Riding between fear,hope,and trust: A study of elite cyclists’ experiences of personal data sharing on TrainingPeaks

Abstract

Cycling exemplifies the intensification of datafication in elite sport, where the drive to enhance performance intertwines with digital technologies that generate extensive personal data about cyclists. While previous sociological research has often conceptualised datafication from the perspectives of surveillance and power, this study contributes new insight by analysing the relational dimensions of data sharing. Specifically, this study examines how elite cyclists experience and handle the sharing of personal data on the digital platform TrainingPeaks. Methodologically, the study is based on interviews with 10 men and 9 women Norwegian elite cyclists. Drawing on Hartmut Rosa's theoretical perspectives, the findings show how cyclists engage ambivalently with personal data sharing on TrainingPeaks. Their sharing practices are driven both by a fear of falling behind competitors and by a hope of gaining support and recognition. Moreover, trust represents a key consideration in navigating the uncertainties surrounding how coaches and team staff may use their personal data. Overall, the findings show how data sharing practices are shaped by how cyclists and others engage with, interpret, and act on personal data. This underscores the importance of including cyclists in decision-making related to data usage and interpretation.

In elite sport, the pursuit of maximising performance intertwines with the use of digital technologies that produce vast amounts of personal data about athletes. The importance of athlete data in elite sport is driven by an ethos of optimisation oriented towards staying ahead of competitors (Toner, 2024). This development reflects a broader societal trend, often described as the datafication of everyday life. Datafication refers to the systematic use of digital technologies to track and monitor aspects of human life, transforming them into digitised information defined as ‘personal data’ (Lupton, 2020: 4). As nearly every aspect of elite athletes’ bodies and performances is now subject to monitoring, Millington and Millington (2015: 140) suggest that elite sport has become indicative of ‘the datafication of everything’.

Cycling is a prominent example of the intensification of datafication, as most professional teams seek to use cyclists’ data to gain a competitive advantage. For instance, the Q36.5 Pro Cycling Team refers to data as their ‘26th rider’, highlighting how this extra ‘teammate’ shapes everything from performance development to recruitment (Zwets, 2025). The emphasis on data is further reflected in the systematic use of digital technologies worn on cyclists’ bodies (e.g. core sensors), bikes (e.g. computers), and in daily practices (e.g. sleep monitors). The volume and variety of personal data, ranging from external workload to sleep patterns, are commonly transmitted to digital platforms, such as TrainingPeaks, ¹ where they can be accessed by cyclists, coaches, and team staff (Rize et al., 2025).

In response to these developments, critical discussions of datafication have developed in the sociology of sport (Jones et al., 2022). Much of current research focuses on how athlete data is used for ‘dataveillance’ (Toner, 2024: 33), highlighting potentially harmful outcomes such as impacts on contract negotiations (Nappert and Bamber, 2023) or reinforcing power asymmetries between athletes and coaches (Greene et al., 2023). While these studies offer important insights, Baerg (2022: 214) notes that ‘little seems to be known about athletes’ relationship to datafication’ and emphasises that the potential for personal data to be shared requires further attention. This study addresses this call by examining the relational dimensions of personal data in elite sport, specifically how elite cyclists experience and navigate personal data sharing on TrainingPeaks. As personal data shapes how athletes respond to situations, but also influences their relationships, particularly those who interpret and act on their data (Toner, 2024), emphasising these relational dynamics are important for understanding the experiences and implications of personal data sharing in elite sport. The research question guiding this study is: How do elite cyclists experience and handle the sharing of personal data on TrainingPeaks?

The following sections include an overview of previous sociological research on datafication in elite sport, outline the theoretical framework drawing on Hartmut Rosa's relational sociology, describe the methods, and present the analysis and findings on elite cyclists’ experiences of sharing personal data on TrainingPeaks.

Previous research: datafication in elite sport

Practices of quantification have been integral to elite sport for decades, exemplified by athletes’ handwritten training diaries through the recording and comparison of aspects such as duration, distance, intensity, and perceived effort. Since the late 20th century, digitalisation has intensified these practices by expanding what can be measured, archived, and analysed through wearables and digital platforms (Toner, 2024). In endurance sports such as cycling, TrainingPeaks is one such digital platform, synthesising diverse forms of bodily and performance data into standardised metrics and analytical features. In this way, the use of TrainingPeaks can be understood as amplifying datafication and what Toner (2024: 5) terms ‘the cult of measurement’ in elite cycling. Insights from science and technology studies further underline that neither technologies nor data are neutral, but sociotechnical constructs shaped by, and shaping, the contexts in which they are used, with consequences that vary across individuals and groups (Skirbekk, 2025). Digital sociology complements this perspective by highlighting how the increasing shaping and structuring of everyday life through data raises social, political, and ethical questions, including issues related to selfhood, power, and surveillance (Lupton, 2015). Still, research on datafication in elite sport has mainly focused on engineering, biomechanical, and medical dimensions, often using experimental methods to develop practical recommendations, such as using athletes’ data to improve recovery programmes (Bourdon et al., 2017), with limited attention on how athletes experience such practices. Despite sociological analyses of datafication in elite sport remaining in their ‘infancy’ (Jones et al., 2022: 170), a growing line of enquiry has begun to address this gap.

For instance, Toner (2024) examined how datafication is often justified and reinforced through promises of performance optimisation. Neoliberal values, such as responsibility, competition, and self-optimisation, shape how athletes are socialised into embracing digital technologies and data to manage performance, reduce risk, and stand out. Similarly, Nappert and Bamber (2023) found that athletes are frequently subjected to datafication early in their careers, fostering the acceptance of data as beneficial. However, Toner (2024) and Nappert and Bamber (2023) also problematise the discomforting dynamics associated with the normalised expectations of the production and circulation of personal data. Athletes often report a loss of control over their data, particularly due to uncertainty about how it is interpreted and used. This perceived lack of control, combined with few opportunities for athletes to resist datafication practices, can generate stress and insecurity (Toner, 2024).

The power asymmetries between the athletes and practitioners within teams have also been examined. These studies demonstrate that coaches and team staff often hold authority over athletes through their control of data (Greene et al., 2023). For example, Williams and Manley (2014: 829) found that the extensive use of ‘data has become a seemingly critical conduit between coaches and players in terms of establishing a climate of control’, one that regulates players’ behaviour against the expected standards. These asymmetries can be linked to a broader surveillance culture in elite sport, where athletes’ personal data are negotiated as a source of disciplinary power, often drawing on Michel Foucault's panoptic model (Jones et al., 2022). Aligning with Williams and Manley (2014), Jones (2019) found that GPS data rendered football players more compliant by shaping their behaviour according to team expectations. Other studies have similarly problematised the implications of ‘dataveillance’ (Toner, 2024), where personal data are used to monitor athletes. Corr et al. (2025) showed that tracking performance, sleep, and location encroached on athletes’ private lives and occasionally fostered dissensions. Likewise, Manley et al. (2012: 306) found that data was used to categorise and benchmark athletes, reinforcing expectations of a ‘professional attitude’, which often compromised autonomy, privacy, and well-being.

Collins et al. (2015: 1088) challenged the tendency to frame digital technologies as purely controlling, arguing that data can also support athletes in achieving ‘personal and collective goals’. Studies examining how athletes use and make sense of data have shown that they often critically assess its credibility by comparing it with bodily sensations and contextual factors such as weather and terrain. Rather than solely dictating behaviour, data can function as a form of guidance that athletes interpret in relation to their experience (Toner et al., 2023). Yet the distinction between support and control becomes more blurred when elite athletes are required to share data on digital platforms as part of their professional obligations, as these data can serve both as opportunities and constraints, for instance, when used as criteria in selection to competitions (Røsten, 2025). Similar dynamics, relating to how data circulates on digital platforms in ways that both support performance and enable forms of social comparison, surveillance, and control, are discussed in studies of platforms such as Strava (Couture, 2021). Røsten and Rogstad (2025) found that elite cyclists balanced the sharing and withholding of performance data on Strava, including heart rate and power output, across training and competitive periods, which reflected tensions between engaging with fans and protecting competitive advantages.

Even when datafication is framed as representing athletes’ best interests (Collins et al., 2015), the varied uses of personal data demonstrate the importance of context in shaping their experiences. For instance, sharing training data may be acceptable for monitoring purposes, whereas sharing sleep data may be perceived as invasive (Corr et al., 2025). Such nuances in how athletes perceive data sharing underscore the importance of examining the relational dynamics that shape athletes’ experiences and their trust in how their data are used in decisions made on their behalf.

Theoretical framework

To examine how elite cyclists experience and handle the sharing of personal data on TrainingPeaks, this study draws on Hartmut Rosa's relational sociology. Sport scholars have emphasised the value of Rosa's framework for understanding sporting practices, for example, by reimagining the intensification of datafication, surveillance, and standardisation of athletes’ performances (Bjørndal and Espedalen, 2026), or by conceptualising resonant relationships between athletes and coaches (Reinold, 2026). In contrast to these studies, this analysis draws on Rosa's analytic descriptions of modernity to examine the relational dynamics of datafication, rather than the normative dimensions of his theory that present resonance as an alternative to the acceleration processes of modernity.

Rosa (2020) argues that modern societies are driven by a persistent desire to make the world controllable across four dimensions: visibility, accessibility, manageability, and usefulness. In elite sport, datafication reflects these dimensions by rendering athletes visible through data, accessible via digital platforms, manageable through analysis and prediction, and useful for improving performance (Toner, 2024). This pursuit of controllability fosters a particular stance towards the world as ‘a point of aggression’, where ‘everything must be known, mastered, conquered or made useful’ (Rosa, 2020: 6). Rosa (2020) explains the normalisation of relating to the world as a series of points of aggression through two interrelated principles. The first is dynamic stabilisation, where modern societies require constant growth, innovation, and acceleration to maintain their current state. Individuals increasingly experience this escalatory pressure as a fear of falling behind. Rosa (2020) describes this as a slippery-slope phenomenon, wherein maintaining one's position demands continuous improvement. In elite sport, this manifests in how data shape athletes’ (market) value and (future) potential. Failure to improve key metrics, such as maximum oxygen uptake or sprint velocity, can lead to loss of status or exclusion from teams and competitions (Nappert and Bamber, 2023).

The second principle is the promise of expansion, which Rosa (2020) describes as the belief that life improves when individuals expand their reach to what is available, attainable, and accessible. This triple-approach logic shapes everyday actions and explains the appeal of digital technologies. For elite athletes, engaging with digital platforms makes their data available (uploaded), attainable (measurable), and accessible (to others), potentially opening opportunities for tailored training, competition participation, or new contracts (Røsten, 2025).

However, relating to the world as a point of aggression has implications for how individuals relate to themselves and others. Rosa (2020) uses the concept of alienation, a mode of relating to the world marked by a lack of meaningful connections and responsiveness, to encompass all these issues. Datafication can contribute to alienation, especially when the digital representation of an individual is disconnected from their experiences (Rosa, 2020). In elite sport, this can emerge when athletes are excluded from data analysis and decision-making (Greene et al., 2023) or when their data are interpreted as failing to meet normalised expectations (Jones, 2019), leading to discomfort and a diminished sense of agency (Reinold, 2026). Rosa (2020) suggests that cultivating an attitude of listening and responsiveness could offer a more meaningful way of relating to the world. Taken together, Rosa's perspectives provide insight into how efforts to make athletic performance measurable, predictable, and controllable are constituted by social practices and relationships, useful to understand datafication processes.

Methods

Semi-structured interviews were used to examine elite cyclists’ experiences of sharing personal data on TrainingPeaks. This study is philosophically grounded in an interpretivist approach in which interviews are understood as co-created accounts of sociotechnical practices, where experiences and the meanings attached to them are constituted and rendered intelligible through conversations. This stance is accompanied by a relativist ontology, recognising that participants may experience and interpret reality differently, and a constructionist epistemology, where knowledge is understood as developing through interaction between researcher and participant (Thagaard, 2018).

Sample

Cyclists were recruited through strategic sampling (Thagaard, 2018) based on two criteria: (1) competing for senior Norwegian national teams and/or teams at the Union Cycliste Internationale (UCI) continental level or higher, and (2) experience using digital technologies in training. This approach helped develop an understanding of elite cyclists’ engagement with digital technologies and the data produced through these interactions, enabling the exploration of how such data shape the decisions they make for themselves and those made by others. A limitation, however, is that strategic sampling may omit a broader variation of experiences, as it draws on a targeted sample of Norwegian elite cyclists. In addition, the absence of coaches and team staff further narrows the perspectives included, although this reflects a deliberate focus on cyclists’ experiences with datafication. The participants were identified through initial contact with the national governing body of cycling in Norway. Each participant was contacted directly via email. The participants in this study were 10 men and 9 women, Norwegian cyclists aged between 19 and 34 years. All have cycling as their primary professional commitment, with training and competition organised around UCI-rated races and championships. In their earlier careers, their engagement with TrainingPeaks developed over time, typically either when they began pursuing cycling more seriously during their teens or when they signed with senior teams, where data sharing became a formalised expectation. The participants were given pseudonyms, which are presented in Table 1.

Table 1.

Participants’ characteristics.

Pseudonym	Age	Level
Alexander	19	Continental
Benjamin	19	Continental
Caroline	20	Continental
David	21	Continental
Ethan	21	Continental
Fred	21	Continental
Gordon	21	Continental
Harvey	22	Continental
Isaac	22	Continental
Josie	22	Continental
Kylian	23	Continental
Linnea	23	International
Matilda	24	Continental
Noah	25	Continental
Olivia	25	Continental
Patricia	26	Continental
Quiana	32	International
Rachel	32	International
Sophie	34	WorldTeam

Data collection and analysis

Semi-structured interviews were conducted digitally between June 2024 and February 2025 using Zoom Workplace due to practical and logistical constraints. Many participants resided outside Norway and navigated intensive schedules. The interview guide was structured into: (1) sporting backgrounds, (2) use of digital technologies and data, and (3) implications of datafication for cyclists. For example, participants were asked to reflect on the practices of tracking and sharing personal data, including how they manage the extent to which others may monitor and use their data on digital platforms like TrainingPeaks and Strava. The semi-structured format allowed for flexibility, enabling clarification and follow-up questions (Thagaard, 2018). The participants’ engagement and interest led some interviews to expand beyond the interview guide, with durations ranging from 60 to 150 min.

All interviews were audio recorded with participants’ consent, transcribed verbatim, and processed in accordance with ethical approval from the Norwegian Agency for Shared Services in Education and Research. Collective qualitative analysis was the starting point for the analytical process, and is a strategy considered valuable for fostering creativity and strengthening analytical rigour. While resembling thematic analysis in its procedural steps, this approach emphasises the joint reviewing, mapping, and sorting of empirical data (Eggebø, 2020). Together with a colleague experienced in sociological research on sport and technology, we organised a workshop to systematically work through and discuss the interview material. Prior to the workshop, we summarised the interviews. We then discussed preliminary understandings and collaborated on writing focused interviews. This formed the basis for mapping key points, topics, and themes, including theoretical connections and ideas for potential articles.

One suggestion developed through this collective process was the cyclists’ experiences of sharing practices on digital platforms, drawing on three overarching themes: (1) elite sport and demands, (2) data, and (3) visibility. For example, visibility is centred on how cyclists became subject to various forms of evaluation, ranging from coaches and team staff to fans and anti-doping authorities. This collective qualitative analysis laid the foundation for the subsequent individual phases. I re-read the interviews and summaries with particular attention to cyclists’ engagement with TrainingPeaks, selected quotes and developed further analytical categories. This interpretive process was shaped by Rosa's relational sociology. In line with the constructive and interpretive framework, both semantic coding, which focused on the explicit content of cyclists’ statements, and latent coding, which involved interpreting underlying meanings, were applied. For example, several cyclists expressed reflections such as ‘to take the next step, I must’. Here, Rosa's slippery-slope dynamics served as a sensitising concept that guided the interpretation of this semantic emphasis on ‘must’ as an expression of experienced optimisation pressures. Through this iterative process between quotations and interpretation, three analytical themes were constructed: (1) optimisation endeavours and fears of falling behind, (2) fostering connections and hopes of recognition and care, and (3) managing uncertainty and trust as a basis for sharing personal data. These themes structure the Findings and Discussion section.

Findings and discussion

To examine how elite cyclists experience and handle personal data sharing on TrainingPeaks, the analysis is structured into three parts. The first and second parts focus on how the fear of falling behind competitors and the hope of gaining support drive cyclists to share personal data. The third part focuses on trust as a key relational dynamic that enables cyclists to navigate the tensions between fear, hope, and uncertainty about how others interpret and use their data.

Optimisation endeavours and fears of falling behind

A recurring reflection among the cyclists was the emphasis on ‘marginal gains’ as a necessity for competitive success, where digital technologies and data were often framed as decisive factors. As Noah reflects, the extensive tracking and sharing of personal data on TrainingPeaks is considered a marginal gain in the pursuit of performance optimisation:

At this level, it is about the marginal gains. A bike computer connected to a […] core sensor can provide an extra percentage by optimising training. Similarly, a smartwatch that tracks sleep […] can be useful for assessing the recovery. […] If I share [data] with coaches [on TrainingPeaks], it can help me improve even further (Noah).

Noah's response illustrates how bodily functions and activities can be transformed into sharable data, making performance more controllable, predictable, and manageable (Rosa, 2020). Sharing personal data serves a dual purpose, enabling both monitoring current performance and identifying areas for improvement. Although cyclists acknowledge that performance can be optimised through training and recovery strategies using metrics such as body temperature and sleep patterns, they remain aware of the inherent difficulties in controlling and predicting performance:

Each road race has its own life. You never really know how it will turn out […] and who will perform best in the end (Kylian).

This paradox highlights what Rosa (2020) describes as a contradiction between controllability in theory and uncontrollability in practice. The pursuit of performance optimisation by tracking and sharing personal data (cf. Noah) is often constrained by the inherent unpredictability of outcomes in practice (cf. Kylian). This contradiction highlights that, if total control over performance is achievable, the pursuit of control would be redundant, as cyclists would already be performing at their optimal levels. Conversely, if data could fully control performance, it would require both measurements to produce reliable data and a clear understanding of what constitutes optimal metrics, which often proves elusive in practice (Toner et al., 2023).

Given this paradox, one might ask why cyclists continue to engage in practices of tracking and sharing personal data on TrainingPeaks when these do not necessarily lead to the desired outcomes. The answer may lie in the nuanced distinction between beliefs and certainty. Cyclists believe in something they cannot fully control or know, yet they remain drawn to the potential of these practices, such as optimising training or assessing recovery. Despite the elusiveness of controlling performance in practice, framing data as sources of ‘marginal gains’, and crucially, appearing to be doing everything possible to retain control, remains central to their understanding of being elite cyclists. Josie reflects on this dynamic, emphasising how tracking and sharing personal data on TrainingPeaks helps make her performances appear as controllable and manageable as possible:

The higher the level, the stricter and more controlled everything is. Eventually, you start tracking almost everything you do […], and everything is uploaded to TrainingPeaks. […] For example, how many grams of carbohydrates, fat, and protein I have eaten (Josie).

She also connects such practices in terms of remaining competitive:

When you go all in on cycling and everyone you ride with is also on-point with those things, you lose quite a lot if you are not on-point yourself. Not all the work happens during training. What you do throughout the rest of your day is also important (Josie).

Together, these reflections show how cyclists often relate to performance as a mode of aggression (Rosa, 2020), where the drive to ‘go all in’ extends beyond training sessions to encompass everyday life practices. For example, this manifests in Josie's pursuit of mastering nutrition by making it visible (i.e. displayed on TrainingPeaks), accessible (i.e. grams of micronutrients consumed), manageable (i.e. sufficient intake), and useful (i.e. improving performance). Furthermore, this highlights cyclists’ experience of the slippery-slope phenomenon (Rosa, 2020), where continuous optimisation, to be ‘on point’ on everything from training to nutrition, is driven by the fear of losing ground to the rest of the peloton. For many cyclists, the fear of falling behind justifies and reinforces the pressure to share personal data on TrainingPeaks. This is reflected in how they often relate to such practices as something they ‘must’ do to progress or perform:

Basically, my entire life is uploaded to TrainingPeaks […], but if I want to take the next step, I must be honest (Harvey).

All data are uploaded to TrainingPeaks, where my coach and team have access. […] To perform at my best, they must know how things are going (Linnea).

By referring to data sharing as a ‘must’, Harvey and Linnea's accounts illustrate how the demands of these practices are challenging for cyclists to refuse. The basic tenet of optimisation is embedded in the culture of elite sport, where the relentless pursuit of marginal gains and win-at-all-cost mentalities are normalised demands (Toner, 2024). Thus, not striving to optimise performance may be interpreted not merely as a lack of commitment but as a deviation from the demands of elite sport. This is expressed in how several cyclists describe the drive to perform as shaping a particular ‘lifestyle’, one that ‘just comes naturally’:

The way you live your life makes you feel that you have nothing to hide. […] It just comes naturally. It is a lifestyle to be and live like an elite cyclist. Many might say that I sacrifice a lot […] but I want to succeed. Then I know that I cannot sleep for five hours and expect I am going to perform optimally (Patrica).

For Patrica, practices that might otherwise be regarded as ‘sacrifices’ are instead perceived as essential elements of performing at the elite level. Her openness to sharing personal data, reflected in the feeling that she has ‘nothing to hide’, suggests that the drive to perform outweighs privacy concerns. As she further reflects:

If anyone is really opposed to others knowing how much they sleep, I wonder what you are doing. For me it is more like: ‘Of course I can give you access to that’ (Patrica).

This aligns with Greene et al.'s (2023) argument that improvements often compromise athletes’ privacy. Instead, data sharing often becomes a way for cyclists to conform to professional attitudes (Manley et al., 2012), shaping their understanding by distinguishing between what is acceptable and unacceptable. Cyclists’ reluctance to share, for example, sleep data, can therefore be interpreted as being misaligned with the expectations and lifestyle of an elite cyclist.

Sharing personal data reinforces the idealised standards, where cyclists are expected to take responsibility for their practices in line with these expectations across all aspects of daily life (Williams and Manley, 2014). However, this alignment is often accompanied by ambivalence:

I am not always keen on my coach knowing when I go to bed. Let's say I go to bed at midnight one night, then I feel: ‘Oh, now I should be ashamed […] because I went to bed a bit late’ (David).

David's experience shows how falling short of these standards, such as going to bed later than deemed acceptable, can evoke feelings of guilt (Toner, 2024). Beyond this, the ideal of the elite cyclist fosters a continuous pressure to improve. This is especially demanding in a context where every cyclist strives to optimise even the smallest performance gains to stay ‘on point’, often leading to a lingering sense of ‘never being good enough’:

I learned that you are never good enough. […] This does not mean that you are bad, as you can still perform very well, but there is always room to push more and get better (Quiana).

As Quiana explains, the pursuit of optimisation becomes a self-perpetuating spiral. No matter how much is improved or how well one performs, there is ‘always room’ for growth. This dynamic reflects Rosa's (2020) critique of dynamic stabilisation, where constant improvement is required to maintain the current state. While this construct of ‘never good enough’ can be psychologically demanding, it is difficult for cyclists to question or resist the pressure to improve, as it is often equated with falling behind. This aligns with findings from previous research showing that datafication reinforces ‘disciplinary expectations’, shaping how cyclists interpret their performance as something that is continually expected to improve in order to stand out (Toner, 2024: 139). Yet the tensions between optimisation endeavours and the implications of datafication are not inherently opposed to joy or meaningful experiences, as striving to ‘push more’ and realising personal ambitions can also be fulfilling. As Bjørndal and Espedalen (2026) note, striving towards goals can be experienced as self-expressive and transformative rather than only productivity-driven. Reflecting this variety, the cyclists in this study express an ambivalent relationship with sharing practices. While they face pressure to optimise their performance, they also highlight the potential opportunities these practices afford for connection and care.

Fostering connections and hopes of recognition and care

The cyclists often describe personal data sharing as a relational practice for connecting with key stakeholders and emphasise that engagement on TrainingPeaks is important to shaping their career trajectories. Their reflections indicate an underlying hope for gaining recognition, as personal data can attract the interests of professional teams, influence competitive opportunities, and help secure better contracts:

It has become common for professional teams to gain access to TrainingPeaks accounts for promising cyclists, often to evaluate their power data and use this to recruit talented riders (David).

[TrainingPeaks] is also about having a network, for example in selection processes for the national team at championships. Therefore, I chose not to remove the cycling federation's access (Harvey).

In contract negotiations with teams, I highlighted things such as my personal records related to power output and peak power, as displayed on TrainingPeaks (Sophie).

These reflections illustrate the appeal of personal data on TrainingPeaks, showing how such data shapes what is available, accessible, and attainable for cyclists to advance their careers (Rosa, 2020). Through TrainingPeaks, cyclists’ personal data becomes available to managers and scouts, creating opportunities such as signing with a professional team. By sharing their personal data, cyclists also make themselves accessible to decision-makers, who can evaluate their potential and readiness to compete in races. Finally, this can lead to attainable outcomes, such as securing a more financially lucrative contract. These dynamics suggest that while data sharing across digital platforms can serve similar functions in enabling engagement with relevant audiences, the concerns and vulnerabilities involved differ. While Røsten and Rogstad (2025) argue that elite cyclists often withhold power data on Strava to protect competitive advantages, on TrainingPeaks, the tension shifts towards the worry of not displaying ‘good enough’ power data to coaches, teams or scouts. As possibilities for recognition, selection, and support reinforce the value of sharing personal data, cyclists explain that it is common for multiple practitioners to have access to their TrainingPeaks accounts:

My coach, my former coach, the team's sports directors, and two other teams have access to my TrainingPeaks account (Josie).

Cyclists frequently highlight that sharing personal data on TrainingPeaks is important for fostering collaboration and receiving support, especially because coaches and team staff are often not ‘physically’ present in their daily lives. Compared to many team sports, where coaches are physically present during most training sessions, this emphasis on data is reinforced in cycling because training often occurs remotely, making data a key part of how cyclists’ everyday routines are guided and evaluated (Jones et al., 2022). As a result, data becomes an important mediator of the cyclist–coach relationship, enabling coaches to monitor, interpret, and respond to cyclists’ performance from a distance. For many cyclists, these practices are tied to a sense of care, understood as attentive practices through which coaches and staff ‘look after’ cyclists by interpreting their data to recognise signs of strain, illness, or overload:

My coach is not around me in my everyday life to see what I am doing. He is based in [another country]. […] TrainingPeaks is important for ensuring the best possible interaction with my coach, allowing him to keep track of everything I do without being physically present (Josie).

[TrainingPeaks] means that riders are now better looked after and cared for. […] When someone is sick or worn down, they are no longer just thrown away; teams consider all the metrics [on TrainingPeaks]. ‘You are starting to get sick’ […] (or) ‘there is something wrong with your heart rate’, so we will take care of you; we will also make sure you do not get overtrained (Quiona).

Josie and Quiona's reflections complicate the often-binary framing of datafication in elite sports, which tends to conceptualise it primarily as either a source of support (Collins et al., 2015) or control (Jones et al., 2022). Instead, their accounts underscore the importance of engaging in diverse experiences as the implications of datafication can vary. The sharing of personal data reflects a complex interplay between caring for cyclists through support and protection and tendencies that may contribute to discipline and control. Even when coaches adjust training and recovery based on a cyclist's current metrics, questions like ‘How are you?’, or more specifically, ‘How are your metrics on TrainingPeaks?’, are often filtered through the lens of performance optimisation and identifying areas of improvement. Thus, despite an emphasis on care, cyclists are frequently told what to do and not do by prescribed plans aimed at maximising output. This creates a form of care that is entangled with control. While cyclists often interpret these interactions as caring, such care can also mask organisational control and interests. Previous studies show that datafication can create compliance by directing attention towards deficiencies, risks, and areas for improvement, thereby positioning athletes’ value in terms of consistent performance (Toner, 2024; Williams and Manley, 2014). This raises questions about the extent to which care supports cyclists’ needs or primarily sustains performance expectations.

Moreover, sharing of personal data on TrainingPeaks risks fostering a narrow form of care centred on quantifiable metrics, typically physiological, while other crucial dimensions of well-being, such as mental health, may be overlooked. As Nappert and Bamber (2023) and Røsten (2025) argue, an overreliance on specific data in selection and recruitment processes risks overlooking athletes whose potential is not made entirely visible through quantifiable metrics. Similarly, although sharing personal data on TrainingPeaks may contribute to caring for the physical aspects of cyclists’ performance, it risks neglecting other needs that are not subject to metric evaluations. These concerns align with Gordon's reflections, which emphasise how current quantifiable metrics are often prioritised by coaches and teams:

It is a fixed habit to track all training, along with things like sleep and heart-rate variability, so that it is uploaded to TrainingPeaks. […] I think the worst part is that if it is not displayed [on TrainingPeaks], it seems like you have not done it because that is what coaches and teams assess (Gordon).

As Gordon highlights, the way personal data is represented on TrainingPeaks can influence decision-making, as those with access tend to value and evaluate what is ‘displayed’. This incentivises cyclists to share personal data if they want to compete at the elite level. Consequently, their ability to resist these practices is limited (Toner, 2024), as they often feel compelled to make themselves available, accessible, and attainable through their personal data. The dynamics of these sharing practices also suggest that cyclists who are unable to display sufficient (or measurable) value through their data may be vulnerable to being overlooked in decision-making processes. A strong emphasis on, for example, power data can narrow evaluations of performance and potential, increasing the risk that cyclists whose technical or tactical skills are less readily captured become overlooked (Røsten, 2025). This further supports the finding that elite athletes are embedded in structures marked by power asymmetries that shape how they are managed and evaluated (Greene et al., 2023). On the surface, however, privacy settings on TrainingPeaks offer cyclists a sense of control over who can access personal data:

Those who have access to TrainingPeaks can see everything; however, I can also block and remove people (Alexander).

I control the TrainingPeaks account, so I decide who has access (Kylian).

While cyclists may technically manage access on TrainingPeaks, Alexander continues to explain that this sense of control is often ambivalent, shaped by embedded structures that limit the possibility of opting out of personal data sharing:

Two other professional teams have access to my TrainingPeaks. They asked to see my training data, and it would feel wrong to say no. If I want to ride for that team one day, I cannot really refuse it because it would be taken as a rejection. […] It is exciting that they are following my progress, but at the same time they see everything that I do every day (Alexander).

This challenges the common assumption that cyclists can control who has access to their TrainingPeaks accounts. Instead, their agency is shaped by the social organisation of elite cycling and external expectations from coaches and teams, whose interests may not always align with those of cyclists. Alexander's remark, ‘I cannot really refuse’, illustrates the ambivalence between these conflicting interests. Although he is excited by the possibility of having his name and performance recognised by two top-level teams, the simultaneous exposure to more sensitive data, such as sleep and daily routines, feels unsettling. Although this points to how personal data sharing can reproduce power asymmetries (Greene et al., 2023), cyclists also describe these practices as shaped by relational dynamics, depending on how others engage with, interpret, and act on their data. Here, trust seems to be a key dimension, connecting the hope that those with access possess the competence and care to engage with data responsibly to the fear that personal data could be misunderstood and misused.

Managing uncertainty and trust as a basis for sharing of personal data

Sharing personal data can be a vulnerable and risky practice for cyclists, as it entails a loss of control over how their data may be analysed and used by others. In the interviews, many cyclists emphasise trust as a key relational quality that enabled them to share personal data with coaches and team staff. In this sense, trust refers to having positive expectations about others’ intentions and actions despite having limited control over how one's data may ultimately be used:

I do not feel that it affects me much that coaches and other teams see my data [on TrainingPeaks] […] because I know they are serious people who want the best for me (Noah).

Individuals who have access to my TrainingPeaks are those I have built trust and relationships with. They use it only to help me […], and I trust them (Matilda).

As Noah and Matilda explain, a relationship is formed between cyclists and those who access their personal data on TrainingPeaks, where cyclists give up control to those they perceive to act in their best interests. Despite the underlying imbalance of power in this relationship (Jones, 2019), cyclists expect those with access to support them in achieving their goals. Mathilda elaborates:

I feel like they [the individuals with access on TrainingPeaks] are helping me reach my goals and dreams. I feel like they are giving me something rather than me giving them something. […] It is like going to a doctor. Doctors do many strange things that might feel uncomfortable, but they act professionally. You accept that they do things that might otherwise feel intrusive, because it is for your own good (Matilda).

Matilda's comparison with visiting a doctor illustrates how personal data sharing is often accepted when it is perceived as serving the cyclist's own benefit. Just as doctors perform various examinations to make diagnoses, coaches can use personal data to tailor training programmes. Although data on sleep or heart rate may feel intimate and intrusive, the trade-off of sharing such information in exchange for optimising performances and support contributes to an understanding that it is ultimately for the cyclists’ ‘own good’. To reinforce this point, Matilda reflects on two different coaches she has worked with:

I did not respond to the training that [former coach] planned. I felt like [former coach] did not read me or my data correctly. […] So, I switched coach […] and that year, I became better than I had ever been. […] I honestly believe [new coach] spent more hours analysing my sessions and numbers [on TrainingPeaks] than I spent on the bike that year (Mathilda).

This highlights four key dimensions of trust. First, cyclists expect those with access to their personal data to be competent in supporting and guiding them. Second, trust only becomes relevant when something of value is at stake. Third, trust connects cyclists and those with access; however, it can also create a distance between them. Fourth, trust is not granted but is built through experience and ongoing evaluation (Øverenget, 2025). As Matilda explains, her decision to switch coaches was driven by a misalignment between her experiences and performance, as she did not respond to the training and felt that her former coach's competence and interpretation of her data lacked understanding. While such accounts nuance previous findings that the application of data often silences athletes by fostering a ‘climate of control’ (Williams and Manley, 2014), this may also reflect the specific context of individual sports like cycling, which can be less marked by power asymmetries between coaches and athletes compared to team sports (Greene et al., 2023). Although the ability to switch coaches and resist certain datafication practices is likely an exception rather than a general rule in elite sport, Mathilda emphasises that personal data sharing concerns not only access but also how those with access interpret and engage with the data. Ethan and Sophie illustrate how both positive and negative experiences shape these relational dynamics of personal data sharing:

Others might take off their smartwatches when they go out partying to hide it, but for me it is important to have an open relationship with my coach and give him access to the fact that I might have slipped one day (Ethan).

It is a bit strange that my coach has access to my sleep data […] but he is very detail-oriented to provide the best possible support. […] I remember that during a stage race, we were all tired, and he commented [on TrainingPeaks] that I was one of the last to breakfast, probably because my data indicated that I was still sleeping. We just laughed it off […] but if he had kept commenting and joking about it, it could have ruined things (Sophie).

These examples emphasise the relational aspects of data sharing, showing how trust can be strengthened or weakened depending on others’ engagement with personal data. Ethan, for example, appears not to have encountered negative experiences and instead views openness with his coach, even when sharing personal data from nights out, strengthen their relationship. Similarly, Sophie acknowledges that while sleep data may feel personal, she places trust in her coach when his actions reinforce the trust she has given him, and he takes responsibility for maintaining it, such as by providing ‘best possible support’. At the same time, she notes that continued ‘joking’ about her sleep could have undermined their relationship, as it created a sense of discomfort (Toner, 2024). This underscores the ambivalence cyclists often express regarding the blurred line between what is considered appropriate and inappropriate use of personal data. While there is underlying trust and expectation that others will use cyclists’ personal data to support them in achieving their goals (Collins et al., 2015), concerns persist about how these data might be interpreted and used in ways that marginalise them:

Data that appear inappropriate but may not mean anything can lead someone to decide that you cannot race or train. For example, if you only slept six hours one night, that does not necessarily mean you will perform worse […] but when others analyse the data, they might think: ‘Oh, he cannot race’, or ‘he needs to hold back in training’ (David).

David's account illustrates how cyclists can feel alienated when their personal data is interpreted and used in ways that seem disconnected from their own understanding. This disconnect reflects a loss of meaningful connection between the cyclists, their data, and those interpreting them (Rosa, 2020). In David's case, sleep data that appears ‘inappropriate’ (i.e. 6 h) may fail to meet normalised expectations of recovery (e.g. 8 h), leading to decisions that can diverge from his own evaluation of readiness to compete or train. This is also a matter of power dynamics (Jones et al., 2022). When cyclists’ evaluations are overlooked in favour of data that produce competing verdicts, and when there are no opportunities for dialogue and responsiveness, their agency and sense of self-efficacy are diminished (Reinold, 2026). In such situations, the absence of mutual engagement can erode trust. When relationships between cyclists and those with access to their personal data are shaped by distrust, they may reinforce monitoring and verification practices (Williams and Manley, 2014), implicitly signalling to cyclists that they are not trusted. Linnea reflects on this dynamic, pointing out how sharing personal data can create a feeling of being ‘watched’ and judged, rather than supported:

The team also uses TrainingPeaks to assess if we have trained well in selection processes. […] Since they have access, they can constantly monitor what I am doing, but that does not mean it reflects what is ‘true’. […] For me, it has been very important to have dialogue rather than feeling like I am simply being watched (Linnea).

Linnea's account shows how data sharing can generate uncertainty and mistrust, particularly when cyclists are excluded from the interpretive process that determines what is considered ‘true’. While personal data may reveal patterns relevant to cyclists’ training and recovery status, they do not explain the context behind those patterns. For example, their heart rate might be affected by weather conditions, and less sleep could be caused by outside noise. Many highlight the importance of considering such contextual factors when interpreting data, since metrics do not capture the whole or ‘true’ story. While Toner et al. (2023) show that athletes draw on embodied and contextual cues when making sense of data individually, for the cyclists in this study, the remote and interoperable sharing of personal data creates additional interpretive tensions, making dialogue essential for bringing contextual factors into interpretation. As Linnea points out, when data is interpreted in dialogue ‘with’ cyclists, it is often experienced as constrictive and supportive. When it is interpreted ‘without’ them, it can feel imposed and create uncertainty. This supports Rosa's (2020) argument that attitudes of listening and responsiveness are essential for fostering meaningful relationships between cyclists and those who engage with their personal data. Cyclists highlight that trust in data sharing is built when they are asked and included, rather than judged and instructed:

Let us say I sleep poorly one night, then it is more relevant that I call my coach and say: ‘I slept badly. What do you think? Should I conduct the planned session or adjust it?’ This dialogue is more important than simply looking at cold data. The data do not show the entire picture (David).

David's reference to ‘cold data’ underscores the relational aspect of data sharing. This illustrates how personal data can generate discomfort and misrepresentation when decisions are made without the cyclists’ input. Thus, an assumption that personal data represent the ‘entire picture’, becomes problematic when the cyclists’ evaluations are disregarded. In the absence of dialogue that enables joint interpretation, this dynamic contributes to the understanding that athletes’ data in elite sport often serve a disciplinary function (Jones, 2019). Although cyclists highlight the importance of dialogue in shaping their experiences and trust in relation to personal data sharing, translating this into practice is often complex. A central issue concerns how to qualify the consent of cyclists who accept tracking and sharing of their personal data. Within the sociocultural structure of elite cycling, individual choice appears constrained, as data sharing through digital platforms like TrainingPeaks is perceived as a ‘necessity’. This is further complicated by hierarchical relations, where those with access to data often assume authoritative roles, positioning cyclists in subordinate roles where they are expected to comply with externally derived recommendations (Jones, 2019). Although cyclists stress the need for their voices to be heard, implementing more participatory approaches to interpretation and decision-making within practices that tend to prioritise optimisation, competition, and evaluation may constrain the possibility of acknowledging relational values (Bjørndal and Espedalen, 2026). This suggests that while initiatives and regulations aimed at securing athletes’ digital rights are important (Toner, 2024), they may not be sufficient unless accompanied by changes in how relational dynamics around data sharing are shaped, in ways that foster mutual recognition and more inclusive practices. In this regard, Reinold's (2026) resonance-centred framework between athletes, coaches, and sport, through features of affection, self-efficacy, transformation, and semi-controllability, may help envision more responsible uses of athlete data. For example, such an approach foregrounds responding to athletes’ differences and vulnerabilities, and a willingness to engage with diverse interpretations and the uncertainties that personal data may carry, rather than assuming a single authoritative reading. However, its normative and not yet empirically examined basis makes practical translation within current elite sport contexts both uncertain and challenging.

Conclusion

This study examined how elite cyclists experience the sharing of personal data on TrainingPeaks. The findings show that cyclists engage ambivalently with personal data sharing, driven by a fear of falling behind competitors and a hope of gaining recognition and support. These dynamics create a self-reinforcing cycle in which data sharing becomes increasingly expected and embedded in cyclists’ everyday practices. Although data sharing is often framed as being in cyclists’ best interests, it also shapes the relationships between cyclists and those with access to personal data. Here, the hierarchical structures of elite sport position coaches and team staff in roles of authority, while cyclists may be compelled to follow externally derived recommendations. These dynamics foster relations that are effective for enhancing performance optimisation through measurable metrics, but can also create uncertainty and narrow the scope of athlete care to physical indicators at the expense of broader well-being. Cyclists emphasise the importance of trust in managing these uncertainties and stress that personal data should be engaged with through dialogue rather than unilateral monitoring. They call for more participatory and co-interpretive practices, where their perspectives are included in the decision-making processes.

Overall, this study contributes to broader discussions on how elite athletes engage with tracking and sharing personal data through digital platforms. The increasing reliance on data in decisions that shape athletes’ practices and career trajectories has prompted calls for more comprehensive approaches to data governance in elite sport (Toner, 2024). Although the varying purposes and uses of athlete data complicate efforts to develop coherent data governance frameworks, this study advances a foundation for such frameworks by drawing on Rosa's relational sociology to show how athletes’ experiences are shaped by those who interpret and act upon their data. A relational approach also opens further questions about how more responsible and responsive uses of athlete data might be constituted.

Building on the findings, the doctor metaphor described by Mathilda, centred on data acting in the athlete's ‘best interest’, requires careful interpretation, particularly when considering the relational dynamics that shape who holds authority and who is included (or excluded) in how data is understood and used. Coaches and team staff often assume a role akin to that of doctors, where data is framed as a foundation for recommendations intended to enhance athletes’ performance, much like medical data guides treatment decisions. This dynamic places substantial implicit responsibility on those interpreting data, since athletes depend on coaches and team staff for everything from guidance to competitive opportunities. In turn, this dependency may increase coaches and team staff's control over athletes, further reinforcing the athletes’ dependence on them. Such dependencies can create dynamics in which athletes’ care becomes vulnerable to paternalistic relationships, where their data can be misunderstood and misused. An example of this is when coaches and team staff omit athletes from the interpretation of their data. While this might be intended to avoid burdening athletes with additional responsibilities, excluding them nevertheless becomes problematic, as it risks marginalising athletes’ knowledge of how their data is understood and used. As a result, data may wrongly suggest a decline in performance, with consequences for competition selection, training adjustments, or recovery measures. This example underscores how athletes’ exclusion from data interpretation and decision-making processes reinforces hierarchical power asymmetries and increases their vulnerability. In turn, it highlights the need for data governance frameworks to extend beyond ethical guidelines and formal rules so they also address the relational dynamics that shape who holds interpretive authority and who can participate in decisions about athlete data in elite sport, as this is essential for ensuring that athletes remain active participants in processes that impact them.

Taken together, these insights also have a theoretical contribution as they point to the value of combining Rosa's relational sociology with sociological perspectives on power to examine how datafication shapes relations and practices in elite sport, opening directions for future enquiry. Issues about power in the interpretation and use of data are central in relations between athletes, coaches, and sporting organisations, as athletes are dependent on coaches’ authority while coaches are also shaped by broader organisational constraints, creating reciprocal dependencies. Analysing the ways power functions in these relationships can thus offer insights into how expectations for continual improvement associated with Rosa's dynamic stabilisation are negotiated, imposed, or contested.

Building on these contributions, this study's limitations also highlight potential directions for future research. As it is based on a sample of Norwegian elite cyclists, future research could explore similar relational dynamics across other elite sport contexts. Likewise, including coaches and team staff in the sample of future studies would provide important complementary perspectives, particularly considering how athlete data is interpreted and used in decision-making. Here, the doctor metaphor could also be drawn upon to identify how supportive intentions may blur into forms of control or manipulation when intertwined with datafication. As personal data becomes increasingly embedded in elite sport, understanding how athletes experience datafication is essential. This study underscores the importance of recognising athletes as active participants both on and off an increasingly digitalised sporting field.