Between clicks and colleagues: the Janus-faced nature of data in the media industry

Abstract

This research delves into the dualistic nature of datafication, revealing its capacity to both illuminate and mask realities within media organizations. Datafication, the strategic use of data in organizations, offers insights into, and transforms, how work and individual outcomes are perceived and evaluated. This study examines how data-centric practices—embodied in metrics, rankings, and algorithms—can skew the playing field, creating scenarios where workers may feel (un)fairly (dis)advantaged or (mis)represented. The findings indicate that while data can provide valuable insights fostering cooperation and learning, it can also create competition and apprehension among journalists. Through the concept of “coopetition” we show how media professionals employ data opportunism to leverage information while also grappling with data skepticism amid potential misinterpretations and increased competition. Hence, the findings contribute to the research on datafication by demonstrating how workers utilize interpretative flexibility depending on what the data and metrics make visible.

Media and communication scholars have called for critical perspectives on how datafication shapes our social world (Couldry, 2020; Flensburg & Lomborg, 2023). As work will be increasingly measured, it is pertinent to understand what the widespread use of data communicates and how workers rationalize datafied environments (Treem et al., 2023; Vitak & Zimmer, 2023). This study responds to the call for a better understanding of workers’ attitudes and experiences related to workplace datafication (Holtzhausen, 2016; Vitak & Zimmer, 2023) by examining the rationalizations of datafication by journalists. Datafication plays a major role in journalistic work and influences how journalistic contents are produced, highlighting tensions between journalistic quality and commercial interests (i.e., click-based editorial work; Dodds et al., 2023) and the use of data in journalism (i.e., data journalism; Borges-Rey, 2016). Studies have suggested that the datafication of journalism introduces a gamified system in which “data and metrics become the main component to assess journalists’ capacities, through automated quantification and the competitive playfulness” (Ferrer-Conill, 2017, p. 706). It has been noted that the gamification of media work creates friction between work and play such that technological assemblages both empower and exploit employees, who seem to have little choice but to play or revolt (Ferrer-Conill, 2018). Importantly, while organizational communication researchers argue that the question is no longer whether datafication will change the way we work (Leonardi & Treem, 2020), the theoretical and empirical understanding of how datafication shapes employee experiences is currently lagging. This is an opportune problem as work and workers may sometimes be misrepresented or disadvantaged through data representations (Leonardi, 2021).

Broadly, the accumulation of data and its strategic use in organizations is a phenomenon referred to as datafication (Flensburg & Lomborg, 2023). Advances in computing technologies have led to a rise in digital monitoring (Newell & Marabelli, 2015), as mediated work environments produce a glut of data as by-products of work, which can be more easily monitored and analyzed using emergent technologies, including algorithms (Leonardi, 2021). However, beyond implications of organizational control, the datafication of work requires a focus on what is made visible and how through data, to understand workplace experiences and relationships (Zorina et al., 2021). Importantly, datafication makes work activities easily visible, necessitating a focus on how the visibility of work activities becomes aggregated, evaluated, and interpreted (Faraj et al., 2018; Treem et al., 2023).

In this study, we scrutinize the rationalization of datafication in a media organization. This approach allows us to explore the diverse, flexible, and possibly competing interpretations of data and what it represents. The performative visibility facilitated by datafication may incur a range of interpretations in organizational settings (Leonardi & Treem, 2020). For instance, in journalistic work, datafication through audience metrics is pivotal for assessing the interest in news and journalistic output (Christin, 2018, 2020). Real-time analytics foster competition among journalists through immediate viewership insights (Dodds et al., 2023; Petre, 2021), yet such data also allow journalists to tailor content to audience preferences (Moyo et al., 2019) and potentially learn from peers (Leonardi, 2014). Hence, importantly, datafication potentially introduces enjoyment through game-like features such as leaderboards but may also give rise to exploitation, surveillance, and control dynamics (Ferrer-Conill, 2018). Journalists may remain skeptical at best and fearful at worst in accepting user data as part of their daily work (Hendrickx et al., 2021). We explore the contradicting interpretations of datafication, highlighting elements of cooperation and competition that emerge from our interview data. This simultaneous presence of cooperative and competitive logics is referred to as coopetition (Brandenburger & Nalebuff, 1996; Corbo et al., 2023).

As such, we seek to make several contributions to media and communication scholarship. First, the findings highlight the dual essence of data and metrics in journalistic work by capturing the delicate balance between cooperation and competition arising from the use of data. These insights shift our focus from the ways in which datafication informs journalistic quality or editorial decisions to the ways in which data permeate the social fabric of news organizations, shaping journalists’ work experiences. In doing so, we highlight that datafication may signal various competing logics for workers (Treem et al., 2023). Second, we uncover that workers utilize interpretative flexibility to manage the conditions and consequences of datafication. The findings show that the visibility of data and metrics informs varied, often conflicting, rationalizations among journalists. Depending on the situation in which workers are confronted with data and metrics, and what the data represent, workers interpret their datafied work experiences through the frames of data opportunism and data skepticism.

Theoretical framework

Datafication in (media) organizations

As work and communication increasingly happen through technology, datafication (Flensburg & Lomborg, 2023), the process of collecting, analyzing, and visualizing information, has increased accordingly. Datafication has been found to play a role in quantifying work (Berman & Hirschman, 2018) and workers (Armstrong et al., 2023). It has been suggested that, along with the increasing datafication in societies, we should be concerned about whether the quantified data of a person’s behavior is able to paint a realistic, objective picture of their identity (Armstrong et al., 2023; Fritts & Cabrera, 2021; Moore & Robinson, 2016). This is important, as increased digitization and datafication have been linked to organizational surveillance and control, often following the metaphor of panopticon by Foucault (1977). The studies of organizational communication have applied the metaphor to identify the structural factors leading to digital surveillance (D’Urso, 2006) and to understand the effects of surveillance on issues such as privacy boundaries (Watkins Allen et al., 2007). However, datafication has increased the value of seeing and being seen (Treem et al., 2020), meaning that visibility of work can be beneficial to employees’ career development through possibilities to signal commitment (Cristea & Leonardi, 2019), and invisibility may be the one causing more concern for employees (Hafermalz, 2021).

In media organizations, datafication involves tracking a wide range of information about audiences, content success, and journalists themselves. Numerous studies have explored how audience analytics reshape journalistic and managerial practices. For example, editors now often base publishing decisions on audience metrics (Lamot, 2021; Tandoc, 2015), and journalists’ performance is evaluated through the metrics of their stories (Lamot & Paulussen, 2020). While increased audience insights can be beneficial, they also raise concerns about journalistic autonomy (Tandoc, 2014, 2015) and the potential for metrics to outweigh journalistic values (Neheli, 2018). Datafication is generally seen to negatively impact news quality (Fürst, 2020), intensifying conflicts between professional and commercial logics (Ferrer-Conill, 2015) and raising ethical issues surrounding data use and algorithmic journalism (Porlezza, 2024).

Within media organizations, the quantitative turn represented by datafication has given rise to gamified systems (Ferrer-Conill, 2017). Gamification refers to using game design elements (e.g., leaderboards) in non-game contexts (Deterding et al., 2011). Media organizations are rife with game design elements represented by audience metrics, rankings, and leaderboards (Dodds et al., 2023; Petre, 2015), which are designed and perceived as addictive (Petre, 2021). Importantly, while gamification may afford fun and enjoyment, it equally presents opportunities for exploitation, surveillance, and control. As illustrated by Hafermalz’s (2021) notion of a fear of exile, gamification can lead to a situation where the concern of being forgotten or left out overpowers the concerns regarding surveillance and thus, motivates employees to play along. Indeed, gamification in work environments introduces workers to a system ruled by sociotechnical norms embedded in daily practices that often do not afford an opt-out (Ferrer-Conill, 2017).

Research has demonstrated that game-like elements, such as performance ranking systems, can reduce employees’ willingness to cooperate (Chambers & Baker, 2020) and may even result in sabotaging colleagues’ performance for their own benefit (Hartmann & Schreck, 2018). Yet, other studies draw less bleak conclusions, for instance, Ferrer-Conill (2017) noted that: “while the journalists did not perceive the quantification and gamification of their journalism practice as something negative, the fact that system dynamics shape their production patterns is a cause for concern (p. 717).” Even though journalists might find game-like metric systems motivating and informative, the concerns lie in the notion that these systems have the potential to exacerbate potential conflict between journalistic values, the quality of journalistic work processes, and the commercial value of clicks and metrics.

In organizations, gamification is increasingly adopted as a form of indirect or “soft” labor control (e.g., Kellogg et al., 2020), promoting self-responsibility and self-discipline (Krzywdzinski & Gerber, 2021). Traditionally, game theory suggests that employees may be cooperative or adopt a more competitive mindset, for instance, refraining from information sharing (Lischka & Garz, 2023; Nash, 1951). We neither vilify nor valorize gamification. However, cognizant of the notion that “in journalism data has a more troubled history, given the long-established skepticism of journalists regarding (technological) change” (Porlezza, 2024, p. 1168), we adopt a critical perspective that uses gamification as a navigation pin to theorize about the often-opposing implications of datafication in organizations. The design principles of gamification often result in cooperation and interpersonal competition among those subjected to gamification (Zhang et al., 2023).

Rooted in game theory, coopetition refers to a situation that inextricably grounds logics of cooperation and competition (Brandenburger & Nalebuff, 1996). Coopetition highlights the importance of navigating the tensions presented by contending logics of cooperation and competition. This is important as workers may collectively resist gamification through datafication or seek ways to “beat the system,” while game elements such as leaderboards inherently introduce a degree of competition (Ferrer-Conill, 2017). Hence, while coopetition has not been widely applied to intraorganizational perspectives (Gernsheimer et al., 2021), it provides a framework for understanding whether and how datafication gives rise to intraorganizational coopetition. Newsrooms provide fertile ground to study datafication because data, such as audience analytics, have been an integral part of work for a long time (e.g., Petre, 2021; Tandoc, 2015; Usher, 2013) and are used in decision-making on several levels of the organization, including assessing employee performance (Lamot & Paulussen, 2020). Hence, our inquiry is guided by questions that focus on how journalists respond to the datafication of their work:

RQ1: How does datafication shape journalists’ work in the newsroom?

RQ2: How do journalists manage the conditions and consequences of datafication?

Method

The organizational site

Data were collected from a public-funded, Nordic broadcasting company, which operates an online news platform and several television and radio channels. The company employs over 3,000 people, about a quarter of whom are journalists. The organization relied heavily on audience metrics in journalists’ everyday work and the platforms providing the metrics were made available to all employees. Two platforms were mainly used to provide audience metrics: one that provided real-time metrics of clicks on each story, and one that provided more in-depth information about the audiences approximately 24 h after publication. Some employees also used additional metrics related to their roles, such as data on the organization’s social media activity. According to the interviewed employees, the organization did not provide information to employees on how the data about audiences were accumulated, nor was there formal guidance on data use in everyday work, other than encouragement to follow the metrics. As part of data collection, the authors had informal visits to the headquarters newsroom, where they discussed the policies and guidelines of data use with representatives of strategic management and communication and observed journalists working in the newsroom.

Interview procedure

Data were collected through semi-structured interviews to capture journalists’ individual perceptions, values, and beliefs (Tracy, 2020). We contacted the managerial representatives of the organization, and with their assistance, we compiled a sample of email addresses of employees in different journalistic roles. We reached out to employees via direct emails, newsletters, and announcements in departmental meetings. Participation in the research was voluntary for the employees, and we did not offer remuneration for participation. All interviewees received formal privacy statements, consent forms, and information on the use of the research data in written form. All data were pseudonymized starting in the first stage of analysis.

The first author conducted one-on-one interviews with 25 employees of the company. The length of the interviews ranged from 34 to 64 min, averaging 52 min. The interviews were conducted in the interviewees’ native language. The interview guide (see Supplementary material) consisted of four themes: (a) background information, (b) work and routines, (c) data accumulation and interpretations of data, and (d) changes within work practices through datafication. We asked interview questions such as “In general, what kind of data is present in your work?” From the answers, we derived follow-up questions, such as “How do you use this data?” or “What is the value of these data to you?” Additionally, we asked about discussions regarding data within the organization, the experiences of being evaluated based on data, and whether data informed interviewees’ work decisions, and if so, how. The semi-structured interview guide allowed us to formulate follow-up questions based on interviewees’ answers and on their different roles in the organization.

Participants

Of the 25 interviewees, 18 identified as female and seven as male. Interviewees’ ages ranged from 24 to 66 years, median age being 38 years. They had worked in the organization for nine years on average (range 0.5–25 years) and in their current positions for four years on average (range 0.5–21.5 years). The interviewees represented four types of workers: 13 worked as journalists, each with their own topic area of expertise. Four interviewees worked as layout editors, responsible for folding the online news site, basically, the order of the stories on the page and the final form of the headlines. Three of the interviewees held the title of producer, leading a team in a particular subject area, and five worked as managers, whose responsibilities extended to entire units or organization-wide projects. While the interviewees held a specific title, many of them had very versatile job descriptions and backgrounds. For example, some journalists substituted for layout editors, and many producers and managers had prior experience as journalists. Therefore, while our interviewees represent a variety of different journalistic roles, we refer to them as journalists when talking about them in general. Appendix Table A1 provides background information of the interviewees.

Data analysis

Data were analyzed using a qualitative, iterative approach (Tracy, 2020). The interviews were transcribed verbatim, resulting in 557 pages of written text (Times New Roman, double-spaced, 12 pt). Some of the interviews were transcribed by an external company and thus, the first author started the analysis by closely reading all interview transcriptions to check their accuracy. At the first stage of the analysis, the interview transcriptions were coded inductively, using the Atlas.ti software, to build a broad understanding of the data. The first author used descriptive, first-level codes (Tracy, 2020) to find the prominent and patterned themes, but also unusual insights within the data, keeping the codes informant-centric, close to the language and formulations of the informants (Gioia et al., 2013). Examples of typical first-level codes were “Comparing personal performance with others through the ranking system,” and “Using metrics as an opportunity for sharing support and positive feedback.” The second stage of the analysis was more informed by existing empirical evidence and theoretical perspectives from the fields of datafication (e.g., Armstrong et al., 2023; Treem et al., 2023) and journalism studies (e.g., Tandoc, 2015), as we compared our inductive findings to themes emerging from research. In other words, we engaged in axial coding (Strauss & Corbin, 1998) to come up with themes that deepened and refined our first-level codes by finding the linkages between singular codes and broader concepts. At this stage, first-level codes were transformed into second-level categories, such as “Comparing with coworkers,” and “Aiming for a mutual goal.”

Finally, by comparing the internal coherence of our codes and categories, we identified the final, aggregate dimensions (Gioia et al., 2013). The goal in this analysis phase was to come up with conceptualizations that would provide responses to our research questions, such as how comparing with coworkers and aiming for a mutual goal represented datafied practices. Figure 1 illustrates our data analysis.

Findings

Datafication shaped journalists’ work in the newsroom in three ways. First, the datafied work environment and its gamified elements created an optimal context for coopetition. Second, working in a datafied environment led to coopetitive practices. Finally, workers moved flexibly and situationally between and within the frames of data opportunism and data skepticism to rationalize their experiences around data and coopetition. Figure 2 guides the presentation of our findings and provides an overview of the perceived consequences of datafied work environment.

Conditions of the datafied work environment with gamified elements

Visual elements on digital platforms

The ways in which audience metrics were present in journalists’ work environments included several visual elements that encouraged comparisons and competition. These elements were often gamified and guided to follow certain comparisons that reflected organizational objectives. Although the systems visualized competition, interviewees did not generally describe them as problematic. Instead, they had decided to play along with the metrics and described the gamification of these systems as fun, eye-catching, and even addictive.

The digital platform all journalists used to monitor detailed audience data provided a default layout where the stories that best met the organization’s objectives were the first to appear in the rankings. Journalists described being able to manage their use of the system by selecting certain ranking factors themselves, which allowed them to compare departments, teams, and individual employees by organizational objectives. Additionally, the system allowed journalists to see the ranking of their own stories. Ann described the platform as follows:

We have this app, where everyone can see the total reach of the news and how many clicks stories have by a department and also by a journalist. I can also check how my own stories are gathering readers and how many clicks I have collected since the beginning of the year.

This way, journalists perceived that the ranking system made stories compete on multiple levels, although the default settings directed them to compete for the number one spot in the news.

The real-time metrics system provided several visual elements that made the platform look game-like. These included color-coded texts, balls, arrows, and graphs that visualized how the news platform of the organization was attracting audience compared to different benchmarks, such as the time of the day or place at the fold. Lydia, a layout editor, provided an overview of some of these visual elements:

Let’s say I have a story on the third place of the fold, which is relatively high up. There’s a little ball that tells me that okay, it’s green. It’s a good thing that it’s green, but if it has a number 15, it tells me that this story attracts readers as well as if it were in the 15th position. So not very well considering that it’s at the third place in our fold, so a lot of people see it. There are even nice arrows that show up and down, to simplify this.

While the real-time metrics system was provided to journalists as a browser version and phone app, for some employees it was even more visually present. There was a large screen in the newsroom at the organization’s headquarters that displayed real-time audience analytics. Being constantly around the changing metrics was described as eye-catching, especially for those whose desk was located near the screen. Even those for whom real-time metrics were not a central part of their job described checking the screen regularly. Susan said:

We have a big screen in the newsroom, where you can see the real-time data all the time. In practice, it shows how many people are reading each story at the moment, and it changes all the time. Because that information is there on the big screen, I tend to check it constantly although it’s not necessary for my work.

The ever-changing real-time metrics system was noted as “addictive” not only by those who could see it on the screen all the time but also by other journalists. Oliver described his use of the system: “For us, who like numbers, it’s very addictive, it’s almost like a game in a sense that you know you have a good headline and then you can see the story really taking off.” Other interviewees also said that since the system was provided as a mobile application, it encouraged them to check it constantly, especially if their story was recently published and they could follow how it “takes off.”

Communicative elements

In addition to the visual elements on digital platforms that fostered competitive environment, journalists described two ways in which communication within the organization reflected the gamified working conditions. First, journalists used headline tests to compete against themselves and each other. In headline testing, multiple headline options were developed for the same story, and their effectiveness was tested upon publication to come up with the headline that attracted the most attention from the public in a given period of time compared to the other options. Audience behavior in relation to the different headline options was measured both in terms of volume and whether clicking on the headline also led to reading the story. These parameters determined the “winning” headline, as Mikael put it: “The system tells us ‘Ping, this one won.’ This headline got the most clicks and especially quality-clicks, so people not only clicked but stayed to read the story.” This system was often used as a game between colleagues by betting on a winning headline or writing different headlines to be tested. Oliver gave an example:

I recently discussed with a young intern of ours. It was actually fun, we had a polite, friendly debate about a headline. I thought they had a bad headline, and they thought my suggestion was a bad headline. So, we tested it, and it turned out that their idea, which I thought was bad, was three times more successful than my idea.

While this kind of behavior was described as a friendly competition, headline tests also enhanced cooperation as testing headlines together were seen as an opportunity to make the story more successful.

Another communicative element that created a context for coopetition was the “hit limit” discourse. The interviewees talked about the number of clicks on a single story that determined whether it was successful. These click-based limits were communicated to employees at meetings, in performance appraisals, and in intranet bulletins. Ann described the limits as follows:

If you get 15,000 readers under the age of 29, that’s good. If it’s like 20,000 or 25,000, you’re doing an amazing job. If you get over 30,000 it’s seriously good and if it is over 50,000, we are talking about a real hit.

While exceeding a certain number of clicks was considered a success, lower numbers led to a more critical attitude towards the topics of the stories. Jenna stated: “If a topic we considered significant does not attract more than a few thousand readers, we start to think that it may not be worthwhile to cover in the future.” However, exceeding the “hit limit,” in other words, reaching young readers, was the indicator that was considered the most meaningful. Whereas exceeding (or falling below) any other number of clicks did not automatically lead to rewards or sanctions to an employee, gamification was emphasized by the fact that exceeding the hit limit resulted in the author of the story receiving a monetary incentive. Berta, along with other interviewees, described, “When an article reaches the hit limit, you get the incentive automatically.” This way, exceeding the hit limit also constituted its own kind of game in which getting the immediate incentive could be seen as winning the game.

Consequences of the datafied work environment: coopetitive practices

The conditions of the datafied work environment created consequences for journalists in the form or coopetitive practices. The interviewees perceived that the use of data led to complex outcomes in which there were some recognized benefits, but the excessive focus on audience metrics came with notable drawbacks. Two coopetitive practices emerged: comparing with others and aiming for a mutual goal.

Comparing with others

The ways in which data were used to make comparisons within the organization constituted coopetition. As the digital systems allowed employees to check different rankings between departments, journalists, and stories, consequently, they used the system to compare their stories’ rankings to others, as Jenna explained:

I can check the metrics of my colleagues, which is interesting to do. I don’t do it every day, but if one of my own stories has been published, I want to see how it ranked, how popular it was compared to others.

Comparisons were also made when the data from one’s own stories did not meet the organization’s objectives, highlighting competition between teams, as Julia, a producer, described:

If all the stories have quite small numbers, and our story has been the most read of the day, it helps to put it in context. In other words, the number of clicks on a single story doesn’t say much if you don’t compare it with the total number of clicks on our site.

Furthermore, feelings of success and failure at work were related to how well one’s story did in comparison to others. Ann described her motivations to follow audience metrics through the experience of success: “Maybe it’s related to some basic human need, that good feeling when you see that the story is successful and has a lot of [readers]. It’s kind of an endorphin: how wonderful, that [story] was successful, and I am successful.” Many other interviewees aligned with this, saying that if their stories were read by people, it indicated that they as journalists were successful. However, when one’s own story did not do well in comparison to others, it led to journalists perceiving themselves as less successful than their colleagues. As Eva put it, “It’s easy to have that negative feeling, like, that person is successful because their story was popular, whereas I am a failure because mine was not.”

Sometimes comparing personal metrics to colleagues’ metrics was more directly communicated. In these situations, journalists explicitly perceived as competing with their colleagues. Julia described, how “It’s very important for some people to write the most successful stories. They aim for holding the prize of the most successful one on the team.” Krista described how the explicit comparison made her feel uncomfortable:

Someone asked me, what is my most read story of all times. I felt it was a strange question. Like, I do know the answer to that question, but I’m not sure why we are discussing this. Is this some kind of competition where whoever has the most successful story is the best journalist?

Although Krista found it strange that someone directly asked her about her metrics, she also talked about checking others’ numbers. She described feeling ashamed of making the comparisons, which was related to data envy described by several other interviewees. In these situations, journalists were concerned that they were not performing as well as their colleagues, and the feelings of envying others’ success encouraged them to compare their data to others’ data. As Krista said:

I must be honest: I have sometimes compared my metrics to others. Only because I sometimes feel like I have done less than others, or I want to know where I stand in comparison. I’m a bit of ashamed for it. I think it’s embarrassing that I have checked others’ metrics.

While comparison to others through data was perceived as fostering competition, comparable data were found useful when it allowed journalists to learn from their colleagues’ success.

Instead of envy, some interviewees talked about admiring their colleagues and checking their most-read stories to learn from them. Mikael explained:

I check the metrics of others every now and then to understand how some of my colleagues’ stories are so popular. [-] I think very highly of a couple of writers here, because it seems like whatever they do, it becomes a massive hit every time. Of course I’m interested in how is that possible.

Whereas some of the interviewees talked more about checking others’ metrics for learning purposes on their own, data were even more linked to cooperation in situations where learning from others was organized formally. Sofia said that when a story from her department became popular in the light of statistics, the people making the story were asked to analyze why they thought their story attracted readers. She described this method as follows:

In their meetings, managers and producers go through the most read stories so that journalists and producers could learn from successes and recognize what are the things that work. If one of the news desks creates a big hit, they are often asked to describe their process: What happened and why do they think the story was successful.

Although these kinds of situations seemed to be directed to enhance cooperation, coopetition surfaced as teams were assigned to educate each other in terms of who was seen as successful.

Aiming for a mutual goal

Along with finding the datafied work environment as enhancing comparisons, journalists mentioned working towards a mutual goal of making stories attractive for readers, or oftentimes, to reach the hit limit in terms of the number of clicks. This represented coopetition, because while aiming for making stories successful was often perceived as joint effort, credit and incentives were typically targeted to journalists who wrote the stories, not to others who contributed by changing the headline or putting the story on the top in the fold. While the possibility to see how colleagues were performing encouraged journalists to compare themselves to others, the same possibility also served as an outlet for sharing peer support and positive feedback. Turning the focus into succeeding as a team was also a strategy of some producers. As they had the possibility to distribute topics to journalists, they aimed to do that in a way that allowed as many as possible to have a presumably popular topic to cover. Oliver explained:

We know more or less what makes a hit. To give everyone the experience of making a hit, I, as a producer, have to know our projects so that everyone can be given projects where they can really succeed. When we work together to make a hit, it’s likely that one of those stories will eventually become one. Through that, we get experiences of success and a feeling that everyone is involved, not that there are the best ones and then there are the losers.

Although employees, especially those with managerial responsibilities, put effort into fostering cooperation rather than competition, the fact that the author of the story was the one getting credit and the possible monetary incentive created coopetitive tension between those who worked under the title of journalist and those who worked in other positions. However, the gamification of data through the discourse of hit limit encouraged layout editors and producers to work toward reaching the limit even though the incentive would go to the author of the story. In Sofia’s case, not getting the incentive from reaching the limit did not stop her from working towards achieving it:

If I see potential in a certain story to go over that hit limit, of course I as a [substitute] producer aim to reach that. I might write a better headline or ask the layout editor on duty to run a headline test when the number of readers starts to slow down, to give the story that extra boost. Of course I do things like that, but I don’t get any personal benefit from it.

This example indicates that the possibility of reaching the hit limit was motivating in itself. On the contrary, layout editors could feel like they were letting their colleagues down if they were unable to make the stories successful with their work. Lydia described: “It feels bad if the data is in red all day. Even though you know you are not solely responsible for it, but because you are the one who kind of serves it [to the public], you feel responsible.” Some interviewees mentioned that journalists’ desire to reach the hit limit, maybe even to “win” their coworkers’ analytics, led them to put pressure on layout editors to put their stories on top of the fold. Several layout editors said that if a journalist had noticed from data that their story was attracting attention but was not put on the top of the page, they contacted layout editors and used data as a justification to ask them to put that story on top. In these situations, the competition among the journalists created tensions between journalists and layout editors, who perceived the journalists’ demands as questioning their professionalism.

The navigation mechanism: flexible framing of the meaning and value of data

The experiences of professionals in different journalistic roles showed that datafication fostered coopetition among organizational members. However, the ways in which employees sought to rationalize the data showed that seeking alternative explanations for the data helped them navigate the datafied work environment and the coopetition it caused. We call these strategies data opportunism and data skepticism, in which journalists sought to determine the value and credibility of data through situational attribution and the strategic use of ambiguous signaling. In data opportunism, the ambiguity of data was interpreted in a way that enhanced one’s position in a coopetitive environment or allowed them to continue using data as they preferred, while in data skepticism, coopetition through data was perceived as a threat to one’s position, leading to criticism of data or avoidance of its use. These rationalizations were dynamic interpretations, enabling workers to shift between data opportunism and data skepticism and experience a sense of neutrality in the transitional space between them.

Data opportunism

When data were showing that a journalist was succeeding in attracting the audience, workers did not pay a lot of attention in explaining the data. When the interviewees described situations where their stories or their team had been successful, they described it as succeeding in being relevant and interesting to the audience. In these situations, the interest of the audience was seen as determining the quality of the work. Alex described: “In my opinion, the numbers correlate with quality: If you have a poorly focused story that people don’t find interesting, maybe the quality is not good.” Similarly, Janne, a manager, justified his use of analytics to give feedback to his subordinates as follows: “It’s amazing how often audience finds the good stuff. The data is seriously a quality indicator, it’s not just a number.” Also, when interviewees experienced that they could learn to improve their work by looking at colleagues’ successful stories and trying to take ideas from them, data were perceived as accurately indicating good journalism.

Some interviewees engaged in data opportunism by choosing to follow other data than audience metrics provided by the organization to find themselves successful and to communicate that to others. For instance, Hugo said that he checks regularly if his stories have been used as references for Wikipedia articles, whereas Berta said:

For us, a good result can be that [our] story is presented in other media, or if there is a story that may not be successful in the light of audience metrics, but the government decides to meet because of it, and it leads to a change.

Often in situations where journalists felt that the data did not show them as successful, they would seek alternative data to demonstrate that they were. This illustrates data opportunism in deciding which data to follow or show to others, but also shows how important a role data were perceived to have in the evaluation of employees.

Data skepticism

On the contrary to data opportunism, when data showed that a journalist, a team, or the entire organization was not reaching the audience, employees found a lot of explanations for how data could be rationalized as simplistic, vague, or unpredictable. This way, they strategically used ambiguous signaling around the data: The fact that they did not know the background or accuracy of the metrics left room to choose their own interpretation and reduce their perceived value. The alternative rationalizations they used in situations where the data were not in their favor aimed to reduce competition as a whole (because these data are unpredictable, no comparison should be made at all) and downplay others’ successes (because these data do not correlate with quality, the person with better numbers should not be considered better than me), indicating data skepticism.

In some situations, journalists aimed to find the reason for poor numbers from the behavior of the audience. For example, the interviewees stated that if the weather was sunny, people simply would not read any news and thus, stories published that day would not get readers whether they were interesting or not. This kind of rationalization was used also by the managers in explaining why the entire organization was not reaching the numbers that would meet organizational objectives. Thus, rationalization worked as a way to avoid comparisons based on data altogether. Additionally, the value of the data for assessing the quality of certain stories or employee performance was downplayed by explaining how audience could be interested in “irrelevant” things, or by emphasizing that the data did not inform why a particular story was popular. Suggesting that one’s own story that gathered fewer readers might have a higher qualitative value than someone else’s story that was popular, indicates an opportunistic approach to the ambiguity of the data, although the attitude towards data use itself indicates skepticism. Mark described:

We know how many people have read the story, but we don’t know if someone has ‘anger-clicked’ a story open, thinks it’s trash, and is angry for spending time reading it. And then, maybe a story that has attracted fewer readers has had someone experiencing a huge social awakening because of it?

As this example illustrates, journalists used the ambiguousness of the data to downplay the value of numbers when they found their story was not successful even though they personally found it important.

Sometimes rationalization of data was done by focusing on the structure of the organization. The interviewees explained how journalists could not affect the layout, departments could not decide on which topics to cover, and data were unable to illustrate the amount of work put into a story. The coopetition between layout editors and those in other roles was often present in the ways in which data were rationalized, by especially journalists stating that the story they had written was not successful because it was treated “badly” regarding publishing time or place on the fold. Hugo explained:

The number of clicks says nothing about the quality of the story. It can indicate whether your headline was good, how long the story has been on the front page, or what time it was published. These are all things that, well, the headline is something I can maybe influence, but otherwise they are things I cannot influence myself. I cannot put my story first on the front page.

While journalists often explained stories that were not successful by the actions of others in the organization, layout editors used the ambiguousness of data as a way to explain to themselves—and to the journalists—that data did not reflect the amount of work put into the story, whether by the journalist or by the layout editor. As Sofia said, “It just doesn’t always help—whatever I do, people may not read the story.” In these situations, the unpredictability of audience interest was emphasized to highlight the fact that seemingly poor data did not indicate poor quality.

Finally, to argue against the competition between departments or teams within the organization, employees engaged in data skepticism by saying that certain teams had more topics to cover, or topics of a certain team were more likely to be clicked. Krista talked about how data were reflecting the number of topics journalists could cover and not the skills of a journalist:

To me, data says more about the volume of people doing stuff. There are surprisingly big differences within the organization, in terms of what the volume is and what kind of topics people write about. I don’t think that the people who have good numbers are top journalists.

Overall, the ways in which workers rationalized the data present in their work environment indicated that the use of data was opportunistic, meaning that data could be interpreted and valued according to what was most useful to the person or their team. When this was not possible, journalists turned to data skepticism, trying to find reasons not to use the data based on its shortcomings and ambiguity.

Discussion

The aim of this study was to increase understanding of the complex phenomenon of datafication and the competing logics surrounding its interpretation in organizations. We were especially interested in how journalists respond to the impacts of datafication and the intense quantification of their work. This study contributes to the emerging literature on datafication in organizations by providing empirical insights into the competing logics of cooperation and competition and associated coping strategies guided by contrasting frames of data opportunism and data skepticism (see Figure 2). In our data, these rationalizations helped journalists manage the datafication of their work, either glorifying or disparaging data and what it represents. We choose to frame these responses as “rationalizations” because editorial staff are increasingly confronted with a datafied work environment—an environment that has not emerged through their own initiative but is instead shaped by management’s changing demand for data-driven decision-making. This shift places journalists in a position where they must manage the opportunities and challenges of datafication, often without having had a say in its imposition.

Having audience analytics visible to everyone on a large screen in the office and on employees’ mobile devices provides a possible context for panoptic awareness, where the presence of a control-allowing system creates a sense of being under constant surveillance (Foucault, 1977). Our findings demonstrate a context where visibility does create pressure, but also invisibility leads to negative experiences, such as injustice. The incentive mechanisms, which were based on rewarding the journalist who wrote the story for getting clicks on it, failed to acknowledge the work done by the layout editors. Our findings speak to the need to understand visibility not only as a possibility for surveillance and control (e.g., D’Urso, 2006), but also as an opportunity to address the fear of exile (Hafermalz, 2021), where being seen is desired and the greater concern is about being forgotten or ignored.

The growing presence and importance of audience metrics have been well demonstrated in previous research (e.g., Christin, 2018; Petre, 2015, 2021; Tandoc, 2014, 2015), but it has also been suggested that in public service media organizations, data could be ignored (Ferrucci, 2020; Karlsson & Clerwall, 2013; Usher, 2013). In contrast, our results illustrate a case where employees in a public-funded media organization make extensive use of analytics, and the datafication of journalistic work gives rise to coopetition also in a public-sector organization. First, our findings show that data provided a framework within which journalists competed against each other, as well as against themselves, by aiming for first place in rankings or fighting to get statistics “green,” i.e., exceeding a set threshold. Second, the datafication of journalistic work provides insights into “what works.” Our findings align with prior research (e.g., Lamot & Paulussen, 2020; Petre, 2015) in that journalists use insights from analytics to try and decipher a recipe for success. Succeeding in attracting the audience evoked positive emotions and could lead to a monetary incentive for the journalist who wrote the story. When one succeeded, others, in turn, could feel envious. Yet, successful stories often require interdependent work: A “good story” requires an interesting headline, successful content, and an optimal position on the fold, all of which were the responsibility of different people. Hence, the datafication created a fertile ground for coopetition.

Our findings contribute to the literature on datafication in organizations by illustrating the different, often contradictory rationalizations of data within an organization. Through the widespread adoption of analytics systems, journalistic work represents a context in which datafication provides exceptional visibility into individuals’ work and outputs (e.g., Lamot, 2021; Petre, 2021). This visibility gives rise to opportunities for vicarious learning (Leonardi, 2014) and cooperation among colleagues. At the same time, the visibility into individuals’ work and their outputs bred a rivalry and gave rise to a competitive environment. Understanding this is key for management in journalistic organizations, who, in their strong belief in audience analytics, might be blind or ignorant to the negative effects of datafication among journalists.

Furthermore, our findings demonstrate that journalists cope with the complex logic of datafication depending on the specific context by adopting framing practices of data opportunism or skepticism. Our interviewees strategically used the ambiguity of what data represent because it allowed them to interpret data differently in different situations. Treem et al. (2023) argue that ambiguous signaling of data is a risk to employees if they are not able to communicate their actions through data due to different interpretations and validations of it. In our study, journalists rationalized their interpretations of data and metrics depending on the result. For instance, data opportunism was evident by celebrating the value of data and audience metrics when they were in favor of certain journalistic outputs.

Ironically, the same data and metrics were relativized, second-guessed, or even discarded as superficial or misguided when indicating poor performance of journalistic outputs. This way, journalists utilized ambiguous signaling, as it enabled them to decide upon how to rationalize and make sense of data in different situations. This indicates that while data were strongly present in communication across the organization through celebrations of success stories, flexibility in the interpretation of the value of data was important for workers to cope with public scrutiny of individual performance through visible data and metrics.

Our findings contribute to the literature on coopetition, providing empirical evidence for coopetition logics within media organizations. Prior research on intraorganizational coopetition has typically focused on teams, organizational units, or subsidiaries (Gernsheimer et al., 2021), but we argue that more focus should be put on studying individual employees’ experiences of coopetition. As work will be increasingly measured in the future (Treem et al., 2023), the consequences of such change in work relationships require new kind of knowledge and skills to manage the boundary between cooperation and competition. Our findings suggest that the flexibility of interpretative logics associated with the gamification of work helps employees in coping with the “wins and losses” made visible through datafication.

The findings of this study need to be considered in light of several limitations. First, data were collected in one case organization, which limits the generalizability of the findings. In addition, while journalistic work provides a promising context for studying datafication for its long-standing embeddedness of data and metrics, the distinct nature of journalistic work may also present unique operating dynamics. Hence, our study provides an opportunity for further research into datafication and worker interpretations across organizational contexts and industries. Finally, we have relied on interview data to generate insights into data interpretations. While semi-structured interviews provide a deep insight into the experiences of individuals, relying on interviewees’ self-reported perceptions can overlook some relational dimensions within the organization. The findings demonstrate that data interpretations are highly contextual. As such, future research might benefit from utilizing ethnographic methodologies and observation of situations in which data interpretations take place in organizational communication to generate richer descriptions and understanding of situated data interpretations.

Supplementary material

Supplementary material is available at Journal of Computer-Mediated Communication online.

Data availability

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. The data will be shared on reasonable request to the corresponding author.

Funding

Funding support for this article was provided by the The Helsingin Sanomat Foundation (20220085).

Conflicts of interest: The authors declare that there is no conflict of interest.

References

Appendix