Abstract
This article analyses the development of the participatory online social movement and health intervention Hello Sunday Morning (HSM) from 2009 to 2016. HSM has developed a media platform comprised of blogs, social networking and self-tracking apps with the aim of changing individual and cultural drinking practices. The case of HSM enables us to reflect on the participatory culture of digital media platforms. I argue that participants don’t just construct the sociality of the platform – its practices of speaking and being heard – they also play a crucial role in generating exercises that are coded into the platform’s interfaces, protocols and algorithms, and they produce data that informs the platform’s ongoing development. The article draws on interviews with participants, analysis of the platform since 2009, and my own critical reflections on participating in research, evaluation and design activities.
Introduction
At the beginning of 2009 a young Australian named Chris Raine began an experiment. He was working in an advertising agency on a campaign against youth binge-drinking, when he realised none of the ideas he pitched would change his own behaviour. In a profile in The New Yorker Ceridwen Dovey (2015) explains: This got him thinking: Why had it become socially acceptable for his Gen Y peer group to spend every weekend alternating between being drunk and hungover? He wondered what it might be like to spend time in what he calls ‘that vast unknown land of sobriety.’ After a breakup, and a blowout New Year’s Eve, he decided, at 2 A.M. on January 1, 2009, to give up drinking for a year.
Raine catalogued his experiences on a blog called Hello Sunday Morning (HSM). The blog recorded a self-devised and directed experiment with his identity, body and health.
The blog, I argue, was also the beginning of an experiment with the symbolic and technical qualities of digital media. He was writing a symbolic narrative about drinking culture in Australia from his own vantage point as a young Australian man. The blog had value as a public form of expression that gave voice to the way alcohol consumption was woven into his cultural experience. At the same time, the blog was a technical device for recording, reflecting upon and regulating the body (Schull, 2016). In this respect, the blog followed a longer history of analogue technical devices used to track the body such as diaries, scales, and wristwatches (Crawford et al., 2015; Schull, 2016). Since 2009 HSM has grown from Raine’s solo blog into a digital media platform and health intervention with over 110,000 participants.
This article tracks HSM’s development since 2009 as it moved from a single-author blog, to a network of blogs, to a bespoke social networking site, to a social self-tracking app. HSM’s development over the past decade enables us to reflect on the ongoing construction of the participatory culture and platforms of digital media. Participation in health-related blogging and online communities has been conceptualised as the affective labour of producing online sociality (McCosker, 2017; McCosker and Darcy, 2013). Following this way of understanding health-related blogging, HSM participants are affective labourers who jointly constructed the sociality of an online platform by producing narratives, images and interpersonal support.
In this article I argue that by analysing HSM’s development we can discern how the participatory culture of digital media produces not only opportunities to speak and be heard, but also shared exercises, practices and routines (Couldry, 2012; Gregg, 2018). Much debate about the participatory qualities of digital media have tacitly imagined it as a public forum of symbolic expression and debate (Carpentier, 2011; Jenkins and Couldry, 2014; Schafer, 2011). I aim to give emphasis to those dimensions of our participatory culture where we experiment, reflect and instruct each other on how to live, and how to look after the mind and body. Digital media are participatory in the sense that they are not just a forum of debate, they are also a gymnasium of daily exercises (Gregg, 2018; Sloterdijk, 2014). By examining how HSM participants produce shared practices, I argue we can understand an important, yet often overlooked, dimension of participants’ affective labour on digital media platforms. Participants don’t just create the sociality of the platform – narratives, interpersonal support and recognition – they also play a part in generating the practices, exercises and techniques that constitute HSM as a digital health intervention or self-improvement technology. In this way, we can think of participation as a crucial element in the production of platform interfaces, protocols and algorithms. The work of ‘programming sociality’ (Bucher, 2018) is a process that unfolds over time, drawing on the creative open-ended expression and deliberation of users as creators of narratives, practices, affects and data. A programmed sociality is a dynamic production wherein sociality is both steered by platforms and an integral resource in the creation and sustaining of the platform.
I characterise HSM’s development over the past decade in three phases. While each of these phases corresponds with a period in HSM’s institutional and technological development (blog, social networking site, app), more importantly each phase is also distinguished by a change in the role participants play in the ongoing development of HSM’s participatory culture, practices and platform. The phases describe HSM’s process of becoming a platform. The product of their participation, I argue, is both symbolic and technical. The narratives and exercises they shared with one another double as data that inform the ongoing development of the platform itself.
Phase 1: Devising a social experiment
The first phase was a sole-author blog begun in 2009 by the founder Chris Raine. This phase continued in 2010 when Raine invited peers to replicate his experiment, out of which emerged a network of affiliated bloggers. During this phase participants used blogs as a technology for agreeing on how to ‘do’ an HSM as a practical, reflective, social and mediated set of exercises. In this phase we see the beginning of HSM’s ‘platformisation’, before it could materialise as a purpose-built social networking site or app, it needed to be imagined as shared practices and protocols.
Phase 2: Programming platform sociality
The second phase (from 2010 to 2015) began with the development of a custom-built blogging and social networking platform. By 2012, this platform had many of the characteristics of social networking sites, including user profiles, the capacity of users to ‘follow’ one another, comment on each other’s posts, and some algorithmic recommendation of content and connections to users. From 2012 onwards, more structured forms of participation were introduced to this platform, such as offering users tools to log their goals, alcohol consumption and mood, and for them to see data visualisations of their progress. Many of the practices established on blogs in phase one were built into the platform architecture in phase two. It is in this phase that HSM meets the threshold requirements for being a ‘platform’ because it began to collect and analyse user data, and use that data to inform the shape of platform features (van Dijck et al., 2018).
Phase 3: Training participants and algorithms
The third, and current, phase, beginning in 2015, is marked by the development of two mobile apps. The first app was a fundamental break with the blogging-oriented participation of the social networking and blogging website, it marked a strategic shift away from open-ended expressive participation. Using the app, participants logged their alcohol consumption and mood on a weekly basis, set challenges that they aimed to complete and viewed data visualisations of their progress relative to others. This app was discontinued in late 2015 and replaced with a more clinically focused app allowing participants to share their progress in short comments, log their mood, post their thoughts, undertake algorithmically recommended cognitive and behavioural exercises, and have access to online coaches.
The first two phases (2009–15) of HSM were organised around user-generated blogs where participants documented, and reflected on, a self-directed and social experiment. From 2012 onwards data-driven platform features and mobile apps have played the role of collecting data about alcohol consumption, processing it, visualising it and feeding it back to participants. This has both oriented participation toward the collection and visualisation of data, and assisted HSM’s efforts to evaluate the effectiveness of their platform as a health intervention. Throughout these phases, HSM has variously described itself, and been described by others, as a blog, a website, a social networking site, a health intervention, a social movement, a brand, and an app. This ambiguity about the form HSM takes is an important part of the story; it highlights its lively experimental character. Furthermore, it highlights the mutability of media platforms. Part of the value of studying HSM over a 10-year period is to observe how its form, techniques and modes of participation change. HSM highlights how media platforms and their participatory culture are in a continuous state of co-production.
HSM’s combination of symbolic expression and logging information about the self is similar to other uses of digital media to experiment with, and document efforts to reduce, alcohol consumption like Dry July and Feb Fast (Robert, 2016). Beyond alcohol, HSM is similar to the larger culture of using digital media to document efforts to change the mind and body, from wellness blogging to fitness and mood trackers (Lupton, 2013, 2014; McCosker, 2017; McCosker and Darcy, 2013). Like other social and mobile media health apps, HSM is a sociotechnical ‘configuration’ (van Dijck and Poell, 2016), ‘underpinned by tacit assumptions, norms and discourses already circulating in the social and cultural contexts in which they are generated, marketed and used’ (Lupton, 2014: 4). Part of what makes HSM a significant case is the combination of open-ended forms of expressive symbolic participation like blogging with technical data-driven forms of participation in experiments like goal-setting, self-tracking and algorithmically recommended cognitive-behavioural exercises. These symbolic and technical data-oriented modes of participation are interrelated. For instance, the symbolic narratives of the blogs doubled as rich data that informed the design of the platform, and the data participants log is used to produce symbolic visualisations of their progress.
I contacted Hello Sunday Morning’s founder Chris Raine in 2009 because I was interested in his attempt to use blogs to animate a creative process of deliberative exchange among mostly young Australians about drinking culture. My original question was: Did Hello Sunday Morning illustrate the practices of speaking and listening that were being attributed to the participatory web (Couldry, 2010, Crawford, 2009)? I interviewed Raine and then undertook 16 open-ended interviews with the first wave of bloggers during 2010. I was then a participant in planning and design workshops, meetings and conversations from 2010 to 2015. In 2010, HSM provided me funding to undertake a semi-automated text analysis of 13,000 blog posts from 2900 participants. In 2014, I collaborated with researchers to undertake a behavioural and demographic analysis of HSM participants (Carah et al., 2015). In 2015, HSM provided funding to develop an evaluation method for their website and mobile app. I outline this several-year engagement with HSM here as a necessary declaration of my insider-outsider status as a researcher. The aim of this article is not to present the findings from this range of research activities, or to evaluate the effectiveness of HSM, but rather to present a critical and reflective analysis of how HSM developed as a media platform. I draw on interviews I conducted with early HSM participants, my own observations of the HSM platform, and correspondence with HSM to confirm, or add detail to, my observations. This article is primarily my own critical account of how I came to understand HSM’s development, and my argument about what it teaches us about the development of digital media platforms and their participatory culture.
In this article I explore how the participation of users informed the development of HSM as a media platform through each of its phases. First, I begin by drawing on the interviews I conducted with the first wave of participants during 2010. I use these interviews in part as a reflective anchor point. They offer a record of how participants during HSM’s most open-ended phase understood the platform and its affordances. I outline how the blogs were used as tools for self-expression, reflection and experimentation. Second, I reflect on how this open-ended participation was gradually formatted and structured via the development of platform protocols by HSM. Third, I consider the development of HSM’s app, which formats participation around algorithmically recommended experiments. I conclude by proposing that HSM illustrates the integral role that participatory culture plays in the development of media platforms.
Phase 1: Using digital media to devise a social experiment
During 2009, Chris Raine’s individual Hello Sunday Morning blog attracted a following of over 1000 people with its reflective discussion of social interactions like dancing sober in nightclubs or asking people out without a drink. This was an experiment in the sense that Raine took pre-defined action, observed, recorded and analysed the results, and then opened them up to peer scrutiny in a public forum. At Raine’s invitation, several readers of the blog joined in and did a ‘Hello Sunday Morning’ of their own. Participants would spend an agreed period – usually three or six months – without drinking and blog about their experiences. By the end of 2010, 77 people had written a post on a personal ‘Hello Sunday Morning’ blog; by the end of 2015 HSM reported they had 40,000 registered participants; in early 2019 their website reports 110,000 participants. HSM is difficult to classify as health promotion, intervention or treatment. When it operated primarily as a blogging platform it challenged dominant health promotion approaches by relying on peer-to-peer forms of communication; and, while participants claimed treatment benefits, the program was generated outside a clinical setting. Since 2015, HSM has moved toward the approaches and structures of clinical health interventions (Kirkman et al., 2018).
The accounts of HSM bloggers during the first phase (2010) can be understood as both the process of speaking and listening to one another and of shaping a larger material structure. That structure was at first a series of socially constructed agreements about their practice made within the affordances of blogs. Over time, those agreements were coded into HSM platform protocols (in the subsequent phase of development from 2010 to 2015). The blog posts that participants wrote animated peer-to-peer exchanges. The posts were also archived in a database where they were later used as data that documented HSM practices that could be codified into the platform design.
In the accounts that follow, participants describe their participation mostly in symbolic terms of speaking and being heard. This is a function of the time at which the interviews were undertaken, before the platform had designed any protocols beyond the open-ended blog and before HSM had used any user-generated material for research and design purposes. This meant that neither myself as interviewer nor participants specifically discussed the role their participation might play in the larger development of HSM as a platform, nonetheless these participants described their role in using blogs to shape HSM as an agreed-upon set of practices.
Most participants were first-time bloggers exploring the technical capacities and cultural possibilities of a new medium. Together they established norms around expression and deliberation that were largely located within the technical and social affordances of blogging on WordPress and, in many cases, posting those blogs to personal Facebook profiles. Participants described HSM as a ‘social experiment’ where they learnt ‘a lot about how people felt about alcohol’ (HSM blogger 1). The HSM blog was a key tool in enabling participants to reflect on how their actions exposed socially constructed norms around drinking to themselves and their peers. As they engaged in these acts of open-ended reflection and deliberation they were also forming protocols that could be coded into the media platform. For example, there was much debate among early bloggers about whether a period of three months abstinence was necessary to ‘do an HSM’. Some participants wanted to participate by moderating their drinking rather than abstaining or by undertaking a shorter period of abstinence. Participants also deliberated over whether a ‘slip up’ by having a drink meant a participant had to reset their period of abstinence. While these discussions on the blogs functioned as deliberations among a small network of participants about the rules of participation, they gradually became coded into the platform infrastructure. By 2012, when new participants signed up and hit the ‘start’ button on the platform they were offered one of two options: a ‘3 month’ or ‘12 month’ HSM. And, later, they could also choose to declare an intention to ‘moderate’ or ‘abstain’.
The value of HSM for the bloggers I interviewed was for them to reflect on, change and learn how to articulate their changing sense of identity. Their individual HSM blog was a key tool in enabling them to reflect on the uneasy relationship they had created between themselves and their peers. The blog offered a space to contemplate and examine their identities, cultural world and aspirations. The act of stepping outside prevailing cultural norms could cause real friction within their peer groups. Several bloggers described how they ‘copped a spraying’, a colloquialism for being verbally abused, when they announced their participation on Facebook. In part, these reactions demonstrated how drinking was embedded within social relationships. But importantly, participants told these stories as part of demonstrating the value of their HSM blog as a space where this friction could be expressed and worked through. Pennay et al.’s analysis of HSM blog posts (2016) and Fry’s (2014) analysis of blog posts and interviews with HSM bloggers made similar findings. And McCosker’s (2017) analysis of posts to BeyondBlue online anxiety and depression support forums also offers comparable insights into the value of peer support on anonymous online forums.
The results of this reflection were evident in the accounts participants offered about their drinking culture. A blogger (HSM blogger 5) explained that: from where I stand it looks like a lot of young people, especially young males don’t have any form of initiation into the world of manhood. So they see that initiation as being turning 18, getting a license, being able to go out and drink, and there is nothing, they’re lost. I agree that I was lost.
A participant (HSM blogger 6) also referred to the need for a ‘rite of passage’ into adulthood. HSM blogger 2 related the experience of growing up in a country town where ‘the only thing to do really was get drunk or smoke cones’ where no one ‘presented me with a choice’. HSM did not offer a prescriptive set of values, but rather a technocultural infrastructure through which these participants could tell stories, ask questions, and express how they feel in their own language. Participants’ capacity to ‘bounce ideas’ and ‘build a path’ for each other (HSM blogger 6) was afforded by the technocultural form of blogging as the act of writing open-ended and self-reflective expressions, and anonymously posting them to a relatively closed social network where others would read them.
This first wave of participants was crucial to forming HSM as a set of practices that could be ‘scaled up’ and codified in the development of a HSM platform. Participants’ blogs theorised how HSM worked and created a set of norms for ‘doing’ an HSM. Participants’ blogs are records of, and reflections on, their actions. For example, a key convention to emerge was setting a goal. Chris Raine encouraged early participants to state goals in their first blog post and this became a convention across the platform. The goals that emerged included getting fit, saving money, or getting better grades. These shifted HSM from saying ‘no’ to alcohol to also saying ‘yes’ to something else and gave participants something positive to blog about in relation to their critical reflection on drinking culture. The goal helped to format the use of the blog as a positive record of accomplishment.
Participants explained that the practice of writing and reading helped them to reflect and develop their own voice. Evident in their accounts is that voice is both an individual and social accomplishment (Couldry, 2010). HSM blogger 7 felt a sense of shared expression and experience: ‘some of [the blogs] are just like so striking and they hit a nerve and hit emotions that you completely completely understand and y’know, that’s what’s so great about that blogging atmosphere.’ Their voice had value where it enabled them as an individual to create an identity not dependent on the drinking culture of their peer group. That value was located in the shared practices they created. By publishing their blog posts they contributed to the development of a technocultural formation that they recognised in the form of HSM as a movement, defined practice, brand and platform. The joint establishment of conventions among participants during this phase of HSM’s development illustrates Couldry’s (2010) account of voice as being able to give an account of oneself that is reflected in the material structures within which we participate.
HSM participants who wrote blogs in its first five years played a crucial role in creating, describing, standardising and defending the protocols of ‘doing’ an HSM. They collectively gave shape to the experiment by defining its practices. They debated how long a participant had to go without drinking, deciding if a participant who ‘slipped up’ had to start over, deliberated over whether you could ‘do a HSM’ if you didn’t record your progress in blog posts, and collectively produced a series of goals that, by extension, defined forms of self-improvement. In doing so, they generated both rich descriptions of practices that informed the choices designers and developers make, and data that is used to track engagement, evaluate features and codify algorithms. The social agreements that emerged among participants about how to ‘do an HSM’ gradually became buttons, dashboards, visualisations, drop-down lists of choices and algorithms. The ‘programmed sociality’ (Bucher, 2018) that began to emerge on HSM came, in part, from participants using blogs to create and record techniques for individual and cultural change.
Phase 2: Programming platform sociality
In the accounts above HSM participants describe a period in which participation was devised by users in open-ended ways. They were using blogs to experiment, explore and socialise practices with each other. Couldry (2012: 35) suggests that we should attend to what people are doing in relation to media. We should examine how their practices build a shared sense of ‘living in a world’ by constructing routines and relationships that address our needs and express our values. We use digital media to organise our life and relationships, search for information, and express ourselves – to search, show, presence and archive (Couldry, 2012). Digital media construct opportunities to express ourselves, collect and archive those expressions, and draw on them to inform the development of platform features and protocols. But, that doesn’t necessarily mean we have agency. We can see our voices reflected back to us in the form of digital media in ways that may or may not value our voice in the sense of giving us agency over the platform (Couldry, 2010, 2012; Dean, 2010). Our practices then are not just things we do with media, they are also things we do to media. Participants can come and go from particular apps and platforms without having a stake in how the platform develops. Regardless of whether their voice is ‘valued’, as they log in and log out, participants leave behind their practices as data that inform the decisions of the developers and owners of platforms (Dean, 2010; Schafer, 2011).
HSM was like ‘most web 2.0 platforms [that] started out as indeterminate services for the exchange of communicative or creative content among friends’ and over time developed more strategically programmed forms of participation (van Dijck, 2013: 6). In van Dijck’s (2013; van Dijck et al., 2018) formulation, a platform is ‘fuelled by data, automated and organised through algorithms and interfaces, formalised through ownership relations driven by business models, and governed through user agreements’. In HSM’s case this process involved translating the practices that were collectively devised in phase 1 into flows of data, interface features, and recommendation algorithms in the subsequent phases. HSM’s design process was informed by strategic objectives and frameworks drawn from a mix of branding, health promotion, clinical intervention, and the ‘agile’ culture and methods of tech start-ups. From 2010 onwards, when it first acquired funding from government philanthropic and corporate organisations, HSM became an organisation with the strategic goal of recruiting participants and changing drinking culture, it began to develop a platform ‘programmed with a specific objective’ (van Dijck, 2013: 6). Like commercial platforms, that objective was to grow user engagement. HSM’s shift from an open-ended blogging platform to a self-tracking app can be read within this frame. In one sense, the shift from self-expression to self-tracking more broadly reflects changes in how we express and log information about ourselves using digital media. But HSM’s shift from ‘narratives’ to ‘interventions’ also reflected their strategic objectives to simultaneously build user engagement and generate data that demonstrated changes in participants’ drinking behaviour.
HSM’s three phases illustrate how platforms are inherently unstable, editable and programmable (Helmond et al., 2019; van Dijck et al., 2018). Platforms are engineered to channel, and constrain, our practices (Alaimo and Kallinikos, 2016; Srnicek, 2016; van Dijck et al., 2018), carefully organising participation ‘along specific activity corridors (such as sharing, following or tagging) that heavily stylise and shape user interaction’ (Alaimo and Kallinikos, 2016: 176). As ‘programmable digital architecture’ they ‘organise interactions between users – not just end users but also corporate entities and public bodies’ (van Dijck et al., 2018: 4). In HSM’s case, the development of the platform was informed by early investment and in-kind support that came from social innovation, government, civil society and public health organisations. HSM also ‘evolved’ within a larger platform ecosystem. For example, by initially building on the WordPress platform or, more recently, partnering with Google to improve the recommendation algorithms of their app.
Programming HSM as a platform involved codifying the open-ended participation of the early waves of bloggers as a set of platform protocols. Following van Dijck’s (2013: 31) framework, ‘governing protocols provide a set of instructions that users are forced to obey if they want to partake in the mediated flow of interaction’. These protocols are made visible to users via interfaces that contain ‘technical features (e.g. buttons, scroll bars, stars, icons) as well as regulatory features (e.g., the rule that a personal profile is required before entering the site), and these features actively steer connections between users and content’. In the accounts above participants articulated the role they played in the deliberative invention of HSM as a set of practices for intervening in both individual drinking behaviours and a larger drinking culture. These practices were reproduced via the act of writing and sharing blog posts.
From 2010, HSM gradually began building a bespoke blogging and social networking platform. From 2010 to 2012 this platform was hosted on WordPress and so followed the basic structure of a blogging platform. In this period, the first way in which the platform programmed sociality was by establishing an official space within which HSM participation took place. To participate, users had to establish a profile on the HSM site. This is important because it meant that HSM would not become a set of portable practices that users could undertake on their own social media profiles, but rather practices that would be confined to a specific media platform. Doing an HSM was sociotechnical in that it meant not just undertaking a ‘social’ activity, but doing that by establishing a profile on the HSM site and recording progress there. Locating participation within a custom-built platform enabled HSM to monitor engagement, collect data from participants, manage and experiment with HSM as a health intervention and a lifestyle brand.
In 2012 HSM transitioned from the ‘off the shelf’ WordPress site to a purpose-built web platform using Ruby on Rails (personal communication with HSM designers, 2017). This version of the platform enabled HSM to develop a series of protocols that governed participation in designated ways. These included setting the length of an HSM, nominating goals from a pre-loaded list, weekly check-ins where participants logged mood and alcohol consumption, and algorithmic recommendation of content. These platform protocols emerged as a solution to a strategic problem. As HSM surpassed several hundred participants, mutual interaction via the blogs became unfeasible. As in most online communities, the majority of blog content began to be produced by a minority of highly engaged participants. Most participants only wrote one blog post when they first began their HSM. Although this pattern of participation is not uncommon in social networks, it produced strategic challenges, which were addressed in the development of several platform features, and eventually resulted in developing a mobile app-based platform not organised around blogging. Furthermore, as the group of participants grew larger and more heterogeneous they became less able to narrate shared protocols of participation to one another. And so platform protocols, rather than shared narratives, became the key device for structuring participation.
One problem created by the majority of posts being made by first-time bloggers on the platform was that the chronological feed of content users saw when they logged on was repetitive. Organised chronologically, the home feed was mostly posts from first-time bloggers announcing their intention to do a HSM. These ‘announcement’ posts were not engaging to other users. With the 2012 rebuild, HSM developed algorithmic recommendation of content and user connections to counter this problem. When users logged onto the platform they could choose between ‘popular’ and ‘your wave’. The ‘popular’ feed featured posts by bloggers who generated high engagement in the form of likes, bookmarks and comments. The ‘your wave’ feed generated a customised set of engaging posts written by bloggers at a similar stage of their HSM. This algorithmic recommendation reinforced HSM practices and protocols in various ways. The popular feed meant that highly engaging bloggers were more widely read, and their practices were therefore more regularly adopted as norms of participation. The ‘your wave’ feed offered bloggers a standardised chronological narrative of an HSM for them to follow. The idea for this feed came from semi-automated text analysis of blogs that indicated a typical structure for an HSM narrative (Carah et al., 2017). For instance, if the first blog post set out intentions and goals, posts during the second month of an HSM might reflect on efforts to change, and posts toward the end of an HSM might sum up achievements and offer advice to others. This feed acted as a template for how participants might then write their own posts at each stage of their HSM.
To develop structured practices that would generate engagement and data from the majority of participants who did not write blog posts, HSM codified goals and created a weekly ‘check-in’ protocol. Figures 1 and 2 illustrate how goals were codified on the HSM website by 2014. Figure 1 illustrates how goals were recommended to participants at sign up, and Figure 2 illustrates how goals were ‘suggested’ to participants once they had signed up. These goal ‘formats’ emerged via the open-ended expressions of participants. In the first phase (2010-2012) participants described their goals in open-ended ways in their blog posts. During 2012, a list of goals was generated inductively using semi-automated text analysis, and these were rendered into a pre-loaded list of goals users could select from. By 2013, in the iteration visible in Figures 1 and 2, the goals also included ‘social’ elements. Users could see how many users have selected that goal, and once selected are more likely to see posts from users with whom they share that goal in common. This enabled users with high affinity to see and engage with each other on the platform.

Goal dashboard (2014).

Individual goal with social engagement from users (2014).
In 2014, HSM developed a weekly check-in (Figure 3). The check-in was the first device that structured a regular form of participation that was not blogging. A notification was sent to participants each Sunday asking them to complete a short questionnaire on their alcohol consumption and their mood. Their progress was then displayed to them in a data-driven visualisation that both demonstrated change in their mood and alcohol consumption over time, and positioned their alcohol consumption relative to population norms and health guidelines (Figure 4). The ‘check-in’ made participants who did not blog visible to the platform and generated dependable flows of data. By drawing on measures from established alcohol consumption tests the ‘check-in’ enabled HSM to evaluate user progress against public health norms (see Kirkman et al., 2018).

Visualising consumption in the check-in (2014).

Giving users standardised advice (2014).
In 2015, with support from the Vodafone Foundation, HSM created a mobile app to engage users who never blogged but accessed the website via mobile devices for the weekly check-in. The design for the app emerged from a workshop with users, health, technology and design experts at Google in late 2014. The app represented a deliberate move beyond blogging as the primary form of participation. While the blogging platform remained open, it was no longer actively promoted or developed. The app opened with a standardised questionnaire about alcohol consumption and then used the profile of the app user to algorithmically recommend activities like ‘go to the gym’, ‘go on a sober date’, or ‘catch up with a friend for breakfast’. The app sent a push notification ‘check in’ to users each Sunday to answer a brief questionnaire about consumption, record completed challenges and nominate new ones. The app visualised their consumption and wellbeing in graphs, and aggregated the data about all users to evaluate the efficacy of the platform (see Figures 5–7).

Logging mood on the app’s weekly check-in (2015).

Seeing a visualisation of your drinking on the app (2015).

Seeing a visualisation of your mood over time on the app (2015).
The app attempted to engage young binge drinkers by structuring participation around social challenges that targeted their lifestyle. The app did attract these younger participants at first, but after a few weeks, these participants stopped engaging. The active participants on the app began to reflect the older dependent drinkers from the website. In HSM’s view, these users started ‘hacking’ the app to match the forms of participation they had developed on the website. They ignored the challenges themselves, but used the ‘comment’ protocols of the challenges to conduct their own conversations with one another (personal communication with HSM designers, 2017).
Phase 3: Training participants and algorithms
Toward the end of 2015 HSM closed the first app, and set about designing a new app called Daybreak, which launched in September 2016 (Tait et al., 2019). Daybreak deliberately targets high-risk and dependent drinkers, rather than young binge drinkers. The app is organised around mood-tracking, cognitive-behavioural therapy and online counselling. The focus on mood comes from HSM’s internal data analysis of user participation on earlier web and app versions of their platform, which demonstrates that participants who improved their mood were more likely to sustain reductions in alcohol consumption (personal communication with HSM designers, 2017).
When Daybreak users check in they see four icons: a yellow sun, a green sun behind a cloud, a blue rain cloud, and a purple cloud with lightning bolt (Figure 8). Users ‘tap’ one of the icons and are invited to ‘share’ a thought that is posted to a feed in the app coloured with the ‘mood’ icon the user selected. The app explains that ‘sharing your mood . . . will make Daybreak a lot more useful’. Once users enter the app they can access a feed of ‘shares’, which are like tweets, a list of recommended ‘experiments’, a personal profile with their set ‘goal’, and a facility to talk to an online ‘coach’ who is a qualified counsellor. The app is free in Australia, where it is now funded by the Department of Health. Users outside of Australia must pay an annual subscription fee.

Visualising mood and ‘sharing’ how you feel on the Daybreak app (2016).
‘Shares’ function as both symbolic expressions and data within the app. At the level of symbolic expression they are posted in a feed that other users can read and respond to. The choice of mood icon and share generate an algorithmically recommended ‘experiment’ for users to undertake. The ‘experiments’ are cognitive-behavioural therapy exercises written by a clinical psychologist. These experiments are recommended based on user profile, expressions, alcohol consumption level and mood. For instance, if the user selects the purple ‘thunderbolt’ icon and then uses the words ‘anxiety’ or ‘worry’ in their share, the app might algorithmically recommend a deep-breathing or mindfulness ‘experiment’ (personal communication with HSM designers, 2017). The development of a deep learning model that makes these algorithmic recommendations is funded through a Google Impact Challenge grant.
HSM describe the aim of this partnership with Google as follows: to ‘build a deep learning engine to determine the precise cognitive-behavioural therapy solution a person needs to deliver it at precisely the right time. We propose to do this by teaching a supervised learning model using the text, location, mood and outcomes data we collect from the Daybreak app, and fine tuning it over time’ (personal communication with HSM designers, 2017). This deepens the use of algorithmic recommendation to shape user experience. Where the algorithmic recommendation of blog posts on the social networking site used basic criteria set by programmers, the recommendation engine of the app draws on a larger range of data points relating to individual users (mood, location, alcohol consumption history, and symbolic expression) and learns over time. The feed of ‘shares’ in the app is customised to link together users who are predicted to ‘positively influence one another’ (personal communication with HSM designers, 2017). As participants engage with the app they generate data which trains those recommendation algorithms. The app, in comparison with the blogs that preceded it, is more deliberately oriented toward training the participants.
Participation as voice, practices and platforms
When I first began observing HSM, and interviewing the early participants, I thought they were using digital media to produce social relationships, identities and cultural practices. Over time I found myself revisiting this question of ‘what are participants doing with digital media?’ In the case of HSM, the products of their participation seemed to extend beyond the forms of sociality their written expression created, they also produced exercises that became codified features of a platform. We produce voice where we use digital media to give expression to our experience, to narrate our feelings and relationships, and to shape the material conditions of our lives (Couldry, 2010, 2012). We produce practices where we jointly create exercises through which we can improve ourselves and our social world (Nafus and Sherman, 2014; Schull, 2016). Participatory digital cultures are ‘acrobatic’ and ‘athletic’, to borrow Sloterdijk’s (2014) terms, as well as ‘expressive’ and ‘symbolic’. We produce platforms where our expressions and practices generate sociality, double as data that inform the development of features and protocols, and are employed in the ongoing training of classification and recommendation algorithms.
In the case of HSM, the blogs, social networking site and apps that have emerged and morphed are at each stage a product of the dynamic process through which actors produced an experiment – a set of practices, rules and modes of evaluation. Those actors include: HSM and their team of designers and programmers along with clinicians, government, corporate and philanthropic funders who set expectations of what good forms of participation and impact look like. And, at each stage, users who make ‘implicit’ and ‘explicit’ (Schafer, 2011) contributions to HSM as narratives, exercises and a technology.
When participants in the first phase debated with each other and the HSM founder Chris Raine, about the ‘definitions’ of HSM, they were explicitly contributing to its design. At the same time, some of the contributions made by participants are implicit, in the sense that they are made through the ordinary use of the platform (Schafer, 2011). For instance, while users explicitly criticised HSM for introducing gamification features in phase 2, the data they generated by using them indicated to the developers that they did actually increase engagement with the platform. Or, when Daybreak users express their feelings and then undertake recommended clinical exercises, their data gradually accumulates and fine-tunes further recommendations.
Each of these actors – HSM’s founders, programmers, designers, users and funders – were contributing to HSM’s platformisation within the larger evolution of digital media (from blogs, to social networking sites, to self-tracking apps), imagining its affordances for ‘prevention’ and ‘treatment’. The shift toward ‘treatment’ was informed by a combination of the availability of funding for treatment over prevention, together with the ability to codify and report on treatment outcomes, and participants increasingly choosing to use HSM as a treatment technology.
Participants in HSM’s first phase exercised some agency by creating social practices; that agency was circumscribed as practices were codified into technical features of the platform. Nonetheless, while participants had less direct control over the creation of platform features, their accumulated contributions as ideas, feedback and data were still crucial to HSM’s ongoing development. This is an ambivalent process in the sense that as participants develop practices that may improve their lives, they simultaneously contribute to the programming of platforms they do not control (even though those platforms might more effectively improve their lives and the lives of others).
Conclusion
In their entry on ‘data’ in a recent Keywords for Media Studies, Gregg and Nafus (2017) argue that the current moment is ‘the last gasp of what has been called “participatory media”’. This reference to the ‘last gasp’ is metaphorically apt; it conjures up an image of the ‘voice’ and symbolic expression as the dominant concern in debates about participatory culture. They suggest that the ‘media studies to come will need to explain our involvement with data and their capturing devices as an accommodation, a co-habitation, a shared breath, mutual dwelling’. This way of thinking neatly captures the decade-long development of Hello Sunday Morning. It began as an open-ended network of bloggers, organised around written expression, reflection and discussion. The blog posts were efforts to narrate the self, to support others, and to imagine different cultural practices.
Reflecting back now, while HSM’s first wave of participants imagined they were creating a new kind of drinking culture, perhaps more specifically they were creating a new kind of drinking technoculture. They were contributing to a set of interdependent narratives, exercises, protocols and algorithms with the explicit purpose of managing the relations between our identities, bodies and alcohol consumption. HSM’s participants are part of a historical process of making ‘sociality technical’ by translating their lived experience into narratives and data that inform the development of the platform features, protocols and algorithms (van Dijck, 2013: 13). This is a participatory technoculture in the sense that expression, experience and action is always-already implicated in the process of programming sociality (Bucher, 2018). Following Gregg (2018), I suggest that participatory digital culture has been imagined in mostly expressive rather than athletic terms. Participatory technocultures like HSM are expressive and athletic, in the sense that they are characterised by exercises of self-optimisation, and regulated with processes of feedback and control. We’re athletic when we use digital media to undertake and evaluate exercises that improve our physiology, mood, attention, relationships and productivity. Participatory technocultures steer, guide, program and structure.
Participation with digital media involves the continuous collection and processing of data in order to transform the media platform itself, the people who are entangled with it, and the world it plays a part in creating. HSM participants stand as a useful metaphor for all of us who log the expressions of our mind and body on media platforms. When we do so we enrol ourselves in data-driven experiments, and via them, in the larger social process of building technocultural ‘life-support systems’ (Gregg, 2018; Sloterdijk, 2014). The stakes then are not only about whether there is control over the platform but whether practices alienate and atomise, or sustain and vitalise, our efforts to improve our selves and make a world together (Gregg, 2018). Critical accounts of digital media platforms then need to focus not only on how we ‘narrate’ our experience, but also the role we play in ‘programming’ the infrastructure that is used to experiment with social reality, our minds, and our bodies.
Footnotes
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author received funding from Hello Sunday Morning.
