Facework in Confessional Videos by YouTube Content Creators

This article analyses the role of facework in the discourse of confessional YouTube videos by female fashion and beauty content creators, where they disclose personal problems, and offer viewers advice. It uses thematic analysis to identify discursive tactics that protect viewers’ face. The article argues that the parasocial nature of the connection that these videos attempt to establish with an audience that content creators know little about makes it important for them to reflexively adapt to these viewers’ needs for fellowship and autonomy. Their disclosures may be intended to create closeness, but at the same time they need to cater for distance and prepare the ground for this content to be received as well as possible. However, just like the connection that the videos seek to establish, the facework they contain is also parasocially situated: the videos speak to an imagined viewer’s need for inclusion and this viewer’s possible objections, as these are perceived by the content creator. The article contributes to a better understanding of the construction of closeness in this genre of mediated discourse.

the audience. It analyses videos by female fashion and beauty content creators, in which they open up about personal problems and offer their viewers advice on how to overcome similar situations. The article applies qualitative thematic analysis on the videos of two very successful and established content creators, and finds that the intimate disclosures being made are embedded in a discourse that anticipates an imagined viewer's reactions, and protects their positive and negative face by extending them fellowship and by avoiding imposition. The article makes a contribution to research in this area by identifying the ways in which the YouTube videos analysed verbally negotiate their viewers' response, prior to actually receiving that response.
The article argues that the parasocial nature of the relationship between content creators and their viewers makes facework imperative in these videos. Confessional videos may be intended to create closeness, but at the same time they need to cater for, and carefully bridge distance. As opposed to a social relationship where an intimate disclosure is usually made to someone the speaker knows and has an established relationship with, YouTubers have limited knowledge of their audience and cannot monitor their reactions to their videos in real time. The article finds that, while the analysed YouTubers perform intimacy by revealing personal experiences and emotions, at the same time, their discourse second-guesses how viewers will take these and pre-empts negative interpretations by deploying face-saving tactics.
By exploring the role of facework in confessional discourse in these YouTube videos, the article contributes to our understanding of the construction of intimacy in this context, and highlights that this is not only about showing vulnerability, as most literature acknowledges, but it is also about imagining and reflexively adapting to the viewer and their reactions.

Influencers, Authenticity and Intimacy
Influencers are everyday people who have gained visibility by regularly creating and posting content about commercial products (for instance fashion, beauty products, gadgets and travel) and about their personal lives on social media (Abidin, 2015). These online content creators are an example of the demotic turn (Turner, 2006) enabled by reality television and social media, which allowed ordinary people to become famous among a wider public. Social media in particular enabled the emergence of microcelebrity, namely 'a new style of online performance in which people employ webcams, video, audio, blogs, and social networking sites to "amp up" their popularity among readers, viewers, and those to whom they are linked online' (Senft, 2008:25). Marwick (2015) stresses that microcelebrity is a performative practice which involves creating a self-constructed persona through the textual affordances of social media; thus anyone with access to a mobile device and social media can practise it. Indeed much academic literature analyses the performance of particularly female content creators in relation to how they enact authenticity. As opposed to traditional celebrities who used to maintain a distance from their fans, microcelebrities' popularity relies on revealing aspects of their authentic selves and lives (Marwick, 2013). Authenticity means speaking and acting sincerely, and being open about one's thoughts, experiences and feelingsit involves coming across as a 'real person', not a persona (Tolson, 2010). Content creators' credibility relies on being seen as ordinary people by their followers, who constantly assess and validate influencers' claims to authenticity (Cunningham and Craig, 2017).
To further reinforce their claim to authenticity, influencers mobilise 'communicative intimacies' (Abidin, 2015) with their followers, especially in the longer form medium of YouTube videos. These attempts to construct a mediated intimate relationship are widely seen as instigated by influencers rather than their audience, ultimately aiming to enhance their online persona, and commercial success. Abidin (2015) suggests that influencers' intimacy strategies focus on interaction, reciprocity and disclosure, and involve responding to their followers' messages and inviting them to comment, showing selective backstage aspects of their private lives, and organising real-life meetings with followers.
In essence giving people privileged access to private, backstage information, that is not available to just anyone, marks a relationship as intimate (Simmel, 1950). Influencers may give this access to anyone who follows them on social media, but the act of disclosing personal moments, feelings or information still invites associations of closeness and trust (Schwarz, 2011). Raun (2018) argues that intimacy is so integral to influencers' work, that it is a genre in itself that has come to be expected on YouTube.
In videos where they open up about personal problems, like the ones analysed in this article, influencers perform a vulnerable self who confides in and seeks emotional support from viewers (Jeslev, 2016), thus narrowing the perceived gap between them and their followers (Marwick and boyd, 2011). Disclosing difficult experiences becomes a tool of self-branding by 'mobiliz[ing] their personal stories to build an online persona and cultivate a following' (Harrington, 2019(Harrington, : 1186. Cultivating a following means more views for the content creator's videos and consequently more potential for monetisation through endorsing commodities (Berryman and Kavka, 2017). This is particularly the case for the more popular female fashion, beauty and lifestyle YouTubers, who follow gendered norms of intimacy by adopting a 'big sister' role towards their viewers when making intimate disclosures (Berryman and Kavka, 2017). Authentic disclosure and selfcommodification are thus not irreconcilable in an online environment, but essentially complementary (Marwick, 2013). Content creators invest considerable affective labour to achieve intimacy with their viewers because this can then be capitalised for money or social recognition (Raun, 2018).

Parasocial Interaction
Essentially though, these relationships that influencers attempt to build by performing authenticity and mobilising intimacy are parasocial. The term parasocial interaction was introduced by Horton and Wohl, (1956) to refer to how members of an audience may respond to a real or fictional personality that appears in the media, like they would respond to people they know personally in their social interactions. Parasocial interactions can be with TV presenters, actors, celebrities or soap opera characters, and they are essentially one-sided: the media user feels like they know the media personality, they have opinions about them and may like or feel attached to them, but this is not reciprocated by the media figure, who does not necessarily know anything about any specific member of the audience.
Parasocial interaction has been seen as serving audiences' need for companionship and social identification (McQuail et al., 1972). Indeed people apply similar criteria and principles to these interactions as to their social relationships: for example, they develop more complex parasocial interactions with soap opera characters whose behaviour is predictable (Perse and Rubin, 1989) and with media figures whom they perceive as similar to themselves in background and attitudes (Turner, 1993), just as happens in ordinary social relationships. Giles (2003: 192) notes that some of the academic debates on parasocial interaction have treated it as something 'irrational' or 'maladaptive' and he argues that such an approach removes the concept from its original understanding as normal social behaviour. He thus proposes a model that treats parasociability as a continuum, where 'the more remote interactants are from one another, the more features their relationship shares with a parasocial one' (ibid). This model draws attention to factors like distance, the amount of information the two parties have on each other, and the possibility of meeting someone in person, which lend a degree of parasociability to encounters we don't normally think of as parasocial, such as lectures, conventions or board meetings.
However reciprocity (or lack of it) remains the one characteristic most commonly associated with parasocial relationships between media figures and audiences. Rihl and Wegener (2019) explore parasocial relationships, namely series of parasocial interactions, between German YouTube content creators and their viewers. They conclude that these relationships fit the parasocial definition because they are one-sided and rarely interactive, while the feedback channels the platform allows are generally underused. Chen (2016) also sees the relationship between YouTubers and their audience as one-sided and therefore parasocial. However, in this case it's the YouTubers who 'imagine' a relationship with their audience by interpreting the number of views their videos get as validation for how they present themselves online (Chen, 2016). This is despite the affordances for reciprocal interaction between media figures and their audience on social media, which have the potential of bridging the social and the parasocial aspects in this relationship (Click et al., 2013). This potential may be achieved to an extent in some cases (Click et al, 2013) but very often it is not (Stever and Lawson, 2013).

Facework and Adapting to an Imagined Audience
This article has so far established that content creators have high stakes in mobilising authentic and intimate relationships with strangers, but in a digitally mediated context these relationships are essentially parasocial and one-sided. Content creators know less about their audience than the audience knows about them, yet they need to build and sustain an intimate parasocial connection. This means that it's imperative for content creators to imagine their audience's reactions when presenting themselves online, if they are to create such a connection.
In her analysis of the self-presentation of ordinary social media users, who are not influencers, Robinson suggests that they are constantly aware of other users' response to their behaviour, to the extent that everything they do and say about themselves online is 'inseparable from the audience's anticipated response ' (2007: 96). They constantly 'read' online interactions for explicit and implicit cues, in response to which they reflexively adapt their behaviour. This 'networked me' (Chambers, 2013:71), namely the self that social media users present to the network of people they are connected to online, is thus a collaborative product between the individual and their network.
Such a reflexive adaptation is also found among influencers who share photographs of their everyday moments on Instagram (Duffy and Hund, 2019). They expect critical feedback from followers and this impacts on which aspects of their lives they choose to share. Instagram influencers were found to engage in 'strategic acts of revelation and concealment' based on the reactions they anticipated of their audience (Duffy and Hund, 2019: 4997). Second-guessing these reactions is arguably rendered more challenging due to the parasocial nature of influencers' relationship with their audience. It is not possible to know who may be in the audience for anything that is posted online; and it is not possible to know in which context anything posted will be interpreted (boyd, 2010). These gaps in knowledge need to be filled in by essentially imagining the audience and its reactions.
Anticipating others' reactions and adapting accordingly is at the core of facework (Goffman, 1955) in everyday social interaction. However, the role of facework in content creators' parasocial interactions with viewers has not been addressed in the academic literature so far. Face is the positive image of ourselves that we try to establish in presenting ourselves to others, and facework includes our attempts to protect our own and other people's faces (Goffman, 1955) especially when we anticipate that what we are about to say may threaten them, or when we are uncertain of how what we say may be interpreted. Facework is performed in anticipation of a reaction before it has occurred, not as a response to it.
Brown and Levinson distinguish between positive face or 'the want of every member that his wants be desirable to at least some others ' (1987: 62) and negative face namely 'the want of every "competent adult member" that his actions be unimpeded by others ' (1987:62). In Lim and Bowers' (1991) terms, the positive (or fellowship) face represents our need to feel included and the negative (or autonomy) face the need to not be imposed on. Lim and Bowers' (1991) terminology is narrower than Brown and Levinson's (1987): the former distinguish between the fellowship (our need to feel included by others) and competence face (our need to feel others respect us as capable people), but the latter comprise both of these needs under positive face. In my analysis of confessional videos, all examples and discussion of positive face refer to fellowship and all discussion of negative face to autonomy.
These needs may be intentionally or unintentionally challenged or undermined by facethreatening acts, such as complaints, insults, orders or requests to do things. Facework therefore involves a speaker's attempts to mitigate the possible face-threatening impact of these acts and protect their addressee's face (Goffman, 1955). So in making a request that imposes on an addressee's negative face, we may use markers like 'please' or phrase the request as a question or an observation rather than as an order (Brown and Levinson, 1987). Or we can express solidarity with the addressee to avoid threatening their positive face when making a complaint. In any case, facework is situated socially and exists in interaction with others: we only try to save others' faces in the context of trying to establish relationships with other people, especially people we do not know a lot about. Arguably, at the same time as protecting others' faces, facework helps the speaker represent themselves as a considerate person, who deserves to be included by others, so the speaker simultaneously protects their own positive face.
This article explores what role facework plays in the discourse of YouTube videos, where female fashion and beauty content creators attempt to build parasocial relationships with their viewers by confiding in them. It looks at whether and how YouTubers' discourse in these videos is mitigated by anticipation of an imagined audience's reaction to what they say. The article argues that extending fellowship and protecting viewers' sense of autonomy are discursive steps that contribute to an attempt to build a parasocial interpersonal connection with people they do not know and cannot see.

Method
This study is a thematic analysis of confessional YouTube videos, where two successful fashion and beauty content creators establish communicative intimacies (Abidin, 2015) with their viewers by disclosing personal problems and offering advice to those in a similar position. This type of video was selected because it constitutes an important way of creating parasocial relationships of closeness with viewers. The article seeks to address the following questions: RQ1: Do content creators deploy facework in these videos and in what ways? RQ2: How do these videos' discursive attempts to protect their viewers' fellowship and autonomy faces contribute to an attempt to mobilise a parasocial connection?
The material analysed was chosen after following ten UK-based female fashion and beauty content creators, with different numbers of followers, over a period of 5 years. This allowed me to identify facework as a useful analytical tool to understand their discourse, having noticed that these YouTubers regularly pre-empt viewers' reactions, especially when they reveal their own emotional experiences within vlogs.
This explorative immersion also enabled the selection of the YouTubers and the videos analysed systematically in the qualitative thematic analysis. Although most YouTubers share information about themselves in their regular content, a few additionally create dedicated confessional videos to discuss their personal problems, especially around mental health and resilience, and to advise viewers who face similar issues. The analysis focused specifically on these dedicated confessional videos and the YouTubers who created them, because they provided richer discursive data in which to examine the deployment of facework strategies in attempts to create closeness with viewers. As discussed earlier in the article, confessional disclosure is a key way female fashion and beauty content creators establish authentic intimacy in their videos, and construct a 'big sister' persona (Berryman and Kavka, 2017) that invites the audience to make a parasocial connection with them.
Among the YouTubers I followed, I selected the two who created this type of exclusively confessional/advice video, and who also had over 1 million subscribers each at the time of the study. The intention was thus to study how facework is used in professional content by highly successful and popular creators, not in diaries of amateur or semi-professional vloggers, nor artefacts by ordinary members of the public. Both these YouTubers speak on other public platforms to raise awareness on mental health, beyond their videos, alongside their main role as fashion and beauty influencers. They are well established in their industry, they are represented by major digital talent agencies and have commercial partnerships with major brands. The videos analysed in this article are therefore products by established professionals in the genre, they were made in a professional capacity, not in a private one, and they are not 'a diary posted in a place of perceived privacy' (Markham, 2018: 3).
The analysis does not focus on nor mention the actual confessions content creators make in these videos about their mental health, but on how they use facework to address their audience while making a confessional video. Markham (2018: 2) proposes that when analysing YouTube videos there is not a blanket approach to ethics, but 'contextual or case-based approaches' should be adopted depending on what line of action would diminish potential harm in the circumstances of each study. There is no potential harm involved in this study as no private or sensitive information is revealed in the analysis, which focuses on the discourse of professional media texts in the public domain. However to further ensure there is no possible impact at all, where excerpts are cited from the videos, aliases are used instead of the YouTubers' names (A and B for their names, followed by numbers 1-6 for the video the excerpt was from).
The analysis included all the YouTube videos by these two content creators published within two years (May 2019-April 2021), whose subject was exclusively issues around mental health and advice on how to deal with them. These criteria delivered twelve videos, six for each YouTuber, with a total length of 4 hours, 5 mins and 51 seconds.
A thematic analysis was carried out of the verbal content in these twelve videos and particularly of the points in the videos where the content creator addressed the viewer directly (e.g., through the pronouns 'you' and 'we', or verbs in the imperative form) or spoke of a potential viewer (e.g. through words like 'people', 'everybody'), because the focus of the study is on how facework is used to address an audience. These instances were all transcribed verbatim for the analysis.
Thematic analysis is a systematic method used to identify patterns of meaning across a set of data (Clarke et al., 2019). As a qualitative method, thematic analysis works with smaller sets of data than quantitative methods, and aims to deliver in-depth analysis of discursive patterns in this data, rather than generalisability to a broader population. Clarke et al.'s (2019) six-step process of thematic analysis was followed: the researcher first familiarises themselves with the data through repeated reading of the material, then codes the data by creating initial labels for different features that are relevant to answering the research questions. The researcher subsequently examines these codes to identify patterns of meaning or themes, and checks these themes against the data to ensure they represent it accurately, while refining the themes for this purpose. Finally, the researcher names the themes and presents them, providing the reader with the context needed to interpret them. To strengthen the reliability of the process, detailed notes were recorded during the development of the themes, which documented the connections identified between different concepts, as recommended by Nowell et al. (2017).
The theoretical concept of face (Goffman, 1955;Brown and Levinson, 1987), which was discussed earlier, guided the grouping of the thematic categories that were identified: I specifically looked for broader patterns in the YouTubers' discourse which extended fellowship to their viewers and others that protected their addressee's autonomy. However no pre-assigned variables were imposed on the data but the precise thematic categories (how exactly fellowship and autonomy were articulated in discourse) emerged inductively from the material itself. In what follows, the findings of this analysis are presented alongside excerpts of discourse from the videos.

Findings
The disclosures made in the analysed videos are embedded in discursive tactics that address both viewers' fellowship and autonomy face. The videos address the viewer on the one hand as a member of an established friendship group that the YouTuber hopes to help by making the video, and on the other hand as someone who might have objections that the YouTuber wants to prevent. The videos use a combination of discursive tactics to address viewers' needs for fellowship and autonomy; however, these needs are not articulated by the audience but anticipated parasocially by the content creator. By showing concern for their viewers' fellowship and autonomy needs, the content creators simultaneously protect their own positive face: by showing fellowship and reluctance to put viewers off, the content creators also present themselves as considerate and caring. The two sections below discuss and illustrate the four themes identified in the analysis of the discourse across all the videos.

Addressing Viewers' Fellowship Face
In the videos analysed, the content creators use consistently two main discursive tactics for extending fellowship to an imagined viewer: a) they say that they make their disclosures with an intention to help viewers or because viewers asked for this help; and b) they appeal to a community of viewers who follow them regularly, who know them well and who can support each other. These two themes were found across the analysed videos of the two content creators, and they are illustrated below with typical examples.
The first theme involves the YouTuber addressing the viewers' fellowship face by declaring an intention to help them. This help may result from the content creator's own initiative or as a response to requests they received from viewers on social media: 1. 'I think sometimes it's just like helpful maybe to see that you're not the only person going through something like depression and anxiety'. A6 2. 'I posted a Tik Tok a few months ago and it blew up, it did really well and I was not expecting it at all. [...] So in the comments for that Tik Tok there were lots of people saying that's nice but-great-but how? How, like what was your method, please share'. B7 Connecting with others who face similar problems is a common motivation for making and watching confessional videos (Gibson, 2016;Schuman et al., 2019) and the content creators here appeal to this need for solidarity. The content creator in example 1 addresses viewers' fellowship face by offering to help them feel included ('you are not the only person'). Raun (2018: 106) proposes that appeals which suggest the viewer is not alone in facing a particular situation communicate an impression of equality, by implying the content creator and the viewer are similar. This constructed sense of 'being on the same boat' with viewers they have never met further strengthens the effect of the facework and it invites a parasocial connection.
In excerpt 2, the YouTuber offers her confession as a response to her Tik Tok followers, who asked for details on how she overcame her personal problems, after she posted a video on that platform. The YouTube video was thus made because viewers asked for it. This tactic is not unique to confessional videos. It is common in fashion/beauty videos more generally, where content creators often claim that they make videos demanded by their audience, for instance because the audience requested a specific make-up tutorial. This tactic serves to construct an ongoing relationship between content creators and their viewers, where the latter contribute to editorial decisions by the former. In the confessional videos analysed here, it additionally helps to articulate the content creator's consideration for viewers' needs, which is a way of extending fellowship.
The content creator in example 2 addresses viewers' fellowship face by suggesting that viewers' wants are important enough to dictate which content she makes. In addition this also complements the negative facework content creators engage in, which will be discussed later in the article, and which aims to pre-empt negative interpretations of what the YouTuber is doing with these videos. By emphasising that they make their confession because viewers asked for it, they not only appeal to their addressees' fellowship face by giving them deciding power over their content, but they also protect their own positive face, representing their confession as responsive and helpful, not as a selfcentred imposition.
The second theme involves the YouTubers appealing to an ongoing relationship between themselves and their viewers as well as amongst their viewers. In this imagined parasocial relationship, the audience is already familiar with intimate information about the content creator: 3. 'In another way, and a lot of you know this, I made a promise to myself to have very strict boundaries about my personal romantical [sic] life. I've explained this so many times before'. A2 In example 3, the content creator speaks to a viewer who has been following her content consistently and who knows things about her that she has revealed in previous videos ('I've explained this so many times before'). Again the YouTuber appeals to the audience's fellowship face by extending inclusion, this time to a friendship group, who know the content creator through watching her videos regularly, even though this knowledge is not reciprocated. In previous videos, she indeed explained her decision to stop sharing her romantic life online, due to the impact this sharing had on her in the past. This may at first appear as refusing the viewers and inclusion. However, it is also a form of setting personal boundaries in an interpersonal relationship, which the viewers are here invited to remember and respect ('a lot of you know this'), because this historic relationship between them exists. Fellowship is being extended to the viewer, but this is done within the YouTuber's own terms. This is also the case in the example that follows.
The friendship group formed by the YouTubers and their viewers is sometimes constructed as a community that can also potentially provide its members with support: 4. 'If you have any advice as well, or like any different advice than I gave, that you want to put your two cents into, just leave a comment down below, you can chat about it down below if you want, and just help each other as well. I kind of want to use my channel as a place where we can relate to each other'. B11 The YouTuber in example 4 makes an explicit attempt to build community and interaction between members of the audience, by getting them to respond to each other rather than just to herself. She urges her viewers not just to comment on her video, but to share their own stories and respond to the confessions of other viewers. She presents her channel as a place where viewers may benefit from bonding capital, namely personal and emotional support derived from close relationships (Putnam, 2000). Although traditional social media have been found to increase users' access to information beyond their immediate environment, bonding capital is relatively rare (Ellison et al., 2010). The YouTuber here appeals to viewers' fellowship face, or their need to be included, and tries to instigate a network of support, which is social rather than parasocial.
At the same time, her call to action ambiguously implies either that she might participate in this network by responding to viewers' confessions ('a place where we can relate to each other') or that her contribution to it has already been made in the form of the confessional video, and now it's up to viewers to engage amongst themselves ('chat about it down below if you want, and just help each other'). In reality, the YouTuber contributes only twice to the comments section of the video, both times to thank someone for a compliment, and not to contribute any thoughts on the confessions that were posted by viewers. This lack of participation in the comments sets her apart from the network of fellowship she is trying to instigate. It also confirms previous research, suggesting that the interactive tools of YouTube are underused by content creators (Rihl and Wegener, 2019), thus limiting the potential they offer for social connection between them and their viewers.

Addressing Viewers' Autonomy Face
Confessional videos are part of constructing and branding YouTubers' online selves, so they need to be particularly aware of how these selves are received by the public they are addressing. In representing their 'networked me' (Chambers, 2013), they need to also anticipate and cater for possible negative responses. However, the affordances of social media mean that they cannot know who will view their content or in what context they will interpret it (boyd, 2010). In other words, they do not know how their videos will be taken, but their reception is still of vital importance to them and their brand.
This places content creators in a context with unknown variables, which the analysed videos manage by enfolding their confessions in mitigations. These anticipate and address imagined negative reactions, perhaps in an attempt to prevent them from even materialising. Content creators address their viewers' autonomy face by using consistently two discursive tactics: c) anticipating judgement or disinterest from viewers in what they have to say and d) pre-empting objections to the advice they give, or to other details of what they reveal in the videos. Again, both these themes were found across the videos analysed and are illustrated with typical examples in what follows.
When deploying both these negative facework tactics, content creators appear to presume that those who would view their confessional videos would be annoyed, angered, disinterested or would raise objections. The content creators' discourse is thus reflexively shaped by the possible reaction of its audience (Robinson, 2007), but this is essentially an imagined reaction to what the content creator sees as a potentially face-threatening imposition on their part. They therefore emphasise that they do not intend to impose and that they wish to protect their viewers' autonomy face: In these two excerpts, the content creators imagine a viewer who would judge their overall confession as annoying and who is not interested in their personal problems that they expound in the videos. They pre-empt imposition on this viewer by acknowledging the possibility of disinterest or dismissal of their whole confession ('you probably don't care', 'I don't know how much you want to know', 'people to think I'm just moaning'). In example 6, the YouTuber additionally acknowledges she may be seen as self-indulgent because she complains about the struggles caused by a job that is presumed to be privileged ('I get to sit in my room and film videos') and she uses irony against herself to pre-empt such a reaction ('what a hard life!'). Acknowledging possible negative reactions to her confession verbally prevents imposition on an imagined viewer, who may or may not have that negative reaction in reality.
The same attempts to protect their viewers' autonomy face were found in the two YouTubers' videos, when they were referring to specific parts of their own personal story or to pieces of advice that they give in their videos: 7. 'But thinking of kids going into it [being a content creator] with the level of negativity that is possible on here, that I know of, and the mental toll that it can take… So I am not saying for everyone it takes this, like you have to be resilient and I have been resilient a lot of my life' B8 8. 'I will say, like, this is just my experience, please talk to a doctor obviously, or talk to someone you trust' A1 In example 7, the YouTuber first makes a statement about her own experience as a content creator ('the mental toll that it can take') and immediately qualifies it with a response to an anticipated contrary statement or objection ('I am not saying for everyone it takes this'). She thus responds to an imagined proposition that she is making a generalisation about the mental toll her job takes, and that content creators in general ('like you have to be resilient') and she specifically ('and I have been resilient') needs to be able to deal with this toll. The content creator here acknowledges the possible objections of someone who disagrees with her point, by having an imaginary, parasocial conversation with this imagined viewer.
Finally in example 8, the content creator pre-empts negative reactions by clarifying she is not speaking from a position of medical expertise but from personal experience with the problems she discusses ('this is just my experience'). She pre-empts potential objections to whether she is qualified to give advice by urging viewers to seek medical help if they are affected by these problems ('please talk to a doctor obviously'). In this way she prevents imposition on a sceptical viewer.

Conclusion
Confessional videos are a way for YouTube influencers to mobilise communicative intimacies with their viewers (Abidin, 2015). Presenting themselves as vulnerable and confiding in the viewers (Jeslev, 2016) helps them construct an authentic online persona (Harrington, 2019) and ultimately it breaks down the distance with their audience and makes them appear ordinary and relatable (Marwick and boyd, 2011). At the same time, offering advice to those facing similar situations reinforces female influencers' 'sisterly' role in relation to their audience (Berryman and Kavka, 2017). However, this article has shown that at the same time as communicating intimacy and vulnerability, these videos also tread very carefully around viewers' sensibilities, as these are understood by the content creators. The videos consistently take care to protect viewers' fellowship and autonomy faces, on the one hand by inviting them to feel included in parasocial community, and on the other hand by pre-empting any imposition or disagreement they could cause.
In the videos analysed in this article, influencers speak with the anticipated reactions of their imagined viewers in mind, they affirm their intention to help them with their problems and to offer advice that they have asked for, while encouraging them to support each other. In addition to this, the influencers also take steps to prevent imposition on viewers and pre-empt anticipated negative reactions or objections.
This prominence of facework in confessional videos might initially appear counter-intuitive. In a social relationship, intimacy is associated with disclosing personal moments and feelings in a context of vulnerability and trust (Schwarz, 2011). Facework does not clearly fit into our understanding and experience of intimacy in everyday life. However, the relationship between influencers and their YouTube viewers is not a social one, and YouTubers have to do a fair amount of parasocial imagining of their audience and of who may be in it. It is not possible for them to know how what they say will be taken because they do not know the viewers' personally and cannot interact with them while they are watching these videos. And yet, how what they say may be taken matters: one of the purposes of making these videos is to contribute to building content creators' personal brand. Content creators therefore need to prepare the ground for their confession by mitigating it in discursive tactics that predispose the viewer positively and prevent any potential negative reactions.
The content creators studied in the article thus construct their intimate online selves reflexively through interaction with their audience, but also parasocially by imagining their audience. As Robinson puts it, 'the "I" […] anticipates the cyberother's reaction, thus creating the "me" ' (2007: 104), so the self that is presented online is a construct, which is shaped partly by second-guessing.
Clearly the content creators' success relies on the parasocial relationship they are able to establish with their audience, so they invest significant affective labour to mobilise and maintain it (Raun, 2018), and they use it strategically for self-commodification and branding (Harrington, 2019). Confessional videos, like the ones analysed here, play an important role in breaking down barriers and status inequalities between them and their audience and making them appear more ordinary and relatable, more like friends than like internet celebrities.
At the same time though, the analysed videos are not a raw display of emotion as might be the case in an intimate confession within a social relationship, but rather they offer carefully controlled and mitigated confessions. The discourse in the videos takes care not to offend, annoy, or trigger objections, at the same time as including viewers in an intimate parasocial friendship group, to whom the influencers respond with helpful and supportive advice.
This discourse of care, support and non-imposition may be gender-specific, as both analysed content creators are female and their YouTube channels focus on fashion and beauty content mainly for female viewers. It is possible that YouTubers who do not create such feminised content deploy facework in different ways. Moreover, the level of success the analysed content creators already enjoy and the persona they have established may be factors in how they present themselves in these videos: these are YouTube professionals in the fashion and beauty genre, with over a million viewers each, and they are potentially particularly careful in addressing viewers in a way that supports and does not damage their already established personal brand.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.