, baby-care, fitness, DIY etc.). Cooking was the same game. It’s impressive that she sings really well
especially popular, due to the oft complexity of Chinese and has a very good taste in music.” (P10) This echoed the
culinary skills, as noted by P6, “Since some Chinese recipes findings of live streaming among teens [18], but differs
don’t provide accurate instruction on amount of ingredients from those about video game live streaming.
to be used and accurate time of the cooking, it’s really
Outdoor and Travelling
helpful to watch the streamer cooking in real time, and even A new category that emerged in the responses was the
cook along with the streamer.” desire to see streamers travel and be outdoors (i.e., 67
The streamers usually share such knowledge in a relaxing responses in total, with 26 of them about outdoor activities,
and entertaining style rather than preaching. The viewers such as surviving in the wild, adventures, hunting, hiking,
appreciated this kind of streaming, because they could learn or climbing mountains). Viewers liked the thrilling feeling
something in a fun, interactive way (P3, P10). Viewers also and the novelty of such streams, since such streams took
appreciated the possibility for customizing practices of place in locations where viewers seldom go to, and they
certain knowledge enabled by live streaming. As noted by knew it was real and unedited because it was live. They
our respondents, “There’re different methods for keeping reported holding their breath for the streamer and worrying
health using Chinese medicine, but it’s hard to choose the about their safety, while trying to learn some outdoor skills,
appropriate methods and right medicine for me” (S501). as noted by S404, “I watched a stream about downhill
Live streaming enabled viewers to learn from the climbing at a 300-metre-high cliff, which was thrilling and
experience of the streamer and other viewers, and to ask exciting. The streamer told us about how to do downhill
customized questions and get them answered in real time. climbing and how to warm up through his detailed
demonstration. I learnt a lot from it”.
Videogaming
Similar to North American live streaming on services like Outdoor streams also provide lens to view society from the
Twitch.tv, watching others play video games was another streamer’s perspective, while without live streaming such
common, slightly less popular activity in China (i.e., 88 behavior may be socially awkward. Some streams involve
responses). Most games were the same as on Twitch.tv streamers streaming outdoors while interacting with
(e.g., League of Legend, The Sims), while some were very passers-by, who were in their natural settings and reacted
popular Chinese mobile games (e.g., Honor of Kings). authentically. Viewers liked being able to observe these
people and discuss the interactions that resulted. “It makes
The motivations for watching video-game play were similar
me realize how people from other places differ from me in
to North America: to learn game play skills and strategies,
behaviors, attitudes, and values, which helps me to
to meet other gamers, and to enjoy the game without
understand our society more thoroughly” (P11).
investing too much time playing and mastering it. These
motivations echoed previous research about the importance E-commerce
of forming communities among game play [8] and Another novel category of streaming that was popular came
cultivating the streamer’s own game play attitude [26]. from live streams that were selling (mainly on Taobao, the
Chinese largest e-commerce website). Although akin to
The characteristics of such streams that made them North American infomercials, this is a unique category
engaging related to the impressive skills and strategies of which is seldom seen in previous literature. Products
the gamer, the humorous way they interacted with viewers, ranging from books, clothes, health-care products, skin-care
and their temperament and positive attitude. One products, to even cars were demonstrated and shown in live
respondent recalled this story about the values of his streams. When asked if they would like to try the goods
favorite streamer. “[The streamer] once started a reward recommended by the streamer, respondents reported an
for fans, asking them to send emails to him about their average likelihood of 3.64 (M=3.64, SD = 0.77), which
wishes and he’ll fulfill some of it. Many fans sent emails, indicated that it was likely that they would try. This
but one of them impressed me a lot. He wanted to be a suggests that some viewers were influenced by the streams
volunteer teacher at the charity school which the streamer and may trust the streamer’s recommendations.
has donated to, and asked him for recommendation.
Although that school was in less-developed region and Unlike the traditional pre-taped infomercial, respondents
doesn’t provide decent pay, he would like to be a teacher appreciated the ability to view the product directly, ask the
there.” (S29). This story shows the strong tie that developed streamer to show different perspectives of the product, see
between the streamer and his fans, and the potential greater the process of the crafting of the goods, and ask detailed
social impact that streamers often have on their audience. questions about the product. P4, who is a frequent

Paper 466 Page 6

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

consumer on Taobao, also noted that live streaming “gives Although we did not focus our study on streamers, 164
me chance to get to know other consumers in the stream, respondents regularly broadcasted, and 6 of our interview
and I prefer to trust them rather than the streamer, as long participants had streamed, with 2 of them being regular
as they don’t seem to be shill”. Such e-commerce streaming streamers. When asked about motivations for broadcasting
highlights that live streaming is beginning to integrate with themselves, most reported that they wanted to share their
other mobile apps viewers use frequently, thus bringing life with others, show their talents, or make more friends.
additional opportunities and challenges for these services. It The need for branding, which is shown to be the main
also suggests a potential avenue for the next generation of motivation of streamers on Periscope and Meerkat [33],
infomercials in North America. was a small part of their motivation. Most reported seeking
to satisfy their social needs, “I stream mostly to share my
Miscellaneous
There were also other stream content types mentioned, such experience with others who share the same interest with me,
as “Exaggerating behaviors”, “Pets”, “Celebrity”, and and to get a sense of accompany when being alone” (P14).
“Competitions”. One “Exaggerating behaviors” that Gifting
emerged was about a streamer who live streamed On Chinese live streaming platforms, sending free gifts are
themselves gorging on food, “At first, I was curious about similar to “liking” a social media post [29] or “hearting” a
how she could eat so much. As she ate more and more, I stream [33], but with different visual designs and more
became nervous and scared, since I knew it was live and it opportunities to show one’s personality. Live streaming
looked so real. After she finished I would get a sense of platforms in China also offer the opportunity for further
satisfaction, feeling that eating was the happiest thing in the interaction between a viewer and a streamer, whereby a
world.” (P1). It was the surprise, nervous feelings, and viewer can purchase and send a virtual gift to a streamer
satisfaction that the viewers incurred that kept them during the live stream. When a viewer sends a virtual gift, a
interested in this exaggerating behavior. message is displayed to all viewers of the stream, thus
serving as a public broadcast of one’s admiration and
Interactions Between Viewers and Streamers
In terms of the interactive features that respondents use, appreciation for the streamer. As noted by P14 “the floating
commenting to express self-emotion, to communicate with and flashing gifts on the screen were just like fluorescent
the streamer or other viewers, or to send free gifts to the sticks of a concert, which echoed the streamer’s
streamer, were frequently reported (Figure 3). performance”. To encourage repeated giving, a leaderboard
of the top gift senders is visible to all viewers, and all
viewers are notified when a top gift sender joins the stream.
Streamers can receive a portion of the value of the gifts, and
the rest goes to the platform.
Surprisingly, many respondents reported sending paid gifts
(66%). This was commonly done because the viewers liked
the content, they wanted to show support for the streamer,
they wanted to reward the streamer, or they simply liked the
streamer. Gifting with fancy animations provided viewers
with a means of self-expression. P13 noted that “sometimes
Figure 3. Respondents’ reported use of live streaming features sending virtual gifts works as my special way of saying hi to
in the month preceding the survey. the streamer”. P3 noted, “Sometimes I send gifts just as a
creative Danmu to show my agreeance with the streamer or
Besides commenting, becoming a guest and video-chatting other viewers”. This usage is similar to the use of emoji and
with a streamer during the stream was an important way for stickers [43], which express rich and subtle emotions that
viewers to interact with the streamer. Although it is not cannot easily be conveyed by text alone.
used frequently, some interviewees reported that they liked
the direct and lasting communication channel and the Paid virtual gifts also allow streamers to wade through the
ability for collaboration during the stream that it afforded. large number of messages that may be sent to a streamer
Others found the guest sharing his/her own stories and more easily, because viewers who want to demand attention
performing together with the streamer “fun and fresh” (S91). can pay for virtual gifts. Some viewers sent paid gifts with
Being able to communicate directly can facilitate deeper the intention of catching the streamer’s attention to lower
conversations, as S513 noted, “during video-chatting with the chance of their comments being ignored by the streamer,
her I realized that we had some friends in common but we especially when asking specific questions or sending a
haven’t met in person. After that we became very good special request to the streamer. P3 noted that “Since I paid
friends in real life”. Participants did express some privacy for it, it would stay much longer than text comments. It’s
concerns about being a guest, “someone can identify me really important if you want to catch the attention of the
even if I just reveal my voice” (P10). streamer when there are over 10,000 concurrent viewers
sending comments”. In this way, they no longer passively

Paper 466 Page 7

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

watched the stream, but were active participants, Fan Groups

influencing the content of the stream more so than those Unlike North American live streams, where interaction is
who do not send gifts. For example, P13 noted that “I once typically restricted only to the platform or Twitter,
sent a gift to let the streamer sing a song I like best of her”. streamers also interact with viewers through other channels.
Viewers who sent gifts can also develop a further Instead of using the built-in community in live streaming
connection with the streamer, such as becoming the apps, or leveraging social networks like Weibo, most
moderator of the chatroom and getting a chance to have streamers build fan groups on instant messaging apps such
private contact with the streamer. One story told that “I as WeChat or QQ, which is an integral part of daily life in
really liked her performance and paid for a lot small gifts China [21,43]. The fan groups are group chats, with
for her. Then I became the ‘guardian’ of her chatroom and features to support multiple modalities of messaging
we became good friends afterwards” (S152). beyond text, including emoji, stickers, voice, image, video
messages or URLs to share other web pages. Such
Our data also revealed several problems with gifting, which functionality is currently missing in the commenting
may result in viewers not being willing to pay for virtual systems of live streaming apps. Streamers usually put the
gifts. One problem is that the threshold needed to satisfy a information of fan groups in the banner of their profile
streamer can be raised if s/he continually receives pages or show it directly on the screen during their streams,
expensive gifts. Such behavior, however, is visible to all so viewers can join the group by searching for the group
viewers, which makes some unwilling to pay for more gifts. chat ID or scanning the group QR code.
As noted by P2, “Now I don’t send gifts very often. When
someone had sent gifts worth of 1000 RMB, I think it One reason a streamer (P3) mentioned for not choosing the
doesn’t make sense to send gifts worth of 50 RMB to the built-in group chat function is that “There are a lot of
streamer. She won’t pay extra attention anyway.” Another spams or even trolls in the group. On the other hand, I use
problem is that the benefits of sending gifts do not WeChat and QQ more privately, and I check messages in
accumulate effectively, as noted by P13, “when I have sent WeChat or QQ more frequently than live streaming apps”.
several gifts to the streamer and developed friendship with As mentioned by the two other participants who regularly
him, sending more gifts doesn’t have more benefits to me”. stream, fan groups were used to send notifications about
Some streamers even treat gift-senders differently from going live, asking for feedback about certain streams, or
non-gift-senders, which most of our participants were aware asking the viewers about their interest. The fan groups are
of and said would impact their impression of a streamer. thought to be more suitable for these purposes, as noted by
In addition to providing gifts within a streaming platform, P7, “I normally don’t ask for feedback during my stream,
51% of respondents indicated that they had rewarded since there are too many people sending comments, but few
streamers through other external channels, such as WeChat would give constructive suggestions. People in my group
Red Packets [39], money transfer through AliPay (similar chat, on the other hand, care about me much more, and I
to PayPal), or by physically mailing the streamer gifts. P14 really appreciate the suggestions they gave me”.
noted, “I have ordered food online late at night for my Over half of the respondents (56.4%) had joined the fan
favorite streamer who was working hard on her streams”. groups of streamers. The most common reasons why they
This behavior demonstrates that as viewers and streamers joined were to communicate with streamers and other
develop a friendship, they begin to disclose more personal viewers, to gain more information about the streamer, or
information, such as online accounts or addresses. simply to make more friends in the group. As P14 noted, “I
Gifting is also not a one-way exchange. Several made a friend in the fan group of my favorite streamer. We
respondents revealed that sometimes streamers reward fans regularly chatted in the group and went along well, so one
during the stream. As noted by S193, “the streamer was day I sent a request to him to add him as a contact in
wearing a costume with a lot of red packets on it. We could WeChat and we became friends afterwards”.
get a chance to pick one packet if we correctly answered a Elements of an Engaging Stream
question about content in her previous streams. A lot of Based on our analysis of our data, several elements were
fans were participating”. Two participants (P3, P13) found to increase one’s engagement with a live stream.
mentioned that they had rewarded their fans in the group
chat by sending money through red packets in WeChat or Streamer Personality and Skills
QQ [39]. Interestingly, they did not find that group Many positive characteristics of streamers repeatedly
members were reluctant to open red packets, even though emerged in our data, including welcoming, good-tempered,
previous research has suggested that people are protective polite, honest, authentic, kind, empathetic, patient, grateful,
while getting involved in financial activities with strangers impartial, hard-working and positive. Viewers also liked
[39]. Some streamers also send gifts to each other during a good voice, humor, story-telling skills, inter-personal skills,
stream, as a way of social exchange and maintaining skills to handle awkwardness, and skills to deal with trolls
“guanxi”, the ties between individuals fostered through wisely. Viewer’s identification of these characteristics
exchanges of favors [40], on the platform. requires a deeper level of attention to the stream.

Paper 466 Page 8

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

Atmosphere as S382 noted, “At first I thought he was close to losing the
Many responses mentioned that the most engaging streams game (LoL), but suddenly he came back and won the game
have a good atmosphere that may come from including by his adept operation. I was so impressed and spent nearly
music of good taste, classy background decorations, light of the expense of food for half a month to send him gifts”.
the setting, the relaxing and pressure-free style of the
Factors Items (all in the form of “X is very Factor
streamer, and the polite behaviors and friendly comments of important to the engagement of the loading
other viewers. A good atmosphere may help viewers stay stream”)
watching longer and be more engaged in the stream. Content and Talents and personality of the streamer 0.661
Form of the Interesting or meaningful content 0.797
Novelty Stream Innovated form of live streaming 0.483
Viewers also seem to be more engaged in things they have Alpha = .64 Being able to access to previously 0.644
not experienced, places they have not been to, knowledge (inter-item unknown knowledge
correlation)
they do not have access to before, and unexpected results. Aesthetics The UI and visual design of the app 0.566
Therefore, they like to watch streams about traveling or Alpha = .39 Streamer’s appearance 0.813
outdoor activities, or streams with exaggerated behaviors. Communication Sending and reading comments in the 0.754
with Others stream
Authenticity Communicate with peer viewers 0.794
Viewers also like the authentic nature of live streaming. Alpha = .77 Keep in touch with the streamer 0.576
Although some planning of the stream is essential and Joining the streamer’s fan group 0.589
beneficial, viewers do not like the content and settings to be Emotional Being able to feel streamer’s emotion 0.463
too artificial, but prefer them to be “close to life”. As Reactions Improvisation and uncertainty of the 0.635
stream
mentioned by P5, who liked watching the live behind-the- Alpha = .69 Sending gifts to the streamer 0.766
scenes streams of a popular variety show on TV, live Leading in the leaderboard of gift-senders 0.736
streaming “provides a chance for me to know something
Table 3. Factor Analysis with 4 factors derived
that would be cut on the final version of the show on TV”.
DISCUSSION AND IMPLICATIONS FOR DESIGN
Factor Analysis
We now reflect on our findings of how live streaming apps
Data from 14 questions of 5-point Likert scale related to
were used in China and compare our results with other
live streaming engagement was used to help find latent
research on live streaming. We provide insights to inform
factors that explain the variance found in the questionnaire
future research and design in the HCI community.
responses and the importance of these factors to the
engagement of viewers. Kaiser-Meyer Olkin measure of Our results echoed those found prior in that viewers liked
sampling adequacy is above 0.6 (KMO = .844), and the authentic, unedited nature of live streaming [33] and the
Bartlett’s test of sphericity was significant (χ2 (91) = community fostered from the streams [8]. The content is
1926.9, p < .05), which indicated that the sample was also less important than the way streamer performs it, as
factorable. The factor analysis yielded 4 factors with found in [31] that “content structure is king”. However, live
eigenvalues more than 1 (Table 3). Each item was included streaming practices in China differ in several ways.
within the factor with which it had the highest factor
One big difference is that streams about politics [7] or civic
loading, and we averaged items in each factor.
content [4] are rare in China. Although some streamers
“Content and Form of the Stream” (M=4.29, SD=0.53) was sometimes banter about current affairs, such content
found to be the most important factor for engagement in a constitutes a very small portion of their streams, and they
stream. Thus, the diverse content streamed on Chinese live do so more for gaining attention than for informing viewers
streaming platforms are the key to their popularity. of civic content. We conjecture that censorship is one
Different people can be attracted to streamers with different reason, as found in [21] that people in China seldom post
talents and personalities, diverse content, novel forms of about politics to “avoid placing themselves in potentially
performance, or previously unknown knowledge. difficult situations”. All live streaming platforms in China
require users to verify their identity before going live or
“Aesthetics” (M=3.84, SD=0.63) indicates the importance
commenting, which may have a dampening effect. Another
of visual elements, such as the UI, the appearance of the
reason may be that young Chinese people tended not to be
streamers, and even decorations of the background. This is
interested in democratic political systems [19,41].
also related to “Atmosphere” in the Engagements section.
Interviewees also noted that WeChat or Weibo are used
“Communication with Others” (M=3.66, SD = 0.70) is also more for following news, while live streaming is more for
thought to be important, highlighting the importance of pure entertainment or learning.
social interactions with others during live streams (through
The interviewees perceived little negative content, though
commenting) and beyond (through fan groups).
reported trolls and inappropriate content being streamed on
“Emotional Reactions” (M=3.53, SD=0.68) is an interesting public channels. They noted that trolls are often blocked by
factor, since its composition indicates that gift sending can the streamers or their moderators, and inappropriate content
be quite emotional. Several stories in our data echoed this, is often cut off by the official platform regulators. They also

Paper 466 Page 9

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

viewed regulations and censorship as positive forces to Based on our results, we now provide insights about how to
reduce content that may be undesirable or negative. design for better social live streaming experiences.
We found motivations for using live streaming aligned with Supporting Deeper and Richer Interactions
the “hot and noisy” principle (“renao”) that Chinese people The practice of leveraging fan groups for casual
use social media for “achieving a harmonious ambience of communications within the community indicates that
social life” [36]. People use live streaming for more viewers desire deeper and richer interactions with other
pleasure, more chances for socializing with others, and a viewers and streamers. Besides multi-modal messages,
feeling of being trusted and accompanied by others. The some participants even desire to convey the feeling of
interviewees also mentioned having added streamers or applauding or remotely hugging the streamer, or being able
peer viewers as their contacts on QQ or WeChat who were to feel, taste or smell objects in the stream. Solutions to
strangers before, showing the willingness to befriend and how to make such interactions less intrusive and
communicate with strangers online [36]. Live streaming interruptive, and how to control trolls in the form of new
“provides a perfect place for befriending strangers in a modality, will be challenging to create. For example, future
socially acceptable way” (P1). work can explore how to design customizable multi-modal
gifts to support richer social interactions in live streaming.
We also noticed different attitudes towards watching
friends’ streams. Previous work shows that when viewing Live Streaming within Non-Entertainment Contexts
live events, people interact more on friends’ live streams Our results showed that live streaming has penetrated other
than strangers’ [7,18]. However, when asked about services, such as e-commerce and online education. As the
watching friends’ streams, some participants said they current design of live streaming applications is to support
preferred to watch more popular streamers who were entertainment it remains to be seen how to best redesign
strangers, especially if their friends just showed their daily live streaming apps to support the unique challenges of
life, although they would like to support and trust their these diverse experiences while still maintaining a sense of
friends. The reasons they mentioned were their needs to community, reward, and authenticity. For example, for
keep a distance and respect their privacy, familiarity and educational streams, it suffers from low efficiency and
potential bore of the streams, having other channels for requires intense preparation for streamers. Future work can
better interactions, and awkwardness that might incur. explore how to make it more efficient, and how to design
tools to engage online learners and the crowd and support
Sending virtual gifts and forming fan groups may be rooted collaborative preparation of live informational content.
in some cultural differences. Previous research has shown
that Asian users prefer multi-party chat, audio-video chat, Leveraging Live Streaming for Social Good
and IM emoticons [14], which may be why they use private We found that some streamers have already begun to use
chatroom services that support multi-modal messaging for live streaming to showcase traditional cultural art forms and
community building rather than Facebook or Twitter-like artifacts. We envision this as a chance to preserve
public social networks. Chinese people think highly of endangered cultures by exposing them to a broader
guanxi [40] in social interactions, which may be why they audience to raise cultural awareness. Knowledge and
reward streamers with virtual gifts. They may perceive the experience sharing streams are also beginning to attract
rewarding process not only as a consumer behavior, which older adults to watch and engage in live streaming. We
can be an impulse purchase, but also a social interaction for should thus design better social interactions for these older
circulating guanxi and keeping ‘face’ [40]. Virtual gifts also adults to feel less excluded from live streaming, and design
act as channels supporting social dialogues [16], ‘currency’ community-based mechanisms that would engage them.
of social capital [40], and souvenirs [34]. CONCLUSION
Our findings echo previous research about live streaming as We empirically examined user’s live streaming practices in
a virtual third place [8], or informal public spaces where China to better understand such social-technological
people engage in socializing with others and form phenomenon, which has a large and ever-growing market,
communities [25]. We also show that fan groups are and different content types, compared to those in North
another third place. Live streams are open to all and public, America. By understanding the motivations, practices,
with synchronized communication, and to some extent, social interactions that occur beyond the streams, and the
ephemeral messages (i.e., once the stream is over, the chat important factors of engagement, we identified that the
cannot go on in context and hard to resume), while fan unique challenges and viewing behaviors encountered with
groups are closed and private, with both synchronized and Chinese live streams can be used as blueprint for the future
asynchronized communication, permanent messages, and of North American and European live streaming platforms
multi-modal channels, which complement the nature of live and services. Our work also outlined the importance of
streaming. Viewers keep engaging in these additional third expanding thinking about the current demographics of
places and seem to build more trust from communication users, live streaming within educational and finance
with each other, e.g., they begin to disclose personal applications, and the needs to support deeper and richer
information or even meet up in person offline. interactions between viewers and streamers.

Paper 466 Page 10

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

REFERENCES audiences choose to keep watching on live video

1. Eric P S Baumer, Phil Adams, Vera D Khovanskaya, et streaming platforms? An explanation of dual
al. 2013. Limiting, Leaving, and (Re)Lapsing: An identification framework. Computers in Human
Exploration of Facebook Non-use Practices and Behavior 75, Supplement C: 594–606.
Experiences. In Proceedings of the SIGCHI http://doi.org/https://doi.org/10.1016/j.chb.2017.06.006
Conference on Human Factors in Computing Systems
11. Iresearch Inc. 2017. Chinese Pan-entertainment Live
(CHI ’13), 3257–3266.
Streaming Platform Development Research Report,
http://doi.org/10.1145/2470654.2466446
2017. Retrieved from
2. Ting Cao and Ruichao Ren. 2016. Live Streaming http://report.iresearch.cn/report/201703/2962.shtml
Industry Research Report (in Chinese). Retrieved from
12. Oskar Juhlin, Arvid Engström, and Erika Reponen.
http://www.imxdata.com/archives/10426
2010. Mobile Broadcasting: The Whats and Hows of
3. Juliet Corbin and Anselm Strauss. 2008. Basics of Live Video As a Social Medium. In Proceedings of the
qualitative research: Techniques and procedures for 12th International Conference on Human Computer
developing grounded theory. Interaction with Mobile Devices and Services
4. Audubon Dougherty. 2011. Live-streaming Mobile (MobileHCI ’10), 35–44.
Video: Production As Civic Engagement. In http://doi.org/10.1145/1851600.1851610
Proceedings of the 13th International Conference on 13. Yumi Jung, Rebecca Gray, Cliff Lampe, and Nicole
Human Computer Interaction with Mobile Devices and Ellison. 2013. Favors from Facebook Friends:
Services (MobileHCI ’11), 425–434. Unpacking Dimensions of Social Capital. In
http://doi.org/10.1145/2037373.2037437 Proceedings of the SIGCHI Conference on Human
5. Rebecca Fannin. 2017. Livestreaming Catches On Factors in Computing Systems (CHI ’13), 11–20.
Much Faster In China Than U.S. Retrieved August 30, http://doi.org/10.1145/2470654.2470657
2017 from 14. Shipra Kayan, Susan R Fussell, and Leslie D Setlock.
https://www.forbes.com/sites/rebeccafannin/2017/05/0 2006. Cultural Differences in the Use of Instant
8/livestreaming-catches-on-much-faster-in-china-than- Messaging in Asia and North America. In Proceedings
u-s/#1d51270d4861 of the 2006 20th Anniversary Conference on Computer
6. Daniel Gros, Brigitta Wanner, Anna Hackenholt, Piotr Supported Cooperative Work (CSCW ’06), 525–528.
Zawadzki, and Kathrin Knautz. 2017. World of http://doi.org/10.1145/1180875.1180956
Streaming. Motivation and Gratification on Twitch. In 15. Jae-On Kim and Charles W Mueller. 1978. Factor
Social Computing and Social Media. Human Behavior: analysis: Statistical methods and practical issues. Sage.
9th International Conference, SCSM 2017, Held as Part
16. Hyosun Kwon, Boriana Koleva, Holger Schnädelbach,
of HCI International 2017, Vancouver, BC, Canada,
and Steve Benford. 2017. “It’s Not Yet A Gift”:
July 9-14, 2017, Proceedings, Part I, Gabriele
Understanding Digital Gifting. In Proceedings of the
Meiselwitz (ed.). Springer International Publishing,
2017 ACM Conference on Computer Supported
Cham, 44–57. http://doi.org/10.1007/978-3-319-58559-
Cooperative Work and Social Computing (CSCW ’17),
8_5
2372–2384. http://doi.org/10.1145/2998181.2998225
7. Oliver L Haimson and John C Tang. 2017. What
17. Jinglan Lin and Zhicong Lu. 2017. The Rise and
Makes Live Events Engaging on Facebook Live,
Proliferation of Live-Streaming in China: Insights and
Periscope, and Snapchat. In Proceedings of the 2017
Lessons. In HCI International 2017 -- Posters’
CHI Conference on Human Factors in Computing
Extended Abstracts: 19th International Conference,
Systems (CHI ’17), 48–60.
HCI International 2017, Vancouver, BC, Canada, July
http://doi.org/10.1145/3025453.3025642
9--14, 2017, Proceedings, Part II, Constantine
8. William A Hamilton, Oliver Garretson, and Andruid Stephanidis (ed.). Springer International Publishing,
Kerne. 2014. Streaming on Twitch: Fostering Cham, 632–637. http://doi.org/10.1007/978-3-319-
Participatory Communities of Play Within Live Mixed 58753-0_89
Media. In Proceedings of the SIGCHI Conference on
18. Danielle Lottridge, Frank Bentley, Matt Wheeler, et al.
Human Factors in Computing Systems (CHI ’14),
2017. Third-wave Livestreaming: Teens’ Long Form
1315–1324. http://doi.org/10.1145/2556288.2557048
Selfie. In Proceedings of the 19th International
9. Dave Harley and Geraldine Fitzpatrick. 2009. Creating Conference on Human-Computer Interaction with
a conversational context through video blogging: A Mobile Devices and Services (MobileHCI ’17), 20:1--
case study of Geriatric1927. Computers in Human 20:12. http://doi.org/10.1145/3098279.3098540
Behavior 25, 3: 679–689.
19. Damien Ma. Young Chinese People May Just Not Be
http://doi.org/https://doi.org/10.1016/j.chb.2008.08.011
That Into Western-Style Democracy. Retrieved June
10. Mu Hu, Mingli Zhang, and Yu Wang. 2017. Why do 20, 2017 from

Paper 466 Page 11

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

https://www.theatlantic.com/china/archive/2013/07/yo (WWW ’10), 891–900.

ung-chinese-people-may-just-not-be-that-into-western- http://doi.org/10.1145/1772690.1772781
style-democracy/277885/ 31. Max Sjblom, Maria Trhnen, Juho Hamari, and Joseph
20. Alice E Marwick and danah boyd. 2010. I tweet Macey. 2017. Content Structure is King. Comput.
honestly, I tweet passionately: Twitter users, context Hum. Behav. 73, C: 161–171.
collapse, and the imagined audience. New Media & http://doi.org/10.1016/j.chb.2017.03.036
Society 13, 1: 114–133. 32. Max Sjöblom and Juho Hamari. 2017. Why do people
http://doi.org/10.1177/1461444810365313 watch others play video games? An empirical study on
21. Daniel Miller, Elisabetta Costa, Nell Haynes, et al. the motivations of Twitch users. Computers in Human
2016. How the world changed social media. UCL Behavior 75: 985–996.
press. http://doi.org/http://dx.doi.org/10.1016/j.chb.2016.10.0
22. MiniMatters. The Best Video Length for Different 19
Videos on YouTube. Retrieved August 30, 2017 from 33. John C Tang, Gina Venolia, and Kori M Inkpen. 2016.
https://www.minimatters.com/youtube-best-video- Meerkat and Periscope : I Stream , You Stream , Apps
length/ Stream for Live Streams. In Proceedings of the 2016
23. Ben Moshinsky. Chinese millennials have created a $5 CHI Conference on Human Factors in Computing
billion industry in their search for 15 minutes of fame. Systems (CHI ’16), 4770–4780.
Retrieved August 1, 2017 from http://doi.org/10.1145/2858036.2858374
http://www.businessinsider.com/credit-suisse-note-on- 34. Alex S Taylor and Richard Harper. 2003. The Gift of
chinese-livestreaming-industry-2016-9 the Gab?: A Design Oriented Sociology of Young
24. Heather L O’Brien and Elaine G Toms. 2010. The People’s Use of Mobiles. Computer Supported
development and evaluation of a survey to measure Cooperative Work (CSCW) 12, 3: 267–296.
user engagement. Journal of the Association for http://doi.org/10.1023/A:1025091532662
Information Science and Technology 61, 1: 50–69. 35. Twitch.tv. 2016. Twitch.tv - About. Retrieved August
25. Ray Oldenburg. 1999. The great good place: Cafes, 30, 2017 from https://www.twitch.tv/p/about/
coffee shops, bookstores, bars, hair salons, and other 36. Xinyuan Wang. 2016. Social media in industrial China.
hangouts at the heart of a community. Da Capo Press. UCL Press.
26. Anthony J Pellicone and June Ahn. 2017. The Game of 37. Chris Weller. 2017. Chinese women are creating a
Performing Play: Understanding Streaming As Cultural billion-dollar live streaming industry based on singing
Production. In Proceedings of the 2017 CHI and slurping soup. Retrieved June 30, 2017 from
Conference on Human Factors in Computing Systems http://www.businessinsider.com/chinese-women-live-
(CHI ’17), 4863–4874. streaming-industry-2017-4
http://doi.org/10.1145/3025453.3025854 38. Michael Wesch. Youtube and you: Experiences of self-
27. Emilee Rader, Rick Wash, and Brandon Brooks. 2012. awareness in the context collapse of the recording
Stories As Informal Lessons About Security. In webcam. Explorations in Media Ecology 8, 2: 19–34.
Proceedings of the Eighth Symposium on Usable 39. Ziming Wu and Xiaojuan Ma. 2017. Money As a
Privacy and Security (SOUPS ’12), 6:1--6:17. Social Currency to Manage Group Dynamics: Red
http://doi.org/10.1145/2335356.2335364 Packet Gifting in Chinese Online Communities. In
28. Katrin Scheibe, Kaja Fietkiewicz, and Wolfgang Stock. Proceedings of the 2017 CHI Conference Extended
2016. Information Behavior on Social Live Streaming Abstracts on Human Factors in Computing Systems
Services. Journal of Information Science Theory and (CHI EA ’17), 2240–2247.
Practice 4: 6–20. http://doi.org/10.1145/3027063.3053153
29. Lauren Scissors, Moira Burke, and Steven Wengrovitz. 40. Jiang Yang, Mark S Ackerman, and Lada A Adamic.
2016. What’s in a Like?: Attitudes and Behaviors 2011. Virtual Gifts and Guanxi: Supporting Social
Around Receiving Likes on Facebook. In Proceedings Exchange in a Chinese Online Community. In
of the 19th ACM Conference on Computer-Supported Proceedings of the ACM 2011 Conference on
Cooperative Work & Social Computing (CSCW ’16), Computer Supported Cooperative Work (CSCW ’11),
1501–1510. http://doi.org/10.1145/2818048.2820066 45–54. http://doi.org/10.1145/1958824.1958832
30. Stefan Siersdorfer, Sergiu Chelaru, Wolfgang Nejdl, 41. Mingshu Zhang. 2013. zhongguoren xiangyao
and Jose San Pedro. 2010. How Useful Are Your shenmoyang de minzhu [What kind of democracy do
Comments?: Analyzing and Predicting Youtube Chinese people want]. Beijing: Social Science
Comments and Comment Ratings. In Proceedings of Documentation Publishing House.
the 19th International Conference on World Wide Web

Paper 466 Page 12

CHI 2018 Paper CHI 2018, April 21–26, 2018, Montréal, QC, Canada

42. Baohua Zhou. 2017. 谁在使用视频直播?——网络视

频直播用户的构成, 行为与评价分析 [Who’s using
Live Streaming? Live Streaming User constitution,
behavior and evaluation analysis ]. News Journalist, 3:
52–62. Retrieved from
http://www.cqvip.com/qk/96951x/201703/671502584.
html
43. Rui Zhou, Jasmine Hentschel, and Neha Kumar. 2017.
Goodbye Text, Hello Emoji: Mobile Communication
on WeChat in China. In Proceedings of the 2017 CHI
Conference on Human Factors in Computing Systems
(CHI ’17), 748–759.
http://doi.org/10.1145/3025453.3025800
44. YY 6.0. Retrieved August 30, 2017 from
http://www.yy.com/yy6/

Paper 466 Page 13