Borderline practices on Douyin/TikTok: Content transfer and algorithmic manipulation

Abstract

TikTok continues to be the top downloaded app in many countries around the world as the short video consumption craze continues. But TikTok has also come under harsh scrutiny for its Chinese origins and data security. For TikTok, the journey of globalization has involved a painful contest with governments, geopolitical manoeuvrers, and, ultimately, finding platform regulation loopholes. TikTok’s sister app, Douyin, shares identical digital architectures, but follows different trajectories of development in China. Through interviews with Chinese influencers and media practitioners, along with a content analysis of policy documents and industry reports, this paper identifies and analyzes the borderline practices that have occurred on Douyin – including content transfers, and algorithmic platformization – and evaluates the potential for these practices to be replicated on TikTok.

Introduction

In May 2022, German carmaker Audi was blamed for stealing content from a Douyin influencer with over 4 million followers named ‘Beida Mange’ in a video advertising campaign (Times, 2022). However, Chinese users later discovered that the ‘original’ content from ‘Beida Mange’ was in fact plagiarized content from the Chinese version of Twitter – Weibo. The event sparked widespread discussion of copyright infringement and content transfer on Chinese social media platforms, particularly on Douyin. Based on the rising issues of content transfer, fake accounts and algorithmic platformization identified on Douyin, this paper analyses industry reports alongside empirical findings, to reveal the possibility of its replication on TikTok. It employs platformization theory to investigate how these borderline practices have been platformized on short video platforms. This paper argues that such practices are enabled and to an extent, facilitated by TikTok’s globalization strategies.

Under the complex geopolitical, cultural, and economic context, the development of Douyin and its sister app TikTok involves various elements, among which data sovereignty, content moderation, and platform regulation became well-known issues during TikTok’s globalization. To identify potential regulatory dilemmas associated with short video platforms, particularly Douyin and TikTok, this paper examines problematic issues that have already been identified on Douyin, and explores the possibility of replication on TikTok, or even, onto other digital entertainment apps. It employs platformization theory to examine the platform affordances of borderline practices, their role in nurturing infringing practices, and how these generate bigger issues for platform companies at large. This paper draws on two industry-wide practices that are not exclusive to Douyin and TikTok – content transfer and algorithmic manipulation, to reveal new directions that require regulatory attention.

Borderline practices

To identify unexpected, and in most cases, unjustifiable uses of social media, particularly those occurring on Douyin and TikTok, this paper advances the concept of ‘borderline practices’ to illustrate infringing practices that cross physical and virtual boundaries. Borderline practices are ambiguous or infringing activities facilitated via algorithms and practiced by corporations, platforms or users, that are both legally and ethically questionable, having the potential of crossing visible and hidden boundaries (Su, forthcoming). Borders in this context, is a porous concept, referring to physical boundaries such as national borders, or virtual boundaries such as parameters of community standards, platform rules and guidelines, existing frameworks and constraints. Borderline practices, therefore, examine the repurposing of social media use under the periphery of legal and ethical accounts that could transcend national boundaries. It is a comprehensive term that covers problematic practices on social media platforms in general.

Within the broad scope of this field of research, this paper focuses on the investigation of the sister short video platforms – Douyin and TikTok. In this article, the major borderline practices identified on Douyin and TikTok include content transfer and algorithmic platformization. The broader definition of content transfer refers to content exchange between different apps, particularly Douyin and TikTok, where trending content can be borrowed, remixed, recreated and reposted on the same platform or another. A notable characteristic of such phenomena is that it is not limited to sister apps but has the potential to be replicated between different platforms. Content transfer has three layers: transferring content between platforms, replication of content or framework via companies, or individual behaviour. These practices may entail copyright issues, content moderation challenges, and platform governance concerns. Algorithmic manipulation refers to well-organized behaviours that attempt to trick recommender algorithms for certain purposes. For example, fake accounts can be seen as an embodiment of gaming algorithms, driven by financial incentives, it is often organized, and implemented for the purposes of livestreaming shopping or traffic diversion. These practices involve issues of community standards of livestreaming e-commerce, authenticity (of account and product), and privacy (security of personal information). This paper argues that these borderline practices are platformized and facilitated via platforms and enabled by social media industries at large.

Borderline practices are either orchestrated operations implemented by corporations such as MCNs (Multi-Channel Networks), a well-designed strategy executed by platforms, or user behaviours that reveal precarious nature of platform labour. It is crucial to examine these nascent infringing activities to understand grey areas of platform governance. One of the main facilitators of these practices are corporations like multi-channel networks (MCNs) – online talent agencies with more resources and means to game the industry (including rules, algorithms, and visibility. Originating in western contexts, Vonderau (2016) notes, ‘the notion of MCN refers back to the scale economies of multichannel operations in cable television’ (p. 367). However, different to traditional talent agencies, MCNs are supported by and incubated within the market infrastructure of Google, an extension of YouTube’s Content Managing System (CMS) and appeared around the mid-2000s. They were described as ‘networks of online content clustered under a brand mostly curating short-form non-television content’ (Boyle, 2018: 137). They were also referred to as ‘a natural evolution from YouTube’s Partner Program’ (Smehaug, 2016: 88), seen as content curators, assisting, and managing the content distributed on YouTube. The early stages of MCNs functioned as a significant mechanism for professionalizing amateur content (Cunningham et al., 2019; Zhao, 2016a, 2016b), serving as a by-product or facilitator in formalizing transformations for online content creators. This intermediaries or ‘middleman’ role is another aspect of MCN as a concept, which builds on existing media professions, and adds a digital layer onto it. This is by no means a new phenomenon, but a remix of existing media roles that dealt with new media professions (Lobato, 2016). The emergence and prevalence of MCNs was set against a backdrop of user-generated content proliferation – the age of content creators.

The YouTube MCN model experienced a frenzied growth between 2012 and 2014, attracting millions in investment capital and new acquisitions (Lobato, 2016). While this period of gold rush has passed, many regulatory dilemmas remain. Chinese MCNs in comparison, began to proliferate in 2015, with 160 companies established at the time, the number rose to over 5000 by 2018, and witnessed an exponential growth in 2019, with over 20,000 organizations in the field (Clarie, 2020). The ‘gold rush’ has found a new lease of life in China through the Wanghong economy, where it is suggested that e-commerce monetization has taken up 46% of gross revenue in 2020, surpassing advertisement revenue (43%) for the first time (Clarie, 2020).

The Chinese MCN emulates and yet differs as a counterpart to that of Google’s incubated MCN industry. Similarities can be found in company strategy, functionality, and participatory culture. Global uptake of MCN resulted from the rise of Web 2.0 and associated participatory culture (Jenkins, 2019), in addition, MCNs provide marketing, promotion, monetization and brand setting for a commission, and function similarly worldwide. As a for-profit commercial company, MCNs tend to focus on top players and promising creators, integrating resources for optimal outcomes. These strategies simulate most commercial entities, in prioritizing top products to generate the most income (Gardner and Lehnert, 2016). However, without affiliations to a dedicated platform, Chinese MCNs share more independence and are particularly flexible in terms of platform distribution. This contrasts the YouTube or Google incubated model, where MCNs share a long history with legacy social media, such as YouTube and Instagram. In addition, Chinese agencies thrive on the emerging short video and livestreaming field.

Such phenomena echo with the prevalence of livestreaming e-commerce in Western and Chinese markets, as Chinese livestreaming sales skyrocket over the years, the same business model is not as effective outside the Great Firewall. The key difference lies in ‘proliferation of scams, counterfeit goods, or ghost merchants’, which are common practices under the Chinese e-commerce market. The bolstering position of Chinese influencers, as well as the booming livestreaming economy speaks to re-established trust in commercial transactions (Buchet, 2021). As intermediaries that dangle between different actors – platforms, advertisers, professionals, creators, and Key Opinion Leaders (KOLs), MCNs pose threats in areas like content regulation and algorithmic manipulation. The idea of modifying content to appeal to algorithms is not foreign to MCN companies, but in China some companies are taking modification to the next level. This paper examines how borderline practices are platformized, how these institutions utilize platforms like TikTok and Douyin to achieve their financial gains, and the regulatory dilemmas associated with such phenomena.

Platformization of cultural production

This study employs platformization as the main theoretical framework, to emphasize how borderline practices are facilitated, or even encouraged by platforms, and how individual or organizational behaviours are influenced by them. While platformization demonstrates top-down processes of platform influence over other societal sectors, this paper engages with bottom-up borderline practices, shaped by platform ecologies.

Platformization refers to the systematic, infrastructural, and institutional changes enabled and constrained by platforms. It engages with the discourses of platform power or platform affordances, and explores the implications on social, cultural, and economic aspects. ‘[Platformization] is not a uniform process, but follows different trajectories along the vectors of infrastructure, governance, and practice’ (de Kloet et al., 2019), which demands detailed analysis on different localities and actors. This observation foregrounds the significance of political economy, noting that the process and consequences of platformization varies under different political systems, and therefore, it is a factor that needs to be taken into consideration. It also necessitates the investigation of platformization under the Chinese context, to explore international distinctions from western practices.

The theorization of platformization in the cultural industries also highlights the significance of infrastructures, governance, economic factors, and practice. According to Duffy (2020), platformization of cultural production incorporates the ‘penetration of economic, governmental, and infrastructural extensions of digital platforms into the web and app ecosystems, fundamentally affecting the operations of the cultural industries’ (p. 4276). These transformations within industry are ‘as much about changes as they are about continuities’ (Duffy et al., 2019: 2). These scholars noticed that creative workers are ‘hyper-attuned to algorithmic systems’ and proposed ‘algorithmic precarity’ to illustrate ‘the turbulence and flux that emerge as a routine feature of platformized labour’ (Duffy et al., 2019, 2021; Nieborg et al., 2019). Such discourses engage with a phenomenal concept known as ‘platform contingency’ or platform dependency.

One notable concept of platformization in the cultural industry is ‘contingency’, meaning that cultural productions are contingent or dependent on platforms, contributing to the notion of ‘contingent commodities’, whereby products and services can be revised, recreated, and redistributed according to platform affordances (Nieborg and Poell, 2019, p. 85 as cited by Su 2022). This contingent nature emphasizes the platform power in shaping and curating content, pointing at power distribution, and the implications for cultural productions, but the creative industries in China are also contingent on government policies. Censorship in China represents an institutional and infrastructural framework that regulates online and digital media spaces. It creates a stark contrast with democratic countries, where media content tends to be less censored, but rated and classified. These scholarly discussions have identified that platforms are ‘anything but neutral’ (Cunningham, 2018; Gillespie, 2010), referring to aggregated platform power, in its algorithmic capacity to dictate content creations, and also to circumstances where it is subsumed under political governance. It is ultimately a matter of control, the way it is controlling other industries, and how it is controlled by higher regimes:

The ‘platform’ metaphor, as Gillespie (2010) argues, obfuscates as much as it reveals. It obscures how social media and other digital services, labelled as platforms, not just facilitate socioeconomic, cultural and political interaction, but very much organize and steer this interaction. In other words: platforms are anything but neutral. As cultural production in general and media work in particular is becoming increasingly platform dependent, the autonomy and economic sustainability of particular forms of production is increasingly compromised (Nieborg and Poell, 2019: 86).

There is currently substantial work on platformization in cultural industries (see, i.e. Nieborg and Poell, 2018), issues of platform dependency (van Dijck et al., 2019; Bygstad and Hanseth, 2018) and the consequences of such phenomenon (Just, 2018; Van Dijck, 2020). Awareness on risks and abuses of algorithmic power has also been addressed under different circumstances, such as the algorithmic lore of YouTube, where experts on algorithms demonstrate tips for higher visibility (Bishop, 2020). Scholars like Duffy et al. (2019, p. 105) calls for detailed research on how ‘algorithmic precarity exacerbates the instability of cultural work in the platform era’, and there is little research on how platformization is facilitating the precarity of cultural work. This research looks at the flip side, whereas platformization examines the process of platforms over content, this paper investigates content over platforms, in other words how the processes of platformization can be countered and manipulated. By investigating three layers, economic, infrastructure and governance of the platformization of borderline practices, this research attempts to uncover how these operations have transformed in the platformized society, how they are adding to the precarity of creative labour, and regulatory precautions associated with such phenomena.

Methodology

This paper adopts a qualitative approach that combines ethnographic fieldwork with document analysis. Data were collected from scholarly research and industry reports and tests arguments via interviews. This study recruited 10 interviewees from relevant fields, including MCN employees, Douyin staff, middle tier Douyin creators, industry practitioners, and ByteDance staff. Viewpoints presented here are not definitive, but explanatory, and reflect real cases within the industry. All interviewees have specified their personal stance in the matter, that their viewpoints should not be understood as an official response of the company or industry, but merely their personal understanding of the field. All responses collected and presented in this study should not be re-used and reinterpreted out of the context.

Content transfer: Platform strategy, company tactic and individual behaviour

Platform ecology

Douyin and TikTok are twin platforms, or parallel platforms, that operate under different sectors of one company – ByteDance. They parallel in many levels but differed in governance, monetization, as well as cultural elements embedded in those digital infrastructures (Kaye et al., 2020). For parallel platforms, content transfer can be a common issue - trending content can be borrowed, remixed, recreated and reposted on another or the same platform. It is sometimes identified at a corporate level, and other times initiated by individual influencers. Content transfer, in a way, can be seen as a vital part of TikTok’s globalization strategy. ByteDance adopted the idea of ‘content pools’ to create and cater to well-established cultural flows (Jun, 2020), in terms of how US content has a greater influence over certain regions, and how Korean and Japanese content rose to greater popularity throughout Southeast Asia. Prior to each market launch, TikTok would import a certain amount of content from existing, popular markets that demonstrate greater cultural influence over regions, then deploy the content at targeted areas. This kind of content transfer is enabled by platforms, endorsed by companies, and constrained by censorship and moderation.

Such kind of content transfer is quite usual on Douyin, contributing to an unhealthy environment for creators at large. Many original creators have mocked Douyin in a well-recognized catchphrase: ‘original content will not go viral, but content transfer will do, copy is the essence of Douyin video’ (Biaowai et al., 2022). This phenomenon ties in closely with the platform’s nature or company culture that encourages user participation and interaction. When the algorithmic recommender system provides a simple solution to deal with homogenous content, impersonation, or fake accounts, not flagging transferred content can be seen as a business strategy, or an inevitable flaw.

Copyright enforcement is a challenging hurdle that concerns platforms or content creators, as they dance between the ‘party line’ and ‘bottom line’. For online influencers, some are constantly worrying about political standards or being shut down by both platforms and the state. The short video industry is under researched, under regulated, and is being closely watched by the Chinese state. In December 2021, Chinese authorities fined top influencer Viya an unprecedented AUD$210 million for tax evasion (Madhok, 2021), signalling tightening regulations on live-streaming influencers. These unexpected role models are believed to be in the crosshair of President Xi’s ‘common prosperity’ drive. As a result, an ‘ideological rigour’ was injected to the livestreaming economy, as evidenced when Viya issued a statement later to quit the industry indefinitely (Lily, 2021).

These precedents demonstrate the influence of state policies, as well as the contingent or dependent nature of the livestreaming economy. It is a sector that thrives or dwindles on these factors (Su, 2022). The crackdown on Viya signalled an industrial turn at the end of 2021, forecasting changes in the livestreaming industry: a moment to regulate and meet the Party line. Apart from tax issues, content moderation is also regulated, via measures like content monitoring and state censorship (Qiu 2021). Copyright infringement, an incubator of content transfer seems to be a minor problem under comparison, generating loopholes not properly regulated by relevant legislation.

Organizational or individual behaviour

Based on interviews with video editors at Douyin, it is evident that they would constantly scour TikTok for inspiration. Because of this, content transfer is nothing unusual on Douyin or TikTok; it is either supported by platforms or done by individual accounts, in a way that an MCN employee suggests ‘will not offend anyone anyway’ (MCN Employee, 2021). Content transfer is neither unique nor nascent. It is a common phenomenon that occurs on different platforms. Its insignificance lies in the cultural variety and context, and its successes are particularly difficult to replicate under different sets of rules or cultural norms. This sort of conduct, then, is interpreted more as a hobby than financially motivated:

It is difficult to say why we have transferred content. It is a wide-spread phenomenon, not just on Douyin, but on YouTube or other streaming platforms as well. It can happen between different platforms, or within the same platform. An interesting thing is, sometimes, content transfer can create a hit video, attract more hits than its original video, this means it is creating a context, one that fits in well with the video, hence the popularity. (MCN Employee, 2021)

The motives of content transfer can be roughly categorized into appreciation of content, lack of creativity and financial gains. This is a borderline practice that blurs copyright issues and content regulations. This kind of content transfer refers to copying of the framework, not the details, but definitions can vary under different circumstances. There are companies in China who specialize in content transfer from foreign platforms to the domestic market, or between domestic platforms. One interviewee who currently works at Douyin revealed that: ‘my first job at another company was to transfer content, between different platforms, but the funny part is, the company is still standing, and not bankrupted, which means that everyone in the industry do not really care about copying or transferring content’ (Douyin Video Editor, 2020). Similar views have been expressed by another staff member working at ByteDance: ‘content pool strategy is quite common. Meta or Instagram may have similar tactics - borrowing content from an existing market. However, since this happens at the platform level, copyright issues will be cleared first, to prevent illegal activities, so this process will be supervised by our legal team’ (ByteDance Staff Anonymous, 2022).

However, platform-endorsed content transfer does create an environment that encourages this borderline practice at an organizational or individual level. A common form of individual content transfer is remixing, where creators, often video DJs, combine a variety of original material in a different sequence, creating new meanings and trends in the process. Remix practices on short video platforms are still a subject of controversy and debate. On one hand, video DJs are worried about lawsuits, for their reuse, recreation of original content. But on the other hand, some argue that creating short video remixes, and therefore new sequences and meaning, is a form of creative labour that should be better protected by copyright laws. The process of proving originality and having to seek approval can damage or jeopardize the creativity. Some opponents view the underlying copyright issues in this instance as too restrictive or stifling for creative industries (Lessig, 2018).

A key evaluation on copyright infringement is financial distribution, whether monetary gain has taken place from this behaviour or not. Interestingly, individual content transfer between Douyin and TikTok is not as successful as expected, where differentiation between styles and users’ preferences can lead to a video trending on TikTok but with low views on Douyin. Content can be transferred from TikTok to Douyin, with added layers of censorship, but the process is faster from Douyin to TikTok. A media practitioner who was previously a scriptwriter at a top MCN company expressed his understanding towards content transfer, pointing directly to the issue of profitability:

I think content transfer is absolutely an infringement to copyright owners, without authorization, there is no room for debate on this. However, if I took a photo on the street, and I share it on social media, then I’m not breaking any laws if I’m not selling the photo for traffic. However, traffic equals money in the short video industry because people could make money by transferring other people’s content, and then make money. I maybe a bit extreme, but I think this should be banned at any level. (Script Writer at MCN Company, 2022).

Variations in audience appeal are attributable to cultural differences, ideological differences, or language barriers (though this can be resolved by automated translation software). While these factors are considered to be prominent, cultural differences mark a real challenge according to one former MCN employee:

Our ‘common sense’, made of cultural elements, are different. The same content may turn out to be a total failure on TikTok, due to differences in cultural connotations. You thought they would understand what you are insinuating, and later you find it awkward to realize that your audience did not resonate with your content. This would be surprising, as your ‘common sense’ does not align with your audiences. Also, it will be challenging to monetize content on TikTok as a Chinese creator, this is a topic that concerns MCNs’ monetization strategies. (Former MCN employee, 2021)

Algorithmic platformization and manipulation

A key infrastructure in the platform society is algorithm recommender systems. Algorithmic logic is considered to be ‘the central tenet of platformization’ (Bishop, 2020: 3). It is interpreted as ‘a codified process’ in promoting or restraining content visibility via platform infrastructure, or a recommendation system that is powered by computational abilities (Bishop, 2020: 2). Critics have commented on algorithmic transparency, as the inner workings of algorithms are hidden to users (Bishop, 2020; Meng, 2021; Watson and Conner, 2019); others focus on the contingent nature of cultural production on platforms, raising issues like platform dependency and data-driven productions (Bygstad and Hanseth, 2018; Lin and de Kloet, 2019; Nieborg and Poell, 2019). Recently, based on concerns over platformization and cultural contingency, Duffy (2020) proposed a closer look at algorithmic precarity, particularly the exacerbation on ‘the instability of cultural work in the platform era’ (p. 105).

It is within this context that this paper argues for a systematic analysis on algorithmic platformization. Unlike algorithmic manipulation, which emphasizes how creators or corporations attempt to game algorithms, algorithmic platformization underlines the ways in which algorithms influence content production. This concept touches on the three dimensions listed above, it refers to interdependence between content, algorithms, and creators, and how infrastructural processes impact on content production (in shaping and curating platform preferred materials), algorithmic optimization (how it is influencing and being influenced by content), and precarity (circumscribed creativity and insecurity). Algorithmic platformization emphasizes on the infrastructural algorithmic influence over cultural productions, particularly how the algorithm-powered or based platforms like Douyin/TikTok impact on content creations. For a company that employs algorithms as its competitive advantage, even during global expansion, ByteDance set itself apart from other social media platforms with its unique and powerful algorithm.

Content-based algorithm

Infrastructurally, the content recommendation algorithm of Douyin and TikTok is different from that of YouTube, Weibo, and other social media platforms, in that it is content centric, not community based. Other than the descending distribution model adopted by most social media companies (high visibility for top creators), online celebrities on Douyin/TikTok are treated as equal with entry-level creators. The uniqueness of Douyin and TikTok’s algorithm rests on its supposed valuation of content above anything else, which purports to be significantly different from the algorithm of YouTube and many other platforms, where accounts with higher visibility (more followers and consistently more views) receive more promotion (Brennan, 2020). A content-based algorithm is more about recommending the ‘best’ content, whereas the community-based recommendation is rather about sustaining and maintaining fan communities. The difference in algorithmic platformization has led to divergent business models. As YouTube has struggled over the years (Stokel-Walker, 2019), Douyin/TikTok have grown exponentially (Davis, 2019; Kaye et al., 2020).

The viewing system on Douyin and TikTok trains creators to constantly revisit and innovate their content for recommender algorithms, instead of relying on fan communities to promote or repost content regardless of quality. But it also suggests that platforms are using algorithms to reinforce their corporate values and incentivizing creators to comply. The content centric algorithm of Douyin and TikTok facilitates a competitive environment, pushing creators to constantly create, refine and innovate. Constrained by algorithms, optimization, and promotion, content creators may find themselves struggling with ‘circumscribed creativity’ (Kaye et al., 2020), creativity that has been noted, expected, adjusted and platformized. Content creators express a feeling of ‘love and hate’ against Douyin/TikTok, in the effort that is required of creators, but also the great opportunity that comes along with it. The most well-tuned cultural productions obtain the highest chances of monetization. As creator Teacher Lai indicated:

Douyin is a very scary platform. If I am a creator with 1 million followers, I enjoy the same treatment as a creator that has 5 followers. This means you have to start from the beginning, the algorithm does not care about the number of followers, but the quality of content itself. Everything goes into the traffic pool equally, if your content is not good enough, you cannot even get it to your followers. That is reality, and yes, it is cruel. (Teacher Lai, phone interview 2021)

Content-based algorithms equalize creators and contravene widely adopted platform revenue sharing systems. Creators on Douyin and TikTok simply cannot survive on traffic sharing – something the platforms do not offer. They must seek other ways to monetize content such as product endorsements, advertising partnerships, and livestreaming (Ma and Hu, 2021). The platform supersedes content creators and demands compliance, and yet, it retains its popularity through its secret and opaque algorithm (Ma and Hu, 2021). Aside from wild speculation or political oriented scepticism, ByteDance insists on the legitimacy of its data collection and usage. However, the company’s public claims and the secrecy of its algorithm continue to press concerns among its creators, as well as users. One Douyin creator shared his concerns with the algorithm:

‘I do not think I am exaggerating, but in a way, I do believe Douyin is monitoring their users, maybe not all the time, certainly when you are using the app, they want to detect everything. The aim is to optimize their recommendation system and make the best guess at your appetite. Remember, what they are offering is certain types of the video, not a certain creator, and that is crucial.’ (Douyin Anonymous Creator A, 2022)

Contract performers and gendered material

Algorithmic platformization underlines the role of algorithms in affecting infrastructural, governmental, and economical operations of cultural industries, and how it dictates, stipulates, and generates business strategies, monetization models, content creation/selection, creativity, and regulatory dilemmas. This vertical, top-down process can be overturned, or counter utilized against the algorithm itself, inadvertently helping borderline content, despite its lack of originality and creativity, to thrive and be replicated at a larger scale. MCNs incorporate means and tools to manipulate algorithms by reverse-engineering trending topics, copying styles, and transferring attention (Collie and Wilson-Barnao, 2020). Such practice is nothing illegal but jeopardizes user experience. Facing vicious competition, some MCNs are taking shortcuts, creating the easiest path to monetization by plundering each other’s originality. Contract performers emerged within context, with MCNs recruiting actors to perform scripted lines and scenes, copied from trending videos. As one online creator reveals:

Most influencers under MCN are contractors, not creators. MCNs normally hire editors, directors, and actors to play their parts based on their own script, which is created based on their analysis of the algorithms – what types of content draws attention, what kind of styles get promoted easily. They identify trending videos, come up with a different theme under the same framework, and then hire people to replicate the video. We call this ‘fake accounts’, in that it is not original. This behaviour is market oriented, and user oriented, but not interest oriented. (Douyin Anonymous Creator B, 2021)

Another interesting phenomenon is, as Craig and Cunningham (2019) has identified, that the content is gendered, with pretty girls on the stage and lonely boys behind the screens. One former employee of a MCN company disclosed that their main income is generated from livestreaming, not short videos. The well-known shortcut in the industry is to cultivate entry level influencers on Douyin, and instantly monetize content via gendered materials on livestreaming platforms. Douyin, as a short video platform, becomes the spotlight, drawing attention that can be diverted and utilized elsewhere:

The company that I used to work for is a special case, they foster influencers - normally pretty girls, and use them to collect rewards and gifts during livestreaming. It’s very easy for them to make money in this way, and that’s why the short video department was losing money, and the livestreaming department was making money. I am not sure about industry data, but from my experience of working in the industry over the years, short videos and livestreaming are making different kinds of money. So, we cannot simply conclude that the short video industry is not making money. (Anonymous MCN Former Employee, 2021)

MCN’s manipulation of algorithms can be divided into two categories: content transfer, and traffic diversion. Such instances are also evidenced by individual fake accounts that entail online fraud. For example, in October 2020, Douyin was found to have thousands of unverified fake accounts under one celebrity name, some of which were engaged in criminal or illegal activities. Analysts attributed this phenomenon to Douyin’s platform affordances, and a lack of differentiation between user accounts, noting that detecting fake accounts or impersonation of celebrity accounts should be a vital part of content moderation (Biaowai et al., 2022). However, this finding does not suggest that these practices are taking place regularly at an industry level, but rather occasionally. They are not significant enough to attract regulatory gaze yet not insignificant enough to be ignored.

Normally, users behave legally on these platforms, and MCNs would recruit famous influencers, focus on branding and integrating resources, while identifying smaller, promising influencers, also sharing company resources where available. These ‘edge ball’ practices conducted by individuals, or a few MCN companies (Flew et al., 2019; Su, 2019), are undoubtedly and unfortunately, drawing increasing attention. Since TikTok shares the same algorithm as Douyin, similar issues can also be identified such as content transfer, algorithm-induced homogeneous content, curated materials for promotional purposes, as well as fake accounts. These are common problems faced by not only TikTok, but other international platforms. In 2020, The Guardian found harmful pro-anorexic content being easily accessible and searchable on TikTok, which led to an investigation launched by TikTok, however, this issue persists among youngsters even in 2022 (Paul, 2021), accentuating a longstanding scepticism against algorithms and the call for transparency. The line between algorithmic manipulation and algorithmic platformization is blurring. If gaming algorithms from the creator’s side is seen as manipulation, then how is it possible to justify opaque content selection and promotion processes driven by algorithms, which is described as algorithmic platformization? Such dilemmas are important to consider when examining borderline practices.

Financial incentives behind borderline practices

Content transfer or algorithmic manipulation aim to draw attention and turn a profit. Despite various paths to commercial success, their financial motives are apparent in both public and private domains. Monetization in the public domain refers to traditional monetization in the public setting, such as brand deals, whereas the private sector refers to traffic diversions, bringing followers and viewers to secondary platforms, under a more private setting. Viewers will then be transferred to followers of the creator, a practice that emulates community-building, at a different space.

Accordingly, monetization in the private setting has become more prevalent. A common phenomenon is selling short video classes on Douyin, where for example, English teachers post common tips on learning English, turn their fans into students, then delivering paid lectures via other means: ‘my fans are now my students, and I’m thinking about quitting my job, to be a full-time Douyin influencer, because this makes a lot more money, and I’m working for myself, not anyone else’ (Douyin influencer, 2022).

Other than traditional forms of monetization, such as brand partnerships, virtual gifts, donations and so forth, a growing tendency on TikTok and Douyin is e-commercialization – monetizing content through e-commerce, with a built-in TikTok Shop, or third-party platforms. Despite the vastly different user bases in China and beyond, similar platform features on TikTok and Douyin encourage digital influencers to monetize content in increasingly similar ways including brand endorsement, traffic sharing, revenue, and live shopping. As such, content monetization is highly synchronized between the sister apps.

Conclusion

Traditional platformization studies focus on the platform power, especially the way platforms shape and dictate content creation processes. This study examines reverse processes, how platforms can be utilized for commercial purposes. Platformization of cultural productions are manifested in three layers: economic, infrastructure, and governance, and therefore, this paper uncovers how the system is being exploited on these fronts: algorithmic platformization, content transfer and monetization.

Findings suggest that borderline practices enabled by short video platforms are also a result of Douyin and TikTok globalization strategies. For example, TikTok deploys ‘content pools’ to transfer popular content to targeted regions, encouraging users to transfer content on a fundamental level. On top of this, it has created a culture that tolerates transferred content, where more content means more attention, and higher following growth. Additionally, TikTok shares the same algorithm as Douyin, generating issues of homogenous content, and algorithmic manipulation. The overreliance on algorithms, even during global expansions, is attributable to such phenomenon.

To summarize, this study explored: (1) algorithmic platformization and the way it has been counter-utilized by individuals as well as MCN companies, such as building fake accounts; (2) how users and MCNs utilize regulatory loopholes to transfer content; and (3) how platforms contribute to the synchronization of monetization models. Through the above analysis, future research should continue to explore the understudied corners of the emerging platform industry, and reflect on the digital culture, algorithmic power, consumerism, and capitalism.

Fake accounts and paid digital contractors could be seen as the by-product of algorithmic platformization. Their prominence demonstrates consumerism and capitalism and showcases the capitalist nature of the platform industry. Content transfer from foreign platforms like TikTok, are implemented with careful consideration on party policy, party line, and user preferences. In terms of dealing with these borderline practices, future studies should examine self-governance, or a similar framework to Facebook’s Oversight Board. Alternatively, relying on state regulation, or relevant institutions to oversee community standards can also be productive. The feasible method should be flexible, and adaptable, to incorporate variations and new challenges.