Abstract
The game industry's content production, maintenance of live games, and processes of acquiring production funding increasingly rely on various kinds of data and its rigorous analysis. These new needs and functions have generated emerging forms of work, such as those of the data analyst, data engineer, and data scientist. Through in-depth interviews with 20 Finnish game industry professionals and an analysis of game industry job advertisements, this paper examines the work and identity of game industry data workers. Drawing from scholarship focused on game production, game work, and data labour, this article argues that organisational practices surrounding data professionals reveal the centrality of high-level data work in game studios focused on live service games and that data work is now performed not just by data analysts, but by the entire staff and management. As a precursor to the wider creative industries, we argue that creative work and data work in game companies are gradually converging, due to the datafied work environment facilitating datafied game work and the work of data professionals increasingly intertwining with creative tasks. Complicating the previous game studio hierarchy is the data analyst's dual role as both a subservient support function and a central broker of data. Adding nuance to this, the article argues that an important aspect of the work of bespoke data professionals in game companies is communication, in contrast to the high-level quantitative tasks often associated with analysis.
Keywords
Introduction
The objective of this empirical study is to contribute to the current understanding of data work through an examination of data analysts and other data workers employed in the digital game industry. In their current form, mobile games have become networked services that are run around the clock and require continuous company input (Kerr and Kelleher, 2015). The industry's content production, maintenance of live games, and securing continued investment increasingly rely on many types of data collected from various sources, and following this, from rigorous data analysis. The article introduces the idea of ‘game data work’, focusing on the micro-level data work (Foster et al., 2018) in game studios. By assessing the data-intensive tasks and processes of current game studios, the study contributes to the larger discussion on how constant data flows and data analytics transform the nature of work in creative industries and aligns with calls for research on how datafication impacts decision-making and the definition of work roles in knowledge work (Fischer and Wunderlich, 2021).
The demand for data and the subsequent need for new services have generated new forms of work and novel positions, like those of the data analyst, data engineer, and data scientist. While previously the work input of such dedicated data professionals (DDPs) was often bought from data analytics companies or freelancers, nowadays these new professions can increasingly be found as not only integrated but also crucial parts of the game studio structure. In many ways, in the past decade, data analysis has emerged as a core skill in game development (Kerr, 2017; Van Roessel and Švelch, 2021; Whitson, 2019). To better understand the new forms of game data work and the consequences of them becoming increasingly central for game production, this study asks:
How is game data work currently organised and communicated in game companies, and how do these processes reconfigure game work? How does a close analysis of game data work challenge the existing theorisations about the role of data analytics in creative industries?
The study examines the two interconnected research questions through two data sets: game studio job advertisements and thematic interviews. Game job advertisements are used to illustrate the variety of tasks associated with current-day game data work. Interview data are used to deepen our understanding of game data work in general and to contrast the ‘official’ narrative of the job advertisements with more personal accounts of everyday data work, also paying attention to any kind of invisible data work in organisations (see Møller et al., 2020). By focusing on the everyday interactions and organisational contexts of data work, this study builds on Foster et al.'s (2018) framework of multi-levelled data work. With a focus primarily on the micro-level of data work, the article goes on to show that, while the intensity of data work varies between positions, data work in the creative industries is now extending to all areas of work and to all workers. As such, the study invites a re-evaluation and expansion of the concept of ‘data professional’.
Game companies worldwide are these days inevitably tied within the context of global app economies, data-driven production models, and game development that has shifted from fire-and-forget commodity development to launching continuously maintained and updated services. Despite the centrality of data in new production models, games-specific data work and the roles of data professionals have only been examined in a handful of studies (Van Roessel and Švelch, 2021; Weststar and Dubois, 2023). Like Carter and Sholler (2016: 2317), we believe that by analysing data professionals, who ‘must navigate a complex array of technologies, organizational structures, and social influences’, we can inaugurate new information on emerging communication practices and work cultures within the game industry. Both platformisation of cultural production (Nieborg and Poell, 2018) and the emergence of phenomena like streaming (Taylor, 2018) and esports (Johnson and Woodcock, 2021) highlight how ‘games are now a central component in the convergence of media content, media platforms and technologies, and media audiences’ (Chess and Consalvo, 2022: 159). Videogames and the industry around them have a potential to introduce data-intensive practices and business models that are quickly adopted in other fields. In this respect, studying games can highlight developments that are highly relevant to various creative fields.
This study is focused on the Finnish mobile game industry, known for its various success stories in the data-driven free-to-play business model. As highlighted by Andrejevic et al. (2015), the non-transparent and privately owned nature of data makes it difficult for scholars to gain access to the actual production processes. Here, the Finnish game industry context, famous for low hierarchies and willingness to share information (Sotamaa, 2021), helps, as industry professionals from different backgrounds were eager to participate in the study. While Finnish companies form only a small node in the global circuits of game production, the Finnish industry, generating an annual turnover of over €3 billion, has been often mentioned as the pioneer of the free-to-play mobile games market. This is exactly the industry branch that has been most aggressively developing data-intensive development practices. Therefore, the lessons from the Finnish game companies will provide a timely basis for the larger scholarly discussion around data work in the creative industries.
Background of the study
Today, most, if not all, progressive organisations use some sort of data to inform their business-related decision-making and shape their design processes. This development has been so powerful that current social imaginaries present data utilisation as a dominant operational regime that is both effective, smart, and almost inevitable (Beer, 2018; Turow, 2017). At the same time, surprisingly few studies have focused on the individuals doing the data analysis work. As Ribes and Jackson (2013: 152) remind us, at the end of the day, the flashy promises of data analytics visionaries come back to ‘a squeamish student taking spit samples’ or ‘a frustrated field technician recalibrating a vandalized weather monitoring station for the third time that month’. With this study, we want to add a game worker perplexed by a data analytics system to this list of examples. Therefore, we first need to understand the nature of data work in general and the game studio as a specific context for data work, in particular.
Roughly described, data work can be characterised as a set of processes that include data-related organising, analysing, judging, and decision-making (Foster et al., 2018). While data work is arguably about turning data into action, it is important to understand that ‘[i]n order to make data “actionable”, numerous data streams must first be collected, cleaned, combined, and then ‘mined for insight’ (Whitson, 2019: 794). Various data professionals participate in constructing ‘algorithmic systems for computerised quantitative data analysis’ (Avnoon, 2021: 336), and these activities can include many forms of ‘data science work’ and ‘data analysis work’ (Carter and Sholler, 2016). Therefore, following Kandel et al. (2012: 2918), we define a data analyst as ‘anyone whose primary job function includes working with data to answer questions that inform business decisions’. At the same time, as Foster et al. (2018: 1423) importantly point out, data work is not reserved only for DDPs: A number of professional groups play a direct and indirect role in the micro context relevant to data work. These include workers directly involved in the process of realizing value from data, including data scientists, data architects, data analysts, data visualization experts, and other data professionals; along with a range of end-users who make decisions and take actions on the basis of data. To these groups, can be added those with a more supervisory and strategic interest in data work, including records managers, information governance managers and officers; data protection managers, IT risk and governance managers, and IT security personnel. Interest in realizing value from data is not restricted to data and information professionals, but also includes end-users and other groups.
Andrejevic et al. (2015) have shown how a cultural studies approach can play an important role in explicating the social and political consequences of data analytics. At the same time, they remind us that while cultural studies can be good at picking up on cultural trends, data processing practices can pose challenges to this approach. Therefore, our project draws inspiration also from the production studies tradition (Banks et al., 2015; Sotamaa and Švelch, 2021) that puts an empirical focus on the wide array of processes, practices, and technologies associated with current-day media production. Previous studies have highlighted how audience metrics and other forms of data analytics affect other creative fields like the television industry (Kelly, 2019), the streaming industry (Rasmussen, 2024) or the music industry (Maasø and Hagen, 2020). We will discuss the findings in connection to these studies and explicate the similarities and differences between other close fields.
If we want to understand the specificity of ‘game data work’, we first need to discuss what ‘game data’ actually means. El-Nasr et al. (2021: 3) describe the different games-specific data forms as follows: A great variety of data can be collected, stored, analyzed, and leveraged to gather intelligence throughout the lifetime of a game title or game company. Typical sources of data include behavioral data from games, information from advertising partners and other third parties (i.e., social media platforms), and data collected from infrastructure (such as servers), the development process itself, marketing, and user research.
Dorschel (2021) notes that job advertisements and universities portray the data scientist role as a ‘hybrid’ professional who is both a generalist and a specialist, skilled as both a technician and a communicator. Avnoon (2021) conceptualises data scientists as ‘omnivorous’ workers, who continuously strive to acquire new knowledge and to master several fields sufficiently enough for the required task. With their ability to bridge technical and social skills, data professionals are reported to disrupt traditional workplace hierarchies and consequently challenge high-end professionalism centred just on a single field. In the game industry, the traditional core competencies have included those of game design, game art, and programming. While additional roles like community management (Kerr and Kelleher, 2015) or game testing (Bulut, 2015) have appeared, they have seldom threatened the dominance of the classic triad. However, there are signs that game data professionals have quickly become highly central for game production (Sotamaa et al., 2024), and therefore we need to pay special attention to how game data analytics shape the industry power relations.
In this research, we approach power primarily through work hierarchies, specialisation, and division of labour. According to prior studies, the growth and maturing of the game industry have led to more intricate division of labour, more tightly managed processes and more hierarchical production pipelines (Deuze et al., 2007; Izushi and Aoyama, 2006). Thompson et al. (2016: 324), who studied autonomy in Australian game studios, reported that specialisation and more fragmented division of labour had led to a situation in which ‘work was now more technical, more driven by the hardware and therefore less creative’. While we agree that the datafication of work can further accelerate these developments, we will also explore how the emergence of specific forms of game data work can unravel traditional industry hierarchies and thereby challenge our ideas of data work in creative industries.
Methodology and data sets
As Foster et al. (2018) point out, conditions of data work can be explored in different levels ranging from international regulation to everyday micro-level processes. To understand the experiences that game professionals have with data analytics – and to better understand how datafication is ‘made and unmade’ through everyday practices (Kennedy, 2018) – we wanted to highlight the subjective, lived experiences, and interpretations of our subjects, while maintaining the open-endedness of our data (Graebner et al., 2012: 278-9). In this sense, our primary focus is on the micro-level of the individual, while the connections to the meso-level (game studio as an emerging context for data work) and the macro-level (overall platformisation of game work) are discussed when it feels appropriate.
Our research uses a mixed methods approach with two data sets: analysis of game industry job advertisements and semi-structured interviews with 20 game workers in various positions, including but not limited to DDPs. The two main data sets both play off and complement each other, in that job advertisements – which are viewed as ‘company talk’ representing the ‘official’ viewpoints towards the studio's operations – are contrasted with the more personal views that game workers share in their interviews.
Job advertisement data
Our first data set consists of game job advertisements posted by Finnish game studios, collected between 1 January 2021 and 1 May 2021, from the Finnish website gamesjobs.fi. Companies on the site mainly hire workers for studios located in Finland and occasionally for studios based in other countries. During the research period, the site was checked for job listings once a week. Altogether, 184 new job advertisements were published on the website during the examination period.
In a preliminary round of analysis, we searched for the keyword ‘data’ across all the ads, and then manually went through all the ads looking for jobs with data-intensive tasks. We found 24 ads for data-centric or data-intensive jobs, all of which were included in a follow-up analysis. 1 Ten ads had the keyword ‘data’ in the job title, and all were evaluated to be data-centric jobs. These included ‘data & performance analyst’; ‘data analyst/data engineer’; ‘game data designer’; ‘senior data analyst’ (three instances); ‘senior data analyst/senior team manager’; and ‘senior data engineer’ (two instances). The remaining 14 analysed ads did not feature the word ‘data’ in their title, yet were manually classified as data-intensive jobs based on their descriptions that featured tasks involving data and/or analytics duties. The included jobs were ‘cloud engineer’, ‘CRM [customer relationships management] manager’, ‘game designer’, ‘lead/senior product manager for user acquisition’, ‘senior java/typescript fullstack developer’, ‘senior java server developer’, ‘senior performance marketing manager’ (three instances), ‘senior product manager’, ‘senior server developer’ (two instances), ‘senior user acquisition manager’, ‘UA/performance marketing team lead’, and ‘user acquisition specialist’. These data-intensive jobs illustrate well the expanding role of data within contemporary game development. Corroborating previous research (Van Roessel and Švelch, 2021), many such data-intensive jobs were for general management or mediator roles or for ‘creole professions’ that represent new occupations born in the junction lines between old roles (O’Donnell, 2014). After this first round of analysis, we conducted a second round of analysis where the ads were manually analysed for recurring themes and language reflecting data imaginaries (Beer, 2018).
While job advertisements are partially constructed to act as marketing narratives by the companies, thus being relatively one-sided data, they are an effective way to gauge what kind of skills are presently needed in a specific production sector (Dziobczenski and Person, 2017: 41). Our approach was inspired by a similar study aimed at in-game monetisation experts in Germany (Van Roessel and Švelch, 2021), which combined job ad descriptions, developer interviews, and the analysis of in-game credits to analyse issues such as how many job roles had monetisation duties integrated into them, with a specific focus on monetisation specialists. In contrast, our overall approach is more qualitative, and the analysis of the job ad data serves more of a contextualising and anecdotal role. The data corpus in this study is smaller and no statistical content analyses were run on the job ad data. Subsequently, our approach leans towards similar, more qualitative studies on job roles (e.g., Kerr and Kelleher, 2015) with a focus on the wider organisation of data work in game studios and its position in the production network.
Interview data
Our second data set consists of game worker interviews, conducted online via Zoom in spring 2021 with professionals from the Finnish game industry. The semi-structured interviews were conducted by two researchers and lasted between 28 and 75 min. As a method, interviews allow complex issues to be discussed in a nuanced manner and can encourage informants to greater openness (Alshenqeeti, 2014: 41). Interviews were deemed an efficient method in trying to assess latent or emerging relationships within a work culture, provided interviewers remain sufficiently reflexive. Interviews were particularly constructed to examine workers’ feelings and work practices related to data work, aiming to identify emerging themes. Altogether, 18 interviews were conducted over 5 months, with 20 professionals participating. One group interview included three people who preferred discussing data-related matters (e.g., the division of work) with their work team present.
The interviews were partly informed by a 2020 online survey for Finnish game developers. Initial interviewees were selected from survey respondents who left their contact details. Subsequent informants were selected using the researchers’ industry contacts and snowball sampling. To avoid common snowball sampling biases (see Parker et al., 2019), informants were handpicked from diverse backgrounds: workers both from large and small companies, with different lengths of working careers, a wide variety of work positions – both management and workforce – varying experience with data, as well as different genders and nationalities. The Finnish game industry has always been very transparent in sharing experiences related to development work, and almost all the workers we contacted agreed to an interview, some of them specifically under the promise of pseudonymisation. None of the interviewees reported having an NDA that would limit what they could discuss in the interviews. Among the interviewees were DDPs such as data analysts and user acquisition data workers such as performance marketers, but importantly, also a spectrum of professionals from other areas such as designers, artists, community managers, and company leadership. Like DDPs, performance marketers too are deeply involved with data, and as such should be considered data professionals; however, their work is less involved with the day to day of general staff. As we will argue in the analysis, all management roles need to understand and work with data. Table 1 presents a breakdown of the interviewees. Proportionally, the included spectrum of company sizes represents the Finnish game industry quite accurately.
Breakdown of the interviewed game workers.
This broader range of game workers enabled us to position data work within the wider context of game work. We had two goals. First, we wanted to be able to better understand to what degree people in positions other than DDPs conduct data work (see Van Roessel and Švelch, 2021). Second, we wanted to sidestep any distortions related to the centrality of data work that might have resulted from a homogenous sample of DDPs, such as data analysts. Nine interviewees (9) identified as male, nine (9) as female, and two as non-binary. Five interviews were conducted in English (some of them with non-native English speakers), with the rest in Finnish.
The interviews were transcribed using a professional service. 2 Thematic analysis of the interviews followed the framework outlined by Maguire and Delahunt (2017). The interview transcripts were read by the two interviewers who familiarised themselves with the data and assigned preliminary codes to the data using Atlas.ti software. The data were examined for patterns and themes that might emerge across the different interviews. The themes were then reviewed by three researchers, with the aim of identifying a cohesive arc across the data. This review included interpreting the data for both semantic and latent themes (Braun and Clarke, 2006), discussing the possible themes, discarding non-functional ones, and naming them. While no inter-coder reliability tests or similar formal procedures were conducted, coding practices were critically assessed in regular debriefings. We relied on the positional reflexivity of researchers and saw their different perspectives and backgrounds as a strength of the process (Anderson et al., 2016; Linneberg and Korsgaard, 2019).
In the analysis section, our findings are presented through three main sections. The first examines the job advertisement data, highlighting themes that illustrate what kind of workforce is currently needed and how this reflects industry priorities. The second section discusses the interview data with a focus on workers not specifically responsible for data matters. The final section focuses on DDPs but discusses them in conjunction with the overall workforce due to the close working relationship between them and other workers.
Analysis: Organisation of game data work
We first turn to explore how game companies present game data work, as seen through game job advertisements. This analysis looks to build our initial understanding of game data work. The game job advertisement data set provides us with the ‘official’ take on game data work – i.e., what we can see on the surface – which is then contrasted with our interview data, to identify what game data work looks like on the micro-level (Foster et al., 2018) or ‘on the ground’ (Carter and Sholler, 2016).
Game data workers wanted!
Data is at the heart of everything we do at Rovio. It enables us to continually improve our games and provide incredible experiences for the millions of users who play our games every day. All of this data is only valuable if it can be used to guide our day to day operations. (Rovio's job listing for ‘Senior data analyst’)
Some of the positions 3 were primarily game design-oriented, whereas others focused on marketing or more technical issues such as machine learning. Many of the job adverts mentioned specific cloud services, database management systems, software development kits, data analytics platforms, business intelligence software, data visualisation tools, game development engines and tools, programming languages, version control systems, application programming interfaces, marketing analytics services, advertising platforms and other online services, making visible the vast network of tools, partners, protocols and frameworks utilised in the current-day game data work.
Based on the job listings, the phase and maturity of the company often had an impact on the recruitment focus. Some companies were hiring professionals to build ‘a data pipeline’ from scratch, whereas others expected a familiarity with an ‘established stack’ of services. In addition to the skills related to developing and using elaborate data analytics systems, several listings mentioned the importance of being able to communicate, present, visualise and share insights from the collected and analysed data, corroborating one aspect of the ‘hybridity’ – technical prowess and communication skills – brought up by Dorschel (2021). As a reminder of the prevalent start-up culture, some of the listings also called for a ‘growth mindset’, ‘relentless culture of experimentation’, ‘ability to think big and outside the box’, and a desire to ‘always strive to improve and learn’.
Some of the newly emerging game industry positions can be seen as relatively invisible or low status compared to the traditional core development team of game designers, artists and programmers (Kerr and Kelleher, 2015; Tyni, 2020). However, the job listings do not portray game data professionals as marginal. Instead, the analysed positions were mentioned in terms of being ‘a key hire within the company’, ‘at the core of decision-making’, ‘in the lead role’, ‘a key player’, or as ‘a pivotal and influential position within the company’. While job listings obviously serve a PR purpose and need to make the described positions appear significant, it is still worth noting how often the centrality of these roles is repeated.
Given how some studies discuss the importance of a specific kind of education for data scientists (Dorschel, 2021), one interesting observation was that almost none of the job listings mentioned an educational qualification, and expressions such as ‘strong experience’, ‘demonstrated ability’, or ‘proven success’ were used instead. Often mentioned general skills included things like an ‘analytical mindset’ and ‘attention to detail’, combined with a ‘hands-on mentality’. While the detailed requirements were often platform and tool-specific, a genuine interest in games was also mentioned in many descriptions. What is interesting in these descriptions is that rather than stipulating qualifications, the companies were looking for employees who are ‘passionate’ about mobile games and data-driven development, and who ‘love’ F2P games and improving them based on data. One job advert summarised this mindset as follows: ‘We expect you to be a gamer at heart with ROI [Return on investment] on the top of your mind’. Previous research has already shown that this sort of neoliberal call for uncritical emotional commitment (Consalvo, 2008; Couldry and Littler, 2011) is often visible in game industry job listings (Kerr and Kelleher, 2015), and based on our data, we can confirm that this is relatively often the case also in the context of game data professionals.
Continuing the analysis, we now turn to the interview data. As a contrast to the ‘official’ stance seen in the job advertisements, the interviews reflect a more subjective standpoint of the workers, sometimes aligning with the advertisement data, and sometimes differing from it.
Non-dedicated game data workers in game studios
Previous research has called for more research examining data work in other job roles outside of data scientists (see, e.g., Dorschel, 2021; Foster et al., 2018) and how data are shared in data-centric sectors of the creative industries and to what effect (Rasmussen, 2024). We have previously argued (Sotamaa et al., 2024) that everybody working in contemporary F2P mobile game development needs to understand data and analytics, at least on a basic level. There are several game workers whose job is not to handle data per se, such as visual artists or sound designers. However, their job still involves interactions with some types of data, as all the designed dimensions of a networked mobile game can be measured against user behaviour data (Nieborg, 2016). Additionally, some game studio operations are purposefully built from the ground up to utilise data and analytics every step of the way, with the idea that all areas of development should be ‘datafied’ to make them better controllable. None of our interviewees mentioned a development role that specifically would not handle data. Instead, some work roles were mentioned as most likely needing relatively little data (such as visual artists), but with the caveat that these workers can also access analytics any time they want to, should they feel the need. Consequently, while being non-expert in terms of data and analytics, these workers nevertheless have a relationship with data and are expected to have basic or ‘normal’ data abilities (see also Charitsis and Lehtiniemi, 2023).
This everyday relationship with data is initiated and maintained in different ways. First, there is a clear focus on fostering an environment that emphasises easier communication through data. To facilitate the access to analytics for even the most uninitiated personnel, there are various dashboards and other kinds of tools that make data more accessible and manageable. Dashboards are software tools that present data in an understandable and concentrated form and can be used to make simple queries within that data. Different kinds of dashboards come as ready-made commercial products, which can be customised to better suit the studio's needs; however, they are often built from scratch by the company. Initially, this might happen based on tool needs identified by the staff, with DDPs perfecting them based on their expertise and identifying opportunities in the data pipeline. Indeed, for most of the non-expert staff, interacting with data means interacting with a dashboard. Dashboards extend parts of the know-how of in-house DDPs and professional outside services to the layman staff, as DDPs try to include in them information that would otherwise be asked directly of them. Accessibility of data was seen to be important, and dashboards were typically mentioned to be available for everybody whenever needed. Despite this, some staff members specifically mentioned wanting to have a direct conversation with the DDP when something was unclear. In many studios, DDPs are a constant presence now, needed at hand to advise the rest of the staff wherever and whenever.
Second, studios frequently have general studio meetings where different kinds of presentations and reports are given by the management, senior staff, and DDPs. These presentations make currently relevant data and analytics visible and understandable for the rest of the staff. Through such instances, everyday game data workers are gradually exposed to the latest phenomena in the areas of game data and analysis, while at the same time getting acquainted with the ‘data talk’ used by the DDPs (Sotamaa et al., 2024). Data talk is both a way of talking using data-related concepts, acronyms and other jargon, and a mindset where arguments and decisions are specifically based on data, instead of concepts such as design intuition or personal opinion. For example, when discussing their previous job as a community manager, a producer described how ‘there [in the community management side] too you can use data to justify your worth and the impact of your work, that benefit in economic terms and other ways too’. Data talk is a significant factor in building up a shared understanding about data – and by extension the ‘normalcy’ of a datafied worker (Charitsis and Lehtiniemi, 2023) – undisputedly guiding the design process. Interviewer: Does the studio offer you any kind of in-house training for working with data, or how are those skills usually learned? Game artist: Not any kind of an official training, but our data people have arranged the kind of meetings where we go through [or] learn particular terms: ‘these terms we’re using, this is what they mean, these are what we look at, these are what we’re interested in, these are our targets’. Like where everybody gets that kind of basic, one-o-one, one zero one, this kind of a basic understanding about what we are talking about if we talk about something like a KPI […]. The most important metrics, everybody knows what they mean. Our data people have given us presentations about that. Also, included in the studio on-boarding process, there's a short introduction about how we use, what we use, how to access it [the dashboards], how to navigate it, what are the most important things in there for that particular job title, and so on. It's part of the on-boarding process.
Alongside the uninitiated general staff, there is a section of game studio workers who need to understand data and analytics on an advanced level, even if it is still not at a level for them to be classified as DDPs. This is a group that one of our interviewees called the ‘product people’, meaning various members of the staff who are, or need to be, concerned with the ‘health’ of the game product. Consequently, these ‘product people’ are the primary users of dashboards (in comparison to the less data-focused staff who have fewer needs to follow the analytics software). Additionally, instead of looking at data and analytics that concern just their role, product people often need to combine different strands of data (e.g., from marketing, support, community and quality) and analyse them as a larger entity. Product people include typically the CEO and other management staff of the company, game producers, game designers, and people from marketing. A producer described how ‘all the producers, product owners are expected to understand data, and it's pretty proactively used by everyone who is working with product management – or any kind of management’, and later continued to describe the role of data for a producer: Producer: [D]ata for me, in my current role, it's wider, it's not about the task of a single person, it's about the impact of an entire team. It's concerned about how the entire company uses money, in a way giving a response to whether it [using money that way] is sensible. And, of course, there are also measurement instruments to see how specific workers are doing. It works in many ways, for everything ranging from pepping up a single worker like ‘hey, your work impact raised the graphs this much’, or ‘hey, this game is bringing us this much profit’, or just something like ‘hey, this feature which we’ve been working on together, the players are getting this much out of it’, or, technologically speaking for example, ‘our quality is this bad’.
The work of DDPs
When a fledgling game studio's data operations reach a critical threshold, they hire a dedicated professional for data and analytics. This typically occurs after stabilising revenue and achieving profitability. Recently, due to the growing importance of data-driven business models in the industry, experienced company leaders may hire a DDP during initial business planning, especially if securing investments before a soft launch. Investors now expect credible business plans grounded in data and analytics.
In game studios, DDPs play a crucial role. Even relatively small companies might have several DDPs on staff. Conversely, mid-sized companies may rely on just one data professional who handles advanced data matters for the entire company. These dedicated data roles encompass various titles, including ‘data analyst,’ ‘data engineer,’ ‘data scientist’, and ‘data visualist’. The specific responsibilities associated with these titles can vary significantly: a ‘data analyst’ in one studio might perform tasks that, in another company, would be handled by a ‘data engineer.’ In mid- to large-level companies, data tasks are typically managed by a data team consisting of various DDPs. The hierarchy among these DDPs varies, but in one interviewed studio, a data scientist with a high level of expertise oversaw the process, drawing multiple strands of data together to create custom solutions, whereas in another studio a data analyst was the only DDP on staff. Data engineers often focus on maintaining and servicing databases and data architectures.
Perhaps the most important task of a DDP is to transfer ‘raw’ data into ‘actionable insights’ [senior data analyst], and this way make it something that can be understood, communicated, moved around and acted on by the rest of the company. The most basic level of data analysis is achieved using ready-made, third-party services, typically by the largest platform companies, Google, Apple and Meta. As companies advance in their use of data and analytics, they often transition from third-party solutions to developing their own custom-built analysis tools and dashboards. These bespoke tools allow for more precise content targeting and player segmentation, ultimately enhancing player services. A data scientist described how, when joining the company, the studio used an externally provided tool, but they gradually shifted toward custom-built solutions. To achieve highly precise custom analytics for improving content targeting and player experience, existing off-the-shelf solutions fell short. For instance, Unity Analytics, while useful, didn’t provide the level of precision or data control needed. ‘So, it was this kind of an impasse; it was necessary to get a solution that simply wasn’t available’, he continued. Whitson (2019: 793) highlights that analytics tools promise efficiency but require technical and business proficiency, often necessitating postsecondary degrees in computer engineering and data science. Most companies now combine DDPs with third-party analysis software purchased externally. Some studios had to rely entirely on a third-party software (e.g., for budget reasons), leaving advanced queries unanswered.
To adequately connect multiple strands of data, the highest ranking DDPs in the company need to understand the big picture. The DDPs we interviewed were all highly motivated, constantly educating themselves about new techniques and phenomena in the field (see Avnoon, 2021) through trying to learn as much as possible about game development. Describing the evolution of their work role – initially outside of gaming – a senior data analyst told how in his first job he was just building the technical framework to move data from one point to another, applying some kind of transformation to it and making it accessible to non-data professional staff at the company. In his first game industry job he needed to learn the entire business area of games ‘with really business-driven people’ to be able to provide insights for marketing, media operations, and more. Finally, in his second game industry job he was charged, alone, with building most of the tools and the data environment from the ground up, necessitating his involvement in building the actual game from the beginning, his tools ‘shaping the product’. This, in turn, necessitated understanding what makes a good game. Because all the areas of game making are datafied (and therefore can be bettered through data), the most advanced data professionals need to know how the different elements of the game service work; ‘live and breathe’ the game [data scientist]. If a data professional is a part of shaping the product, they ‘need to understand the business’ and the product, the KPIs and their importance – in other words, how data are leveraged in the best possible ways to help the product. Consequently, because live mobile games are so datafied, data-related tasks like working on the data architecture are fast becoming one of the core development areas.
These observations align with previous research: Avnoon (2021) points out that whereas 20th-century engineering professionals based their identity on highly specified expertise, data scientists base their identity on generalism. The data scientists interviewed by Avnoon felt that in addition to their data skills, they could acquire vast amounts of domain-specific knowledge on the field they were working on, e.g., medicine, finance or law – in this way resembling some of the DDPs interviewed by us. In this eagerness to master the domain of game business, an ethos could be detected that game studio DDPs were curious about problems in the field and invested in making a difference to the financial outcome through their work. All in all, observations in this section indicate that data analysis is now tightly entangled with traditional game development skills, both through DDPs moving towards game developers (i.e., learning those skills) and through the rest of the staff moving towards data analysis via the more and more accessible data.
The central role of communication
While the technical solutions built by DDPs, too, target a more accessible and communicative work environment, it is worthwhile to examine DDPs’ communication tasks more closely.
In a game studio environment, matters related to data are communicated in several ways. Outside of the dashboards, DDPs answer questions from the team members whenever needed, often explaining and educating staff members about the logic, possibilities and costs of different kinds of data queries. For example, game designers mostly know what will benefit the game in terms of design but might lack the expertise or literacy to think of more advanced questions (beyond the usual event-based data-driven design) to ask from the game data. Several of the interviewed management also brought up the fact that there is too much data to analyse, and that DDPs are needed to point out what is most important in this regard. Again, this emerging lock-step relationship between DDPs and the core development roles such as design demonstrates the rising centrality of data in game development.
Direct face-to-face communication between two people (e.g., a DDP and a game designer) was often considered to be the most efficient way of communication, and such conversations were reported to happen whenever needed. The direction of this kind of direct communication is usually from a perplexed game developer to the DDP, with the DDP acting as a ‘support’ service in these situations. Some of the interviews brought up how DDPs were previously located elsewhere from the rest of the ‘core’ game development team, or they were employed as a freelance workforce that communicated with the studio completely remotely. More recently though, DDPs have been working in the same room with the team. This is aimed at speeding up communication, as any problems occurring in games that are ‘live’ need to be reacted to as quickly as possible. More generally, this further highlights the frequent need for data expertise within the day-to-day operations of contemporary mobile game development.
In addition to direct conversation, data matters are also communicated in other ways. In addition to dashboards, there are also several kinds of general instructions, such as PDFs on a company server, that are accessible by all. Internal messaging platforms such as Slack are typically used by employees to ask quick, simple questions from the rest of the team or, for DDPs, to share more general data-related information with the rest of the team. One of the more important ways to communicate data matters is through different kinds of visualisations, usually created by DDPs or product people, as well as other kinds of presentations. These visualisations focus on specific data relationships, often highlighting unexpected connections between design choices and user behaviour. By transforming complex data into more understandable formats, these visualisations serve as a key communication tool between DDPs and the rest of the studio staff, underscoring the crucial role that effective communication plays in the work of DDPs.
Previous research has highlighted how data professionals emphasise the importance of different areas in their work based on the domain they work in (Carter and Sholler, 2016). Data analysis skills take precedence in domains where the accuracy of analysis is highly important, such as in medicine and finance. On the other hand, in organisations where data professionals need to work in close proximity with non-technical staff, communication skills and the ability to translate technical information into actionable insights become central (Carter and Sholler, 2016). In our study data, game studio DDPs sit somewhere in between the two areas, a ‘hybrid’ position highlighted by Dorschel (2021).
In general, the studios we interviewed seemed to operate under a transparency where all the workers had equal access to the studio data through the various dashboards available. At the same time, as the most highly trained professionals, DDPs are the central brokers of data in game studios. They often build the solutions through which data are communicated, and, although not concealing data in any way, DDPs are the ones to tell others what is relevant about the data. While this highlighted the DDPs’ need to sufficiently communicate everything deemed as important, it also highlights how data are disseminated from the high-level professionals to other members of the workforce, and how data change the organisation and hierarchy of communication within studios.
Discussion and conclusions
In the beginning of this paper, we set two goals for the study. We wanted to both empirically explore how game data work is organised and communicated in game companies, and to better understand how it potentially challenges prior theorisations about the role of data analytics in creative industries.
Based on our findings, game data work is similar in many ways to data work in other fields, but also exhibits marked differences. For example, Kelly's (2019) research on the television industry suggests that increasing the usage of real-time analytics encourages the production of particular genres and puts more focus on live programming. As we have shown, ‘live ops’ – being able to configure games in real time – is crucial for current-day game production as well. At the same time, there is little evidence to suggest that only some game genres benefit from the work of data analysts. Instead, and similar to Paul (2024), we argue that the optimal usage and impact of analytics can vary significantly from genre and monetisation model to another.
Screen workers’ interactions with data studied by Rasmussen (2024) include very similar practices to those explored in our study. Workers in creative industries increasingly engage in analysing data from service providers, collecting their own data, and occasionally resisting extensive data use in their work. There are differences though. If the screen workers in Rasmussen's study routinely discussed the lack of access to the global platform holder data, game data workers in our study rather complained about the overflow of data and the lack of time to properly analyse it. In this respect, we did not witness the data divide that is argued to characterise the television industry (Kelly, 2019). In contrast to findings from other creative industries, Finnish game companies strive to share data insights widely among their workforce, promoting a culture of transparency and collaboration. This challenges traditional views that data analytics is a top-down process. For the game studios, data analytics is not anymore solely the domain of specialised roles; instead, data engagement is a collective responsibility across the work organisation.
As highlighted both by the game industry job listings and our interview data, game data work is organised around a vast network of specified tools, services, platforms and partners. Competencies around data analytics appear to be increasingly central to the operations and expected success of the companies. While job advertisements can consciously exaggerate the importance of particular roles, our interview data support the idea that DDPs are crucial staff members for their companies, at the very least in larger companies. At the same time, DDPs sometimes perceive their work as ‘support services’, outside of the core development tasks, highlighting an ongoing negotiation of data work's value within the broader context of game production work. Despite their high technical prowess, it is the communication skills of DDPs that have risen to a very high value. As Netta Avnoon (2021: 342–43) points out, this combination seems to be crucial for understanding the specificity of DDPs: The imperative to combine both ‘hard’ mathematical skills and ‘soft’ social skills is what distinguished data scientists from both the ‘old-school’ technical snobs and non-technical occupations. In their identity work, data scientists maintained the omnivorous symbolic boundary between themselves, who have a mathematical-social skill set, and single-skill occupations such as statisticians and algorithmists, the ‘geeks’ with only mathematical skills, or marketing and sales personnel, who have only social skills.
The results also draw attention to the broadening scope of data work and how data work and creative work seem to be gradually converging. Data work is not anymore confined to data-specific roles, but is integrated into the daily tasks of the development staff and management. All workers are trained to understand data and analytics on a basic level through onboarding and regular data briefings, with evidence pointing to ‘normal’ data literacy (Charitsis and Lehtiniemi, 2023) having become an assumed skill in the data-driven game industries. If Dorschel (2021) describes data scientists as hybrid work subjects who need to combine several skill sets and character traits, we see similar developments in connection with the data workers of the regular staff. For workers, cultivating data skills to complement creative expertise is a way of keeping one's skill set attractive and relevant, as such hybridity is seen as a strategic advantage in a digital capitalist system that values adaptability and optimisation (ibid.).
Everyday issues of power in the workplace can be approached through analysing work hierarchies, specialisation and division of labour. In this respect, the results of our study are twofold. With an increased focus on data analytics, the game industry appears to be more diverse in terms of job descriptions and professional backgrounds. With the advent of roles like game testers and community managers, the industry has already moved beyond the sole focus on a core development team consisting of game designers, programmers, and artists. The new aspect that DDPs bring to the equation is that they seem to have more power and authority than people in non-development roles have previously had. At the same time, most people in management roles (around half of the informants) also worked intensively with data, highlighting how the company leadership does not necessarily change due to the introduction of data, if the people serving in management positions are able and willing to update their approaches and skillsets.
Similarly, it is important to question who provides the overview of the data-focused processes, who controls the access to different data forms, and how power is distributed between different data professionals. As Chia (2022) has importantly highlighted, data-intensive processes and forms of automation have already started to create an underclass of creative game workers whose everyday work is centred around habilitating algorithmic systems. One interesting observation is that DDPs are rarely gamers, and their central role is possibly starting to unravel the narratives of passion-driven labour for the game industry. The idea of ‘passion work’ cherished in creative industries has often led to bad working conditions and structural overwork. While datafication of game production can further intensify development processes and reduce creative autonomy, the influx of data workers from other fields can also potentially help in seeing game development and associated activities more as ‘normal’ or ‘real’ work, this way slowly improving the work culture.
Altogether, the gradual convergence of data work and creative work warrants more studies, perhaps even re-evaluation, on how we understand the integration of data analytics within creative industries. While making a division between macro-, meso-, and micro-levels of data work can be useful on a conceptual level (Foster et al., 2018), an empirical project like ours shows how understanding the nuances of everyday data work necessitates mapping the connections between these levels. We have also shown how it is increasingly important to pay attention to the more invisible data workers along with the more visible data professionals already studied by the likes of Dorschel (2021) and Avnoon (2021). We encourage future research to embrace more sophisticated, context-sensitive theorising towards data work that considers in a nuanced manner the spread of data tasks to other job roles within creative industries.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is supported by the Strategic Research Council [grant number 352523] and Research Council of Finland [grant number 353266].
