Journal Archive

For service providers, previous studies have provided evidence of the positive impact of digital transformation on the business or organization. Internally, studies have focused on the improvement of digital technologies regarding business operational performance, such as automation, financial performance, and significant business improvements [7]-[9]. Externally, businesses or organizations are using digital technologies to enhance the service experience. Some industry studies provide specific data. Shen et al. found that omni-channel service strategies are adopted by whole retailers to enhance the customer experience [10]. Gunther et al. noted that data analytics are being widely used by internet companies in order to meet consumer demand and provide quality service [11]. As a whole, digital transformation is changing the ways in which the real economy operates. It has been suggested that new business models may emerge when all business sectors and functions are affected by digital transformation [12]. This transformation may be the creation of a new business model or the improvement of an existing business model [13]. The current study shows that digital transformation facilitates the iteration of business models and service experience enhancement in companies or organizations. On the consumer side, people are using digital technologies more extensively in their daily lives due to the convenience that they bring. The identity of consumers is changing as a result. Lucas et al. point out that digital technology empowers consumers to engage in value co-creation [14]. People are beginning to move away from being passive agents influenced by the service provider and are actively participating in service encounters [15]. Due to the proliferation of digital devices, the channels and agents that deliver services are iterating and technology-led interactions are gradually increasing. There has been a profound change in the ways in which consumers interact in service encounters. Examples include smart convenience stores and virtual fitness assistants [16]-[17]. As a result, consumers are starting to take more and more responsibility for their interaction behavior and service experience. Furthermore, these changes have generated the emergence of a new type of consumer—the digital consumer. von Leipzig et al. showed that digital consumers are more confident and better informed about their needs, and they set higher expectations for services [18].

For the service offering itself, digital technology has first changed the “human” element of the service. First, it is obvious that technology enhances the capabilities of employees [19]. In another case, technology replaces employees, and technology-led service encounters no longer require the presence of employees [20]. It has also been argued that employees and consumers are taking on new roles, rather than being at the heart of traditional services [21]. Schumann et al. suggest that there is no longer a clear boundary between consumer and employee roles due to the impact of technology [22]. In addition to this, the new features of digital technologies have also impacted the service experience. Scholars such as Lanzolla noted that products and services have the opportunity to achieve new features, higher reliability, and higher utilization due to the excellent connectivity, interactivity, and data computing capabilities of digital technologies [23]. These changes create new challenges for the evaluation of service experiences, which Stephanidis suggests are related to the means and object of interaction. It expands from explicit to implicit interactions and escalates from one-to-one to many-to-many interactions [24]. At the same time, in addition to “human-to-object” interactions, “object-to-object” interactions are also included [25]. The study of the behavior and experience of these interactions is usually focused on usability and user experience, two important concepts in the field of human–computer interaction. They are considered to be the key determinants of any product, system, or service for human use [26]. Currently, user experience research is widely disseminated and accepted, and the concept of usability is gradually regarded as a subset of user experience [27]. According to ISO standards, user experience aims to study “a person’s perceptions and responses resulting from the use and/or anticipated use of a product, system, or service”, including all users’ emotions, beliefs, preferences, perceptions, physical and psychological responses, behaviors, and accomplishments before, during, and after using the product or service [28]. Compared to usability research that focuses on user perceptions and user performance, user experience research provides a more inclusive perspective. It focuses on the user’s emotions or feelings during the interaction behavior and the intangible value that comes from the interaction behavior.

The rapid adoption of digital technology in the fitness sector offers the possibility of digital interventions and personalized support for gym-goers. In the application space, data tracking features such as pedometers, heart rate monitors, and oxygen saturation tests are being added to mobile and digital devices with IoT modules. An example is the mobile application WeRun, which allows users to see the number of steps that they take each day to exercise. In the research arena, several studies have shown the impact of digital technology on fitness behavior. Brickwood et al. evaluated the use of wearable devices in digital interventions. Wearable devices were found to lead to significant increases in daily steps, physical activity, and overall energy expenditure [29]. Katelyn Esmonde explored the relationship between data collection technology and runners’ intrinsic experiences based on the experiences of 10 young female runners using self-tracking devices [30]. Jeroen Stragier conducted a study of 717 runners regarding online fitness community use, and the results showed that the use of online fitness communities facilitated long-term progress in training [31]. These studies focused on investigating the use of specific digital technologies, while only a limited number of studies used an exploratory approach to examine user behavior in the fitness domain. For example, Uttara Ananthakrishnan et al. used data from Safegraph, Google Places, and other websites to compare the digital resilience of gyms and dental centers [32]. Meera Radhakrishnan et al. used data on users’ visits to gyms to explore people’s motivation to exercise, gym participation patterns, and reasons for dropping out of gyms in order to help gym-goers to adhere to their workouts over time [33].

In summary, research has been conducted in a number of areas to explore the impact of digital transformation driven by digital technologies that have changed the service experience and placed new demands on the evaluation of the experience. The important value of user experience in service evaluation is now receiving attention from industry and academia, and the role of digital interventions in the fitness field is beginning to be recognized by gym-goers. However, research on the user experience of digital fitness services is still in the exploratory stage. Most of the relevant studies focus on users’ use of a particular digital technology or digital device, while few studies have been conducted on the comprehensive experience of digital fitness services. The most important factors affecting the service experience in digital fitness scenarios have not been fully described. In terms of research methods, prior studies have mostly used usability testing to evaluate the service experience, collecting data and validating theoretical frameworks or hypotheses through structured methods. Two of the most commonly used methods are user testing and expert reviews. Users’ emotions, feelings, and expectations are rarely explored. Moreover, user behavior is constantly changing according to different specific scenarios [34]. Therefore, we believe that research methods that presuppose a theoretical framework or hypothesis may not be fully applicable to this study. Given the purpose of the work, we conducted an exploratory qualitative study to ensure the validity of the data obtained for studying the phenomenon in a specific scenario. The results obtained may serve as a useful addition to the existing quantitative studies. In contrast to previous studies, this study combined all temporal dimensions of user experience research to collect and code user behavior data in order to form a theoretical framework that can describe user experience pathways. This study focuses on the key factors affecting the service experience in digital fitness services and discovers the actual gap between the service experience and users’ expectations, so as to provide substantial theoretical support and suggestions for the optimization of smart gyms and their digital fitness services.

In China, places that provide digital fitness equipment and digital fitness services are called “smart gyms”. According to the design of Alibaba Group’s “Koubei Smart Gym”, before entering the gym, users can apply for membership, view class schedules, and book classes in the mobile application. Once the user enters the gym, the identity of the member will be automatically recognized, and the type of exercise, exercise data, and exercise trajectory of the user in the venue will be tracked and analyzed in real time to form a personalized fitness program and visual exercise analysis for the user. The smart gym is also equipped with smart showers, smart closets, and other related digital equipment. Smart gyms generally include four main areas: an aerobic training area (set up with treadmills, mountain climbers, and exercise bikes, etc.), a strength training area, a multi-functional gymnastics hall (for group classes), and an activity area for private instruction or stretching. Among them, the stationary fitness equipment is generally equipped with IoT modules and monitors. The multifunctional gymnastics hall and some areas have electronic screens for image and data display. There is also music playing on a loop throughout the gym. In addition, because of the influence of COVID-19, many smart gyms have also started to carry out online fitness services through mobile applications.

We used a procedural grounded theory research approach to build the theory [35]. Grounded theory is a set of qualitative research methods summarized by American scholars Barney Glaser and Anselm Strauss in the 1960s. Grounded theory allows researchers to collect primary data from literature, texts, interviews, questionnaires, and other sources. Afterwards, through sentence-by-sentence analysis and systematic induction of the primary data, a theory is eventually abstracted on the basis of empirical facts to address the serious disconnect between theoretical and empirical research that is prevalent in social science research [36]. Compared to deducing theoretical propositions from existing research, grounded theory focuses more on the reality of respondents’ experiences. There are fewer studies on digital service experiences in specific scenarios, especially fitness service scenarios, and the relevant research hypotheses or theoretical frameworks have not been widely explored. Therefore, this feature of grounded theory provides a bottom-up approach to theory building for our study. Moreover, due to the deep integration between the digital, physical, and social domains, for service experience evaluation work, researchers are required to develop a deep understanding of the real thoughts and behaviors of users in specific service scenarios. Grounded theory is a method for qualitative research and theory development based on the systematic collection and analysis of primary data [37]. This method can help us to extensively explore the deeper connections among user behavior data. For these reasons, we adopted grounded theory as our research method. The bottom-up theory construction helps to generate a descriptive theoretical framework to allow us to understand the user experience in digital fitness services.

The specific research process is shown in Figure 1. After recruiting the sample, we collected data as primary sources through semi-structured interviews. Data analysis of the primary data was completed through three steps: open coding, axial coding, and selective coding. The data were collected and analyzed in a continuous comparison. When no new concepts were found during the data analysis, the theory could be considered to have reached saturation. Otherwise, sampling was continued in order to supplement the data until theoretical saturation was reached. Eventually, our coding scheme passed the theoretical saturation test, and a theoretical framework was developed from it.

Fig. 1. 
Research process

In order to reduce the variation among the study cases, we selected six smart gyms in China that applied similar digital technologies through fieldwork, which were located in three cities in the northern, southern, and southeastern regions of China. The six smart gyms had different branding and visual identity systems, but they were all chains and were representative. As shown in Table 1, they offered similar digital services, such as automatic identification, mobile payment, wearable devices, and IoT fitness equipment. We adjusted the questions that might be covered in the follow-up interviews so that some of the service differences between the cases would not affect the results of the study.

After identifying the case gyms, we recruited two investigators in each city in which the gyms were located, who were graduate students with social research experience at local universities.

They served as hosts and recorders during the interviews and participated in the subsequent data review. During the interviews, the hosts were responsible for organizing the interviews and ensuring the respondents’ willingness to express themselves. The recorders, on the other hand, were required to organize and submit the content of the interviews transcribed by the speech recognition software. At the end of the experiment, the investigators were paid in cash. We eventually formed a six-person survey team and implemented at least weekly online surveyor meetings to ensure that the surveyors had the knowledge and research skills needed to perform their duties. Data collection began in February 2022 and was completed in August 2022. Investigators conducted face-to-face interviews with respondents near the gym. The length of time ranged from 5 to 20 min. During the control period resulting from COVID-19, interviews were conducted by online video conference. The length ranged from 10 to 30 min. Interviewees were reassured that their personal identities would not be disclosed. Interviewees were given a membership to a video site at the end of the interview.

Respondent sampling and recruitment were conducted simultaneously by three survey teams. To recruit respondents that fit the service scenario, we first conducted a small population-based purposive sampling procedure. Respondents were required to have visited a smart gym no more than 1 month prior to the study; to be willing to share their behavioral data and thoughts; and to agree to have the interview recorded. In the data from the first interview, we summarized the main service aspects of the smart gym and the users’ workout patterns, as shown in Table 2. It was found that respondents had different interaction behaviors and service experiences at different service stages. To support the development of the theory, we conducted a second recruitment step based on the theoretical sampling of users who visited the smart gym more than three times per month and used it for more than one month. The second sampling step focused on ensuring that respondents had a full-time experience of the digital fitness service. Moreover, data collection from users with frequent participation in digital fitness services was used to improve the detail and veracity of data in retrospective reports and to reduce the potential for general description and reporting bias. The number of recruits was determined by the principle of theoretical saturation. Theoretical saturation was reached in this study at the completion of the data analysis of 48 respondents. Subsequently, we continued to recruit 10 respondents and conducted interviews and data collection for a follow-up theoretical saturation test. The respondents consisted of 25 males and 23 females with a mean age of 28.23 years. The age distribution in the sample was under 20 years (8.33%), 22–30 years (54.17%), 31–40 years (33.33%), and over 40 years (4.17%).

The interviews were conducted according to the developed interview guide, which is shown in Table 3. The guide was designed to accommodate the purpose of the study and to ensure consistency in the direction of the study across the different survey groups. If the interviewees had any confusion about the questions, the host would clarify them. In grounded theory research, data collection and analysis were iterative. Therefore, without compromising the direction of the study, the hosts were allowed to refine the interview questions based on new concepts that emerged during the data collection process, thus guiding the interviewees to better express their ideas. After completing the interviews, the transcribed and collated interview data were summarized by the transcribers of each group in the form of electronic transcripts, which served as the source material for data analysis.

The same concept has different names in multiple versions of rooting theory. We adopted LaRossa’s naming of the individual concepts in the coding session [38]. Open coding is the first stage of data analysis and aims to build a concept–indicator model. Indicators consist of a word or a series of words, phrases, or sentences in the data. A concept consists of a label or name associated with one or more indicators [39]. In this stage, after line-by-line coding, we obtained 37 indicators and their corresponding 6 concepts.

Axial coding is the second stage of data analysis, which requires intensive analysis of categories (i.e., variables) in order to explicitly examine the relationships between variables [40]. In this phase, six concepts are abstracted into three variables. We describe the service experience path of the user in relation to the division of service stages in user experience research, i.e., before, during, and after the service. The variables are (1) overcoming resistance, (2) seeking new experiences, and (3) strengthening connections. The coding scheme is shown in Table 4. Due to the excessive amount of data from the original interviews, only the coding scheme after axial coding is listed.

Selective coding can be defined as the search for a “story line” [35]. Since each variable tells only part of the story, it is necessary at this stage to identify a core variable that can link all the variables to form an explanatory theoretical framework. In this study, we sorted out the interconnections among the three variables generated in the axial coding and identified “continuous value perception” as the core variable.

During the grounded theory study, theoretical saturation was considered to be reached when the source material simply repeatedly displayed existing concepts. To verify theoretical saturation, as well as the reliability and validity of the study, two additional investigators who were not involved in the coding review were invited to code the 10 transcripts used for the theoretical saturation test. The results showed that no new concepts emerged during the coding process. Twenty-nine indicators of general agreement emerged between the obtained coding results and the previous coding scheme, with a coding consistency of 0.78%. Therefore, we consider that the current coding scheme has high reliability. To further ensure the confirmability of the coding scheme, we sought confirmation from the interviewees. Through online meetings, we randomly invited the eight respondents who had participated in the interviews to review the coding scheme and invited additional comments to confirm that their ideas were reflected in the coding scheme. In most cases, the final coding results that were retained came from all of the above participants. Therefore, we considered that the coding results reached theoretical saturation and no further additions were needed. In Figure 2, the rectangular tree diagram is the visualization result of the coding scheme. The rectangles represent the nodes in the hierarchy, and the hierarchical relationships between parent and child nodes are expressed by the mutual nesting of the rectangles. The chart shows the frequency of agreement between the three selective coding results in the original data. The relative degree of agreement among the results is represented by the relative area of the rectangular regions.

Fig. 2. 
Rectangular tree diagram: nodes compared according to the number of coded references

29 chose trainer-guided, of which 22 used digital interventions. Moreover, 19 chose self-directed, of which 13 used digital interventions. An approximate 73% of respondents in the offline fitness scene used digital services. In addition, 21 people performed online fitness activities.

We collected respondents’ reasons and responses regarding their choice of workout pattern, as shown in Figure 3 and 4. Digital interventions are now widely accepted in trainer-guided workout patterns. Here, 22 of them used digital technologies. These technologies were focused on “session booking” and “recording” sessions to “gain easier communication skills” (77%) and “facilitate data analysis” (68%). The data show that digital technology has shifted some of the service tasks from trainers to technology, but the workout and workout analysis sessions are still trainer-driven. Trainers “can provide a more tailored workout program” (86%), which is the main reason that respondents continued to choose trainers over digital technology for assistance in their workouts. This phenomenon supports the interpretation that trainers still play an important role in technology-driven fitness services and that the value of the services unique to trainers prevents them from being replaced by digital technology, at least in the short term. In addition, the other main reasons are “stronger training supervision” (71%), “trainers are committed to training results” (57%), and “used to traditional methods and don’t want to change” (43%). It can be said that in addition to the customized services provided by trainers, service commitment and strict training supervision were found to be the most valuable services that respondents believed that trainers could provide. The results partially support and extend the findings of existing studies on the value of trainers [33].

Fig. 3. 
Reasons for choosing digital interventions (>10%)

Fig. 4. 
Reasons for not choosing digital interventions (>10%)

Of the self-directed workouts, 13 had digital interventions during the workout. “Workout data visualization” (87%), “automated data logging” (67%), and “more fun experience” (40%) were the main reasons that they adopted digital services. Respondents who chose not to use digital services explained that they “trust their workout plan more” (83%). In addition, respondents felt that “using digital devices while exercising is distracting” (67%), while others cited “not wanting to spend money on wearable devices” (33%). User adoption of digital services is a gradual process, with higher data availability and a fun service experience demonstrating appeal to respondents who work out through self-directed means.

In addition to offline workouts, 21 respondents reported that they had recently also been working out online. The main reasons cited by respondents were “online workouts are free from distractions such as COVID-19” (86%), “better access to trainers” (52%), and lower cost of ownership (43%). Other respondents indicated that they did not work out online because “the physical fitness environment provided a more focused experience” (67%) and “the intensity of the online workout was impersonal” (37%). Relatively speaking, the physical service scene continues to provide a stronger experience and memory for users [41]. However, for users who cannot work out in a gym, online fitness provides them with a new service scenario. Users can complete their workout programs at a lower cost and have access to more professional trainers through online channels.

A theoretical framework with “continuous value perception” as the core variable was developed in the study, as shown in Figure 5 . This process has a direct impact on the adoption of the service. When the user’s service experience meets or exceeds the perceived value, the user tends to stay with the service. Conversely, this may lead to intermittent withdrawal or even abandonment of the service experience. Perceived value in this context refers to the user’s subjective perception of the multiple forms of value that the service has. Moreover, the criteria for assessing this value are diverse and often change with context, time, and place.

Fig. 5. 
A framework of digital fitness service experience for smart gym users

Before the service, users need to overcome the resistance of “technical readiness” and measure the service provider’s service image in terms of security and reliability to establish a “trust relationship”. Afterwards, users experience the new service content through “perceived services” and different forms of “sensory stimulation”. After the service, the online service dominates the connections between users and between users and the service. Mobile apps extend the service lifecycle of digital fitness services and promote long-term usage by integrating the ability for users to “socialize online” and “work out online” with other users.