Idols of Promotion and Authenticity on TikTok

TikTok’s rapid growth in the past few years, especially in the younger demographic, may signal a market shift. With chil‐ dren, teens, and young adults reportedly making up 40% to 60% of its user base, the platform is becoming the strongest challenger to YouTube, Facebook, and Instagram. The most followed TikTok celebrities are mostly young people who have either grown up with the platform or recently extended their popularity from other platforms to reach new audiences. This research investigates the discursive strategies and persona performances employed by the top 25 TikTok celebrities under the age of 25 in both popular content and content marked as advertising. A large sample of TikTok content metadata was collected using API interrogation. From each of the 25 young TikTok celebrities, up to 1,000 videos per user ( N = 22,650) are explored using quantitative approaches. Two subsamples are analysed using visual, rhetorical, and narrative analysis to evaluate the most popular content ( N p = 226) and content marked as advertising using the TikTok ad flagging ( N a = 213). The findings include the identification of seven persona performance types and a significant difference in terms of per‐ formed ordinariness in content marked as advertising.


Introduction
TikTok is a social network type system where users can create, publish, and consume short videos of typically under 60 seconds but ranging up to 10 minutes.The system is unidirectional (users can follow other users), and content is aggregated across hashtags and sounds.The mobile interface centres around the For You Page, often abbreviated as FYP, which displays a feed of algorithmically-driven recommendations that can be navigated one by one.In just five years, the vertical short-video-sharing app has risen in popularity to over 1 billion users, challenging the dominance of YouTube, Facebook, and Instagram as the most downloaded app of 2022 (Ceci, 2023).
Like other popular social media, TikTok is not only a "feel good space" for "silly fun" but also new forms of civic engagement through social media such as "lip-sync activism" or what we could call "duetting and stitching dialectics."TikTok's Chinese ownership and its popularity among very young users (Zeng et al., 2021) have also attracted criticism and scrutiny.During the intensely mediatised US Congressional hearings, concerns were raised around the use of TikTok by children and teenagers.A recent report (Qustodio, 2023) shows that TikTok is the social media app that children spend the most time on-an average of 107 minutes/day-and that TikTok is the most blocked app by families worldwide.The platform attempted to mitigate concerns by introducing features for teens and families in March 2023 (Keenan, 2023).Other issues raised by TikTok's popularity include the representation and self-representation of youth using the platform.Kennedy (2020Kennedy ( , p. 1072) ) points out that the "so-called silly, unashamed and unfiltered girlhood on TikTok, which is epitomised in a figure like D'Amelio, is highly constructed."Hypervisible teenage and young adult TikTok celebrities-some of whom have "grown up" on TikTok-serve as models whose performances are attended to and replicated by young users ("imitation publics") who tend to engage with the platform the most (Statista, 2023).
Concern over social media content targeted at children and teenagers is a catalyst for research into discursive practices emerging on popular video platforms, especially when these develop in conjunction with influencer marketing practices.There is little research into TikTok celebrities and platform-specific influencer marketing discursive strategies.However, existing research into YouTube "kid influencers" (Tur-Viñes et al., 2018) uses established models for content analysis focusing on engagement strategies (McRoberts et al., 2016) and brand presence (Smith et al., 2012) that can be partially transferred to the analysis of TikTok content.
Social media content creators whose fame and recognition have commercial value from a marketing perspective are said to have "influencer" status.Marketing practitioners categorise "influencers" by platform or by audience size (Ruiz-Gómez, 2019): (a) micro-influencers; (b) macro-influencers; (c) mega-influencers; (d) social media celebrities-this latter category being the object of this research.Although research on TikTok communicative forms (Schellewald, 2021;Vizcaíno-Verdú & Abidin, 2022), celebrity (Abidin, 2021); and various influencer/microcelebrity categories (Jaramillo-Dent et al., 2022) has been emerging, and the importance of observing the constructedness of the most popular TikTok celebrities has been argued in previous works (Kennedy, 2020), there is no research into the most followed TikTok content creators worldwide: social media celebrities who reach audiences of tens of millions, beyond the reach of what are usually classified as influencers.
To fill a gap in current scholarship, this research focuses on the top 25 most popular teenage and young adult (under 25) TikTok users and explores discursive practices employed in their most viewed videos to categorise the persona performances that they construct with respect to the desirable values of authenticity and relatability.Another goal is to identify the emergent discursive practices of promotion and self-promotion in terms of adapting persona performance to a promotional context.Considering the platform's memetic processes, TikTok celebrities' performances and discursive practices may constitute templates for other content creators.This study provides much-needed insight into how the platform's young celebrities construct authenticity and relatability, attract and retain users' attention, while also monetising their fame and large audiences (most likely made up of children, teenagers, and young adults who aspire to gain similar status) through ads and paid partnerships.

Literature Review
The short-video social media platform TikTok has emerged in recent years and has the fastest-growing popularity among the younger generation.Scientific interest in the platform initially focused on the platform's affordances and has identified communication forms and memetic processes that engage "imitation publics."With the emergence of new platforms and new types of networked publics, new types of celebrities emerge, and TikTok research meets scholarship on "mass idols" and mediated celebrity dating as far back as the first half of the 20th century.However, our research is also framed by a contemporary research focus on the instrumentalisation of new-found fame and attention for marketing purposes by the liminal social media influencers-both "ordinary" and "extraordinary," simultaneously promoting and self-promoting.

The Affordances of TikTok
Beyond the appeal of similar social media apps, derived from "providing a stage for self-presentation and social connection" (Papacharissi, 2010, p. 304), the affordances of TikTok have been the subject of recent scientific research.Affordances are understood as properties that do not dictate behaviours but shape the users' engagement by configuring the environment in a certain way (boyd, 2010).Bhandari and Bimo (2022, p. 3) reflect on "self-making" as a key set of "intertextual and flexible practices, conventions, and norms of both the production and consumption of visual content" and identify three themes that stand out with respect to users' perception of TikTok: (a) awareness of the algorithm, (b) content without context, and (c) self-creation across platforms.The environment configured by TikTok presents "a very different vision of sociality based on repeated engagement with the 'algorithm'" (Bhandari & Bimo, 2022, p. 10).Schellewald (2021) broadly categorises short videos on TikTok into communicative forms such as (a) comedic, (b) documentary, (c) communal, (d) explanatory, (e) interactive, and (f) meta.More recent work focusing on the platform's affordances identifies the enactment of "TikTok as a 'feel good space'"especially in the context of the 2020-2021 lock-downscharacterised by perceptions of relatability, authenticity, and closeness to the users' "idea of self, taste, and current life situation" (Schellewald, 2023, p. 8).According to the same research, TikTok is perceived as more of a locus for ordinary people than social media celebrities and influencers.Content is consumed in the context of an algorithmic reflection, the FYP feed-a self-representation made up of preferred images and performances of others that can be imitated and shared.Bhandari and Bimo (2020, p. 3) concluded that TikTok is fundamentally different as a social media experience built on "intrapersonal engagement," "directed toward the individual instead of an 'audience.'"Zulli and Zulli (2022) look at "TikTok as a memetic text," with its digital structure initiating "imitation publics" in two major ways: (a) through imitation and replication of specific videos; and (b) through general memetic processes (content/user groupings referred to as "hashtagged publics") such as Straight TikTok, Queer TikTok, Alt TikTok, Deep TikTok, etc.The TikTok structural support for "imitation publics" is perhaps illustrated in the memefication of intergenerational politics through the #OkBoomer memetic trend (Zeng & Abidin, 2021).Recent research (Peña-Fernández et al., 2023, p. 49) shows that in the case of controversial social media debates, TikTok is "a less partisan and more dialoguebased network than Twitter."Literat and Kligler-Vilenchik (2023, p. 1) argue that "TikTok is a vital space to study social movements due to its centrality in youth lives and its ability to give voice to youth political expression," but emphasise the need for balanced and complex approaches that also consider concerns such as misinformation and polarisation fostered on the platform and avoiding the glorification of TikTok activism.
We explore TikTok celebrities' use of hashtags and sounds to identify their positioning with respect to TikTok's publics, as meme founders for "imitation publics," or as mobilisers/participants concerning particular "hashtagged publics."Hence, the first research question is RQ1: What are the main textual and audio discursive practices associated with the performance of the most followed young TikTok celebrities?

Idols, Celebrities, and Influencers
The beginning of the 21st century saw major transformations of the popular media landscape in what Turner (2006) called a "demotic turn" towards "the ordinary," first through reality TV, then through social media fame and celebrity.Lowenthal (1944) looked at the "mass idols" constructed and served for consumption through popular magazine biographies.Duffy and Pooley's (2019) "idols of promotion" are overwhelmingly situated in the sphere of entertainment (film, television, music), following the trend detected by Lowenthal in the 1940s.However, success in the 21st century is owed to promotional skills and self-branding, articulated through three key tropes: "(a) a promise of meritocracy; (b) a spirit of cross-platform self-enterprise; and (c) an incitement to express oneself authentically" (Duffy & Pooley, 2019, p. 28).Concepts such as "familiar strangers" (Milgram, 1977) and "para-social interactions" (Horton & Wohl, 1956) still provide theoretical grounding to many of the contemporary conceptualisations of mediated celebrity (Rojek, 2015).
Traditional celebrity studies have produced various fame taxonomies (Rojek, 2001;Ruiz-Gómez, 2019), with types of fame such as (a) ascribed, (b) achieved, or (c) attributed; and other celebrity types such as (d) celetoids, (e) celeactors, (f) infamous, (g) accidental celebrities, and (h) subcultural celebrities.The age of social media and digital platforms saw the rise of various other nomenclatures and taxonomies for internet celebrities (Gamson, 2011): (a) anticelebrities or accidental celebrities, (b) do-it-yourself celebrities, and (c) microcelebrities.
Research into influencer marketing identifies endorser types (celebrities or influencers) as well as content-control types (content creators vs. paid promoters) and shows that consumers think influencers have "more expertise, trustworthiness, more correspondent inferences, and more authenticity" (Kapitan et al., 2022, p. 347).This research provides a categorisation of persona performances constructed and employed by top young and teenage TikTok celebrities, focusing on features of authenticity and promotion.

The Construction of Authenticity
Members of the networked publics most often bring up authenticity as a criterion for following certain content creators.It then becomes an important value for creators and marketers.In social media-as with tourismauthenticity becomes a "selling point" or "source of credibility" (van Nuenen, 2016).An analysis of three categories of content producers on YouTube-science YouTubers, stay-at-home mothers, and make-up artistspoints out that sometimes tactically adopted "ordinariness and amateurship markers contribute to the online construction of the authentic persona" (Riboni, 2020, p. 22).Riboni (2020) identifies three aspects of authenticity in mediated public discourses: (a) the parasocial features of the constructed media persona (Horton & Wohl, 1956), which social media transform into potentially social (Marwick, 2013)-the social media celebrity is more approachable than the TV celebrity; (b) the degree of homophily between content creators and their audience, contributing to relatability; (c) emphasising the viewers' importance to the content creator through "synthetic personalisation."We can see evidence of this when YouTubers, Instagrammers, or TikTokers address messages in the live chat, respond to specific comments, or simply imply some shared complicity with their viewers.
In Duffy and Pooley's study (2019, p. 41), constructed authenticity is key to self-promotion since "cultivating a true-to-self persona is a form of value-creation."Through self-effacement in TV interviews or social media posts, celebrities construct themselves as relatable, ordinary people.Lee (2020) provides a three-component model for authenticity in (mass-oriented) computermediated communication: (a) authenticity of source, (b) authenticity of message, (c) authenticity of interaction.The model was applied to exploratory research into the self-perceived authenticity of social media influencers (Balaban & Szambolics, 2022).Riboni (2020, p. 36) also points out the paradoxical commodification of authenticity and applies the notion of "emergent authenticity"-from tourism studies (Cohen, 1988)-to designate "the gradual shift from inauthenticity to authenticity" that may happen over time to make an experience be perceived as authentic or more authentic, even if it is not.Goffman's (1959) dramaturgic approach distinguished between front-stage and backstage.The front-stage personas are fabricated, inauthentic social performances, whereas in the back-stage, in intimate settings, people disclose their genuine, authentic selves.However, later interpretations (MacCannell, 1973) consider the possibility that the back-stage self is also constructed as inauthentically as the front-stage.Riboni (2020) points out that this type of "staged authenticity" is dominant in social media discourses.
To explore the constructed performances of popular young TikTokers, we used Goffman's (1959) dramaturgic approach and expanded the classical categories (stage, props, costume) with categories that allow the classification of the persona performance with respect to the platform's memetic processes.In our analysis of discursive practices employed by popular TikTokers in their most viewed videos and also in content marked as advertising, we draw from Frye (1957) to code the persona/protagonist constructed in relation to ordinariness.We apply the systematic representation of the four archetypal mythoi (comedy, romance, tragedy, satire/irony) and five modes corresponding to ethos performance (ironic, low mimesis, high mimesis, romantic, mythic) mapped to degrees of ordinariness performed: less than ordinary, ordinary, more than ordinary, extraordinary, more than extraordinary.The four pre-generic plot structures provide a meaningful way of categorising content and distinguishing between broad categories.Accounting for the liminality of social media celebrities, we try to identify the ethos of each performance analysed to ascertain their variability across popular content and content marked as promotional.Furthermore, we employ a taxonomy of communicative forms that combines the typology identified by Schellewald (2021) complemented by the time-tested taxonomy of modes proposed by Nichols (1991).Thus, the first three aspects of the conceptual model proposed here (mythos, ethos, and form) also fit with the broad memetic aspects defined by Shifman (2013;content, stance, and form).The second research question is formulated with respect to the elements that make up the persona features that can be identified in the popular TikTok performances and to thus identify coherent persona performances adopted by the TikTok celebrities: RQ2: What type of persona performance features are employed in the most popular videos?

Promotion, Self-Promotion, and Emergent Discursive Practices
In an era dominated by digital mobile platforms, Marwick (2013) suggests that our perceptions of celebrity are intrinsically linked to the affordances of web and mobile media technology-rewarding those whose behaviour and use of self-presentation strategies engage people's attention with higher social status.Abidin (2017, p. 1) provides a useful point: for influencers and aspirational influencers, being a microcelebrity is "an established career with its own ecology and economy."Turning influencer status into a career prospect or expectation is perhaps most visible in the case of micro-microcelebrities (Abidin, 2015), "lifestyle" influencers (Abidin, 2016), and child microcelebrities of "family influencers" (Abidin, 2017).Abidin (2017, p. 1) defines calibrated amateurism as a complex practice and aesthetic employed by actors in the attention economy who craft contrived authenticity, which portrays "the raw aesthetic of an amateur." Social media celebrities, "influencers," microcelebrities, and "ordinary users" engage in similar selfaware promotion, especially self-promotion strategies.For both "idols of promotion" and "ordinary users," the projected identities in everyday self-performance on digital platforms are "both personal and employable" (Duffy & Pooley, 2019, p. 43).A similar point is made by Negreira-Rey et al. ( 2022) about the blurring boundaries between "ordinary" content creators, microcelebrities, influencers, and professional public communicators such as journalists.Mellado and Hermida (2021) identified new roles for journalists on social media: (a) the promoter, (b) the celebrity, and (c) the joker.Earlier research that looks at contemporary shifts in the professional roles of journalists (Hanitzsch & Vos, 2018) acknowledges emergent roles that help publics navigate three areas of the domain of everyday life: (a) emotion, (b) identity, and (c) consumption.As the new public communicators fill a void left by the younger generation's disengagement with news media, it is appropriate to use these categories to identify the specific roles that social media celebrities and influencers fulfil.
In her conceptualisation of "calibrated amateurism," Abidin (2017) further details three calibration spectra: (a) selectively toggling technology use from smartphones to professional studio equipment; (b) toggling between "anchor"/main content for which the creator is recognised and followed, and "filler"/impromptu/casual content; (c) correlating performances and apparent spontaneity with reflexive, behind-the-scene revelations.Furthermore, Abidin (2017) points out five ways in which the "amateur" aspect anchors reliability: (a) daily frequency of "filler" content; (b) mundanity and privacy of "filler" content; (c) capitalising on platform affordances and subverting use to show images with and without beauty filters or enhancements, outtakes, and actual-takes; (d) reinforcing or challenging the norms of the platform by willingly breaking the constructed persona or engaging in meta self-parody; (e) turning the occasional behind-the-scenes content into a genre.From Riboni's (2020) analysis of YouTuber discourses, we draw several persona types such as (a) "the ordinary expert," (b) "the passionate non-specialist," (c) "the life guru," and (d) "the expert influencer."Barta and Andalibi (2021) notice that the "normative authenticity" of TikTokers is connected to the perceived anonymity of both creators and audience members, which in turn facilitates sharing emotional content-both positive and negative/traumatic/stressful-and eliciting support, empathy, or solidarity responses.Abidin (2020) attempts to map internet celebrity on TikTok and finds that in the TikTok attention economies, (a) "post-based virality is privileged over persona-based fame," (b) "audio memes are the driving template and organising principle," (c) video editing is a marker of expertise, and (d) observable memetic histories foster competitivity.The music challenge content has also been found to constitute more of a mode of transmedia storytelling rather than a competitive practice (Vizcaíno-Verdú & Abidin, 2022) that allows for self-expression and the construction of ad-hoc communities based on various in-group affiliations.
The final components of our proposed conceptual model are drawn from (a) the domains of everyday life (Hanitzsch & Vos, 2018) within which public communicators construct contexts for the development of parasocial relationships; (b) Abidin's (2017) conceptualisation of "calibrated amateurism"-the "anchor"/"filler" content labels; and (c) previous research into brand presence on social media (Tur-Viñes et al., 2018).In our investigation, violence and sexualisation (as clarified by Poppi & Dynel, 2021) were also analysed to observe the phenomenon of young performers' sexualisation identified by Vizcaíno-Verdú and Tirocchi (2021).In order to identify the changes that occur in the constructed persona and performance when celebrities engage with their audiences as promoters, we reach the final research question: RQ3.What persona performance features are more salient in the context of promotion and advertising videos?

Conceptual Model
The proposed conceptual model introduced in Sections 2.3 and 2.4 is summarised in Table 1.

Data Collection and Sampling
The sample was constructed based on the Tokboard top (https://tokboard.com/users/top)most followed TikTok users by excluding any users above the age of 25, based on platform data as of November 2022.The website is unaffiliated with TikTok and uses automated data collection from the platform.The top was compared for reliability against rankings provided by another independent source (SocialBlade).As this research is focused on teenagers and young adults, celebrities above the age of 25 were excluded, as the literature suggests that greater homophily between content creators and consumers greatly contributes to relatability (Riboni, 2020).
As Table 2 shows, ages range from 15 to 25, while follower numbers range from 36 to 151 million.Fifteen of the 25 are based in the US, while another four are based in Mexico, making the sample more anchored in American culture.There is only one content creator under the age of 18.For each of the 25 users (Table 2), the TikTok API was interrogated using the Clockwork TikTok Scraper on Apify (https://apify.com),and a maximum number of 1,000 video metadata records were extracted for each user.Due to some users having posted less than 1,000 videos overall, the final sample size was N = 22,650.Two subsamples are analysed by coding the categories in Table 1: 1.A popularity subsample: top 10 videos by view count for each content creator-N p = 226 (24 videos initially included in the sample were no longer available for analysis in the coding phase); 2.An ads subsample: content marked as advertising in the metadata ("IsAd" = True)-N a = 213.

Data Analysis
We used frequency and co-occurrence analysis on the metadata to observe the systematic use of certain ways to address TikTok publics as well as the overall use of platform features such as sounds and filters.
Both the TikTok advertising sample (N a = 213) and the most popular content sample (N p = 226) were independently coded by two researchers.Before coding, both researchers underwent a training session covering all coding categories.To ensure consistency, a set of 30 videos from the popularity sample was used for training purposes.The intercoder agreement for the advertising sample was calculated to be 81.22%,while the intercoder agreement for the popularity sample was found to be 86.28%.
The co-occurrence analysis of hashtags and the semantic network of concept codes were generated using KH Coder (Higuchi, 2016(Higuchi, , 2017)).The hashtag co-occurrence analysis was run on the entire dataset (N = 22,650), while the semantic network represents the coded concepts for the cases in the popularity sample (N p = 226).
We used formal concept analysis (Ganter & Wille, 2012) to generate a conceptual diagram of discursive features using the FCA Tools Bundle (Kis et al., 2016).The formal context was based on the dominant codes occurring more than once for each content creator in the popularity sample.

The Overview of Discursive Practices
The most popular young TikTokers produce and publish very short videos (between 10 and 20 seconds), with only  1 shows, this may be due to several videos becoming viralized, as in the case of bellapoarch, who has a relatively smaller number of uploaded videos but has produced three that total over 3 billion plays.Beyond the virality of several videos, charlidamelio, a former competitive dancer, attracts on average the highest numbers of diggs, shares, and comments, emerging as a typical young TikTok celebrity, also capable of extending her aura of social media celebrity to her family and social network.Celebrities with large audiences on other video platforms, such as mrbeast, seem to be able to transfer their success formula to TikTok.Table 3 provides an overview of metadata.
A third of the videos in the sample (7.245, 32%) feature sound or music uploaded to the platform by others (musicOriginal = false).These are trends and challenges that the most followed TikTokers start or in which they participate.Figure 1 shows the most played such videos.In most cases, it seems that the performance constructed by one of the TikTok celebrities clearly dominates the stage set by each sound-this is likely the case when the celebrities are trend starters or meme founders (Shifman, 2013).There are some cases where the views are more evenly distributed among two or more.This is because young TikTok celebrities not only initiate memetic phenomena, they also participate in communal platform rituals, challenges, and trends, adding to their performance authenticity; they participate in such "silly fun" as a performance of their ordinariness, adding, of course, their own interpretation.Even though navigating across trending sounds is one of the characteristics of TikTok, hashtag use has the potential to engage with specific issues and communities, the so-called "hashtagged publics" (Zulli & Zulli, 2022).It seems, however, that the most followed young TikTok celebrities do not engage with specific publics, mostly keeping to a positive, mainstream addressability and using hashtags that encourage parasocial relations, such as #yzfamily or #loveyouall.
Compared to the Ads subsample (N a = 213), we noticed that ads are mostly not marked as advertising in the metadata ("IsAd" = True).This result is consistent with previous results published in a Mozilla Foundation Investigation (Mozilla, 2021), which concluded that "TikTok has a major weakness when it comes to influencer advertising" because ads aren't placed through the platform's advertising system, resulting in much less oversight and more likelihood of users being misled in paid promotions/advertising.

The Personas and Performances
The performances in the most played videos subsample (N p ) vary between the ethos of "ordinariness" and  "more than ordinariness."Authenticity is constructed through alternating "anchor" and "filler" content, participating in challenges and trends.The dominant mythos or pre-generic type is that of comedy-TikTok is celebratory in either poetic or communal modes of expression.However, there are also alternatives, such as the Romantic, larger-than-life exploits of mrbeast, who carries over to TikTok his YouTube content/recipe of displaying status, spending large amounts of money seemingly on a whim.Furthermore, the most followed TikToker in the sample, khaby.lame,uses satire and deconstructs platform content tropes, thus performing authenticity ironically.Figure 3 shows the semantic network of codes and creators.
To group persona features described in the conceptual model, we employed formal concept analysis (Ganter & Wille, 2012) and selected the features that were coded more than once for each of the 25 celebrities to construct a formal context.Formal concepts were then computed using the FCA Tools Bundle, and the resulting concept diagram is shown in Figure 4. Concepts (groupings of the celebrities based on feature combinations) were then used to identify and describe the distinct persona performances presented in Table 4.
Figures 3 and 4 show that in the case of the most viewed videos, several discursive practices emerge as distinct persona performances presented in Table 4.For each of the proposed categories, the intension (list of coded features) and the extension (list of celebrities who match the combination of coded features) of the formal concepts are listed.The formal concept analysis method yielded the concepts represented as nodes in Figure 4.The concepts at the top are more general (as they correspond mostly to only one feature), while the concepts at the bottom are more specific (as they correspond mostly to only one celebrity).The seven persona performances were selected based on representativity, coherence, and good balance between generality and specificity.
To answer RQ2, the persona performances employed by young TikTok celebrities in their most popular videos are mostly comedic and liminal (between "ordinary" and "extraordinary") in the first four of the seven persona performances.However, there are notable exceptionswith satirical/ironic, romantic, and tragic persona performances sometimes emerging either somewhat idiosyncratically or contextually-the latter three.The "ordinary expert," the "plebeians," and the "empowered victim" fit with categories proposed in previous research on the construction of authenticity on YouTube (Riboni, 2020).The "MC" and the "platform sweetheart" are platform adaptations of charismatic vloggers, streamers, and Instagram celebrities.The "star among us" is relatively atypical-perhaps also encountered in the case of mainstream media celebrities using the platform-while the deconstructive "raissoneur" persona sets itself on the outside, in a meta stance, satisfying perhaps a need for critical reflection on the platform's discourses on authenticity and "DIY" ethos.

The Ordinary Promoters
While the mythos of comedy clearly dominates, ethos performance fluctuates between "ordinary" and "more than ordinary" for most of the videos in the popularity subsample.However, the content marked as ads through the platform metadata contains a dominant proportion of "ordinary" performances, as shown in Figure 5. Furthermore, even though the "poetic" and "communal" forms (dancing, nonverbal, memetic challenges) are still dominant, "expository" performances are more common in the ads subsample than in the popularity sample.
Brands are rather seen (through use, visual identity, and text mentions) than heard (through brand sounds or oral mentions), which constitutes an interesting result given the platform's affordances-specifically its memetic sounds and lip-sync video formats.Most of the content marked as ads is not representative of a TikToker's usual discourse, being mostly filler/casual posts (n = 182) in contrast to the 31 posts where the ads follow the expected anchor/main style.To answer RQ3, persona performances in the context of promotion tend to be more oriented towards "ordinariness" and showing rather than telling.

Discussion
In terms of posts' aesthetics, our findings are in line with the ones theorised by Abidin (2021) casual, and sporty outfits, while 82.74% (n = 187) out of the total posts (n = 226) feature home-based activities presented in videos made indoors, moving away from a staged, picture-perfect background.This visual construction of the "calibrated amateurism" of the TikTok influencers is complemented by the self-presentation as an ordinary person disclosing their daily life experiences (n = 69) but succeeding in doing better than the rest because they are smarter, more talented, more beautiful, craftier, and stronger (n = 144).
Nevertheless, it is difficult for internet celebrity aspirants to construct a complex and coherent persona because they are permanently forced by the logics of the platform to seek out, learn, participate in, and engage in what is "going viral" at the moment in order to remain visible to others on the app (Abidin, 2021).Therefore, apart from the memetic emotion-triggering performances, TikTok celebrities also propose relatable performances with multiple characters, such as high school romance, pyjama parties, duet or group dances, and teen pranks (communal n = 79).
Cohesion is also affected by promotional content.The TikTokers' desire to separate their persona from the brands that sponsor them, to maintain an authentic relationship with their community, and the brands' requirements that specific information be conveyed leads to posts being constructed on different templates than the anchor content and less consistent personas.TikTok's Table 4. Persona performances of young TikTok celebrities.

The MC (master of ceremonies)
Performs enthusiastic participation in communal, ritualistic performances such as memetic challenges, setting an example, encouraging certain trends, or just joining an already popular trend as an ordinary member of the platform (comedy, communal, emotion, more than ordinary/ordinary: avani, brentrivera, dixiedamelio, elrodcontreras, homm9k, justmaiko, kimberly.loaiza, riyaz.14, xoteam) The platform sweetheart Delivers poetic and communal performances in the realm of emotion, mostly comedic, using memetic sounds and nonverbal expression, lip-synching, and dancing-a teen idol, eliciting admiration and parasocial relations (comedy, emotion, more than ordinary, poetic, sexualisation: addisonre, babyariel, charlidamelio, domelipa, elrodcontreras, kimberly.loaiza,montpantoja) The plebeian(s) Deliver(s) slapstick comedy performances, often based on/staged in real-life situations, also involving interactions between two or several personas (comedy, emotion, ordinary, performative: babyariel, brentrivera, dobretwins, joealbanese, xoteam) The "ordinary" expert Delivers performances based on expertise/status, sometimes about how to make successful TikTok content and includes reflexive "behind-the-scene" or "peek-behindthe-curtain" sequences (comedy, consumption, more than ordinary, sometimes reflexive: avani, brentrivera, dobretwins, homm9k, justmaiko, youneszarou) The raisonneur Deconstructs and satirises platform tropes and trends, allowing elevation above "ordinary" users in ambivalent stances-critical of platform practices, yet involved (identity, more than ordinary, reflexive, satire: homm9k, khaby.lame,lorengray) The benevolent star Delivers performances of generosity, displaying status in real-life and virtual interactions with others-it is the traditional media celebrity performance (comedy, romance, communal, consumption, emotion, identity, more than extraordinary, participatory: mrbeast) The empowered victim The tragic mythos is rarely found in most played videos.It is, in fact, more frequent in the Ads subsample.However, when it is performed, it is coupled with "ordinary" or "less than ordinary" ethos performance in a self-referential poetic or reflexive form as "the empowered victim," eliciting sympathy and solidarity, apparently instrumentalised for self-promotional strategies (bellapoarch and dixiedamelio in their musical content) "idols of promotion" calibrate "ordinariness" to satisfy marketing requirements and expectations for authenticity and relatability.Brand advertising should account for anchor and filler content templates, persona perfor-mance types, as well as the celebrities' own promotional strategies and the viewers' awareness of social media marketing.As TikTok users tend to watch more content from creators they do not follow, marketers should aim to integrate advertising in the most engaging performances.Integration as a revelation in behind-the-scenes "ordinary" or as a facilitator for the "more than ordinary" or "extraordinary" performances are two plausible approaches.
Based on the affordances of the platform, several of the top young TikTokers (charlidamelio, addisonre, kimberly.loaiza, dixiedamelio, lorengray, avani, elrodcontreras, xoteam, montpantoja, babyariel) engage in performances (dancing, lip-sync, acting) that result in different degrees of sexualisation of the sampled video content (31.41%, n = 71), indicated by body exposure combined with provocative outfit, perpetration and gaze, facial attractiveness, lascivious gestures and movements, similar to the sexual objectivation techniques used in the music industry and advertising.But, as the literature suggests (Vizcaíno-Verdú & Tirocchi, 2021), this self-sexualisation is perceived by the performers as a form of empowerment, while the sexualised content featuring other female performers is assessed as a form of objectivation.

Conclusions
The research complements existing research on TikTok by looking at the most successful young celebrities emerging on the popular short video platform.It is relevant, especially in the context of recent concerns over the platform's management of child users.
The main findings indicate that, as expected, comedic, predominantly nonverbal content dominates the performances constructed by the most followed young TikTokers.The TikTok celebrities do not specifically engage with "hashtagged publics" (Zulli & Zulli, 2022), presumably because they address the mainstream and seek broad popularity.Their construction of authenticity relies mainly on acting out either an "ordinary" or "more than ordinary" ethos.The use of hashtags reinforces the main content (the tongue-in-cheek #learnwithkhaby encourages parasocial relations or signals advertising content), in most cases circumventing the platform's advertising system.It is also interesting to note that a significant part of their performances include self-sexualisation and implied violence-elements that open up paths for future research.
In the context of platform memetic logics, TikTok celebrities' performances are often the origin of founderbased memes (Shifman, 2013).Beyond individual short videos, "imitation publics" may turn persona performances into memes, thus creating templates for the performances of millions of young users.The proposed persona types play out ordinariness in different styles: (a) the "master of ceremonies" is a ritualistic persona that creates or engages with trends and challenges; (b) the "platform sweetheart" thrives in the space of nonverbal dance performances and instrumentalises sexualisation; (c) the "plebeians" similarly instrumentalise violence with their slapstick humour; (d) the "ordinary experts" set themselves up for easy switching into influencer performances.Furthermore, some notable performances fall outside this type of construction of authenticity: (e) the "raissoneur" persona proposed by khaby.lame'ssatire; (f) the "benevolent star" persona transferred to TikTok by the YouTuber mrbeast; (g) the "empowered victim" triggers sympathy and solidarity by revealing or implying uncomfortable personal tragedies.Since TikTok also sets itself apart from other social media systems by allowing users' consumption patterns to construct a fluid, algorithmised self (Bhandari & Bimo, 2022), the persona performances function as both templates and reflections.
When the young TikTokers perform in the context of advertising or partnerships and endorsements, they tend to perform differently than their established main persona-ads rely on "filler" content where the persona constructed is notably "ordinary."Findings confirm and complement previous work by Abidin (2017Abidin ( , 2018Abidin ( , 2020) ) and Riboni (2020).
The conceptual framework constructed for this analysis and the resulting proposed persona performance types constitute the main contribution of this work.The conceptual and analytical approach allows for the characterisation of performances through the analysis of several layers.The size of the two subsamples constitutes a limitation for this research.A larger sample of videosincluding the less successful videos or those marked as ads in the hashtags but not in the metadata-would yield interesting results.The scope of our research was limited to popular performances, and future research is needed to validate the framework and proposed personas on samples that better represent the diversity of TikTok content creators.

Figure 3 .
Figure 3. Semantic network of concept codes and creators in popularity sample (Jaccard coefficient > 0.05).

Figure 4 .
Figure 4. Concept lattice diagram for the dominant discursive features found in the most viewed videos based on formal concept analysis.

Figure 5 .
Figure 5. Brand presence and ethos performance in the Ads subsample.
Andreea-Alina Mogoș is an associate professor at the Department of Journalism and Digital Media at Babeș-Bolyai University.In 2009, she earned a PhD in sociology at Babeș-Bolyai University and a PhD in information and communication sciences at Université Paris 8 Vincennes-Saint-Denis with the thesis Media Representations of the Romanians in the French Daily Newspapers.Her research interests focus on the mixed-methods approach of the textual and visual media frames and representations, and the transformation of media genres.George Prundaru is a lecturer part of the Faculty of Political, Administrative, and Communication Sciences, and teaches at the Department of Journalism and Digital Media, at Babeș-Bolyai University.His research interests revolve around new media, spanning from the content of digital communication (video game and transmedia narratives) to the forms it takes (visual communication and design), and the processes that connect them (user experience design).

Table 1 .
Conceptual model and codes (full descriptions available in additional tables).

Table 2 .
The top 25 young user sample.