Interrogating Creative Theory and Creative Work: Inside the Games Studio

Introduction: Claims and Rationale

The traditional literature on the management of creative work has tended to emphasise the inherent tension between autonomy and commercial pressures (see Thompson et al., 2007). This correct but somewhat static observation has been superseded by the expansion of creative and cultural industries that has provided a rich source for theoretical claims and commentary. Recent years have seen a proliferation of conceptual terminology – the creative class, the cognitariat, immaterial labour, the cybertariat. Such arguments draw on radical (post-Fordist, post-modern and autonomist Marxist) and mainstream literatures that reproduce and extend the idea that ‘[i]n place of an account of the power of capital, it stresses the autonomy and creativity of labour’ (Gill and Pratt, 2008: 5). What’s more, it is presented not only as something happening on one territory, but its wider significance ‘for reflecting on post-industrial forms of work to which future sociological analysis will, in all probability, be increasingly directed’ (O’Doherty and Willmott, 2009: 945).

In this article we argue that there has been an excess of high theory over detailed empirical accounts of actual creative labour in specific creative industries. In interrogating the above claims we seek to add to the relatively small but growing body of work that is grounded in close observation of industry contexts and dynamics (e.g. Christopherson, 2008; Deuze et al., 2007; Hesmondhalgh and Baker, 2008; Hodgson and Briand, 2013; Legault, 2013; Martin and Deuze, 2009; Teipen, 2008; Umney and Kretsos, 2014). After outlining and commenting on the new literatures, we locate the digital games industries within the value chain context before drawing on evidence from interviews with employers and workers in games studios in Australia to examine key work and employment trends, offering a significant corrective to claims concerning creativity, autonomy and attachment.

Literatures, Theories and Claims

In this journal O’Doherty and Willmott (2009: 939–943) present the clearest exposition of recent theoretical claims. This can be summarised as:

The picture in this and other accounts is not wholly positive. Networks and new ‘non-places’ can also be disciplinary spaces and sites for insecurity, long hours and low pay. However, the overall judgement is clear – the outcomes are the result of self-discipline and ‘networked individuals’, many of whom are highly motivated self-entrepreneurs, are not ‘victims of capital or managerial oppression’ (O’Doherty and Willmott, 2009: 941). The explicit target for such observations is ‘neo-orthodox labour process analysis’, with its supposed preoccupation with the wage–effort bargain and capital–labour conflict.

In contrast, O’Doherty and Willmott draw primarily from highly influential autonomist Marxist writings on creative labour (see Gill and Pratt, 2008 for a summary). Capital is seen as increasingly relying on mobilising the subjectivity of labour, with a particular emphasis on the role of information, communication and emotions in the labour process and the production of commodities: ‘the co-operation and subjectivity of labour have found a point of contact outside the machinations of capital’ and function as ‘radically autonomous processes of self-valorisation’ (Hardt and Negri, 1994: 282).

While immaterial labour can be found everywhere, the games industry occupies a special place in such theorisation. In one of the most explicit applications of autonomist thinking, De Peuter and Dyer-Witheford (2005) assert that digital games are an exemplary site for the mobilisation of immaterial labour. Though they recognise wider industry constraints on games’ firms, their view on work autonomy is that: ‘In certain games studios the “management” of collaboration is imminent rather than externally imposed on game labourers’ (2005: 4). They quote one engineer: ‘There’s nobody telling you how to do something. There’s no paperwork getting in your way. There are no set rules that you have to follow – rules that you don’t feel are necessary’ (2005: 2). The outcome is a ‘communism of capital’ in which work is like play, characterised by self-expression and self-transgression: ‘You come in, you see your friends, you get to make video games, and you get to play some. It’s pretty cool. It’s really not even so much like work here’ (2005: 5) To add to its ‘exemplary’ status, the games industry receives major inputs from ‘modders’ – players and enthusiasts who apply their social capital to the modification and development of games with the active encouragement of many of the companies (Postigo, 2007).

However, for autonomists, immaterial labour has a dual character. This is focused on the precarious nature of such labour, giving a radical twist to mainstream notions of mobile free agents; recognition is given to a diverse range of temporary, intermittent and insecure jobs with weak work–life boundaries. This label is not difficult to fit to cultural work, with its high mobility; frequently poor pay; reliance on informal networks to find work and the mixture of anxiety and attachment to creative endeavours and aesthetics. De Peuter and Dyer-Witheford (2005) refer to a dark side of passionate but precarious ‘play slaves’ in the digital games industry, subject to excessive hours and corporate rationalisation. Gill and Pratt (2008: 3) draw attention to the double meaning of precarity – a new moment of capitalism and space for new subjectivities; and the experience of unstable, insecure and contingent work. What stabilises this tension is the high level of aesthetic identification and attachment that creative employees have to their work, industry and occupational community.

Despite the radical language, what is significant are the overlaps and continuities with mainstream literatures with respect to creative industries and labour. Florida’s (2002) The Rise of the Creative Class has been the seminal influence. He argues that in the increasingly fierce war for talent, employers have to provide creative spaces with little or no regulation to attract such entrepreneurial, mobile free agents. Referring positively to Florida, Shorthose (2004: 4–5) talks of an ‘expanded realm of autonomy’ for artistic labour and alternative self-organised cultural forms from below. The centrality of project work is also a consistent theme of mainstream accounts. Grabher’s (2002) influential account of the project ecology of the advertising industry in London highlights the ‘decomposed’ production process – structured around recurrent, episodic projects that are difficult to pre-specify either with respect to content or volume. Similar arguments are made about freelance entrepreneurs moving across projects in different locations in the animation industry (Yoon and Malecki, 2010), and the reliance on output controls rather than standardisation and planning in games because ‘no one project resembles another’ (Teipen, 2008, 322). Finally, mainstream accounts tend to confirm high levels of attachment and identification with work (DeFillippi et al., 2007).

Data Collection and Analysis

Our main source of data was 33 semi-structured interviews conducted with firm managers and employees in a variety of firms of different sizes and participating in the mobile and/or console industry segments in the Australian games development industry. Interviews occurred in two stages. The first round of interviews was primarily to identify the nature of industry dynamics and the strategies firms used to attract work. At this stage we interviewed either the firms’ founder or current owner/manager in 13 firms as well as leading public policy-makers in relevant government departments. The selection of interviewees was driven by theoretical considerations rather than statistical sampling; firms were selected to ensure we had access to different market segments (console (four), mobile (three)) or both (six) and which varied in terms of age of the firm (1–16 years), number of employees (1–100+) and ownership (domestic (11) or international (two)).

In order to capture fully the labour dynamics in the industry, we conducted a further 20 semi-structured interviews with employees/games developers. Employees were selected from a sub-sample of six firms. The developers’ experience working in the industry ranged from one to 20 years, with an average of seven years. Three of the more experienced developers had experience working internationally in the games industry. When referring to these respondents later in the text, we describe them in terms of the firm (1–6) and number of employee (A–E) (for example 1A or 6D).

Interviews with firm owners covered the range of issues necessary to identify key dynamics of the industry associated with financing, design, production and distribution and how these dynamics were affected by the structure of the global industry and international market developments. The interviews with employees/developers working in the industry were also wide ranging and focused on the nature of work and employment within the industry as well as the nature of the labour process and how these were affected by the nature of the industry context as described in the first round of interviews. Once we achieved a high convergence of responses we ceased interviews (Corbin and Strauss, 1990).

Interviews lasted for between one and one-and-a-half hours and were recorded and fully transcribed. The first step of the analysis was to conduct open coding of the interviews. Coding was conducted on segments of the transcripts that formed identifiable ideas, with each segment often coded into multiple categories. The categories and themes identified in the first round of interviews focused on the industry context and were theoretically informed by concepts concerning the structure of the industry, characteristics of transactions, including their complexity and codifiability, and power relations along the industry value chain. The categories and themes identified in the second round of interviews were theoretically informed by labour studies concepts such as autonomy, tenure, security, rewards, job fragmentation, skills, work hours and intensity.

The Political Economy of Digital Games and Entertainment Industries

We have argued for the need to examine creative labour in particular industry contexts. The games industry is the largest entertainment industry globally generating more revenue than film and music with successful console games titles individually producing over US$1 billion in revenue. The video or console game market revenue was US$18.58 billion in 2010 (Fieldman, 2011). There is a complex competition for value capture between console manufacturers, publishers, development studios and retailers (Johns, 2006); and research is beginning to uncover something of the changes in labour markets and processes associated with the evolving industry architecture (Deuze et al., 2007). Publishers, sometimes overlapping in ownership with console manufacturers are the fulcrum of power in the value chain, arising from their control of intellectual property and financing of increasingly costly games development. Independent development studios, such as those that characterise the Australian industry, rely on a business model which involves the development of small licensed products for publishers, particularly in the children’s or sports games markets (Colwill, 2010). This is because most Australian studios do not have the resources required to develop games, which take more than 12 months and involve very large production budgets, without publisher financing (Martin and Deuze, 2009: 283). As such, Australian firms are dependent on publishers and licence holders, mostly located in the USA, to finance their operations. In recent years, a large number of development studios in the Australian industry have either been acquired by US publishers or have been closed (Launay, 2012).

As such, independent development studios – the core of the creative labour process – occupy a marginal position in the global production network, leading to ‘strong pressure on costs and flexibility’ (Teipen, 2009: 18). The structure of the global industry is such that developers need to win work from publishers to ensure survival resulting in power asymmetry between publishers and developers. Firms are constantly pitching for new work and trying to build reputation and trust with publishers. A key issue for this article is the ways in which these power asymmetries and market pressures around cost and flexibility are reflected and reproduced in capital–labour relations within firms.

A further characteristic of the industry with significant implications for labour relates to the high level of uncertainty regarding how the potential product will evolve during the course of development. This puts a premium on budgeting and time scheduling. Firms frequently find themselves in financial difficulty because intensity of competition leads them to under-budget and commit to unrealistic scheduling in order to win projects. Clients have a reputation for changing their specifications and demanding more along the way. A number of respondents argued that ability to manage the subsequent contract variation carefully is a key to their success. That variation is, in part, premised on a double transfer of risk – from client to studio and from studio to workers.

However, a key shift in industry dynamics and business models has recently occurred. A new stage of digitalisation via devices such as the smartphone has created additional access to consumer markets and new opportunities for small developers in Australia and other countries. These devices require new types of games that are less complex and more suitable for casual gamers. Most importantly, developers can now upload their games onto retail sites such as the Apple App store, without the involvement of a publisher. This reduces barriers to entry associated with cost and reputation and therefore the dependence of developers on publishers. However, developers in the smartphone market assume all the risk of development. There is massive competition in the development market for mobile devices (Fieldman, 2011), with an estimated 1000 apps uploaded to the Apple AppStore daily. The heavy concentration in the retail and smartphone devices markets has ensured that developers bear the costs of the development of the vast bulk of games that never succeed on the online market. In the next section we examine how changing conditions in traditional and new segments of the market are impacting on creative labour.

Digital Games: Labour in Context

When respondents were asked to reflect on changes in the industry, two main themes emerged. First, that firms and workers had ‘grown up’. In other words they had moved from the garage to becoming larger, more structured, more managed, more professional outfits, with career paths for core employees. This is also true of other markets such as the UK (NESTA, 2011):

The industry when it started was very ad hoc … we went through a maturing process where the projects were getting bigger and bigger. So something had to change … more structure, systems and processes. (5A)

I think I’ve seen a much more professional approach in a lot of ways taken to actually running it as a business and scheduling properly, managing people properly. (1A)

Almost all of the larger firms in Australia had grown and developed some stability through licensed contract work for domestic and international publishers. Second, the global financial crash (GFC) had, however, led to a severe shake-out in the industry. The consequences of this constituted the second theme. Respondents talked of tighter budgets, shrunken workforces and disappearing firms. A form of ‘squeezed middle’ seemed to have taken place whereby more prominent firms sheltered by continuing relations with successful, mostly US publishers surviving in downsized forms along with small studios that were nimble enough to adapt to new, notably online, markets.

Jobs, Tenure, Security, Rewards

Respondents confirmed what is known generally from the creative industries literature – that employment was accessed through networking and by word of mouth. On the key issue of (in)security that emerged from the literature, a complex picture emerged. On the one hand there had been a medium-term trend towards greater stability. Contract work for publishers has had sufficient continuity to mean that recurrent projects encourage retention of a core of expertise with a predominance of salaried positions, supplemented by temporary workers, ‘Everyone changed over to full-time in 2004–2005’ (3D).

What looked at the time like stable expansion also allowed employees to rise through the ranks of the larger firms: ‘there wasn’t much of a hierarchy and I really like hierarchy. I want to climb, I want to achieve’ (2B). Internal career paths arose not just from stability of work, but differentiation of work role. Artistic categories such as animators as well as technical grades such as programmers all developed associate, senior and lead positions. Greater complexity and coordination requirements also led to an expansion of managerial roles, with a number of respondents reporting pressure to move into them, even though there was little evidence of any training provision.

These trends are not inconsistent with a degree of job mobility across firms. Res-pondents reported the attraction of new projects and the limitations of existing work roles facilitated an ability to move either onwards and sometimes upwards:

You’ve nowhere else to go unless you take on management responsibilities. (1B)

The other thing is that you get the same types of project in a company over and over again, so once you get sick of working on the same thing it’s probably a good idea to go somewhere else. (2A)

Movement across firms in Australia and elsewhere was particularly the case for younger, more socially mobile workers, but most were not freelancers.

However, all respondents recognised that more recent trends following the post-GFC shake-out had increased insecurity. Referring to firms that have gone out of business or shrunk, a lead designer commented: ‘That style of industry does not work anymore. All that work is going to cheaper places like China and India … Australia didn’t even do a great job of it anyway’ (4A).

Those who had set up smaller firms in what respondents called the ‘indie space’ also understood that whatever the other attractions, work life is more precarious:

We do contractors at the moment. We never used to, but it’s only the way the industry is going that we’re forced to really. Because contractors – without it sounding really awful – are easier to get rid of if they don’t work out and we’re not stuck with them at the end of the project. (6A)

Division of Labour, Specialisation and Skills

Trends towards a fragmented and detailed division of labour were broadly confirmed in our interviews. Though there was variation by project, work had become more ‘production line’ and repetitious, with employees often ‘pigeon-holed’ in specialised tasks. This was particularly true for those in junior positions on more complex games:

I opened the same door like five million times and I heard the same sound effects five thousand times. Play ViroShot every single day for the rest of this year and I bet on day 365 it won’t be the most brilliant game you’ve ever played. (1D)

However, specialisation was also true for most working in larger firms and on licensed games:

There’s a common misconception about people trying to be more well-rounded. In smaller companies and indies, yes. Artists can be an animator or a texturer or a character maker or an environment artist. But that never happens in consoles. You’ve got a guy who creates heads and you’ve got a guy who creates bodies and you got a guy who makes environments and you’ve got a guy who animates them. (2B)

A number of respondents commented that work was now more technical, more driven by the hardware and therefore less creative. In the larger firms there was less need for generalists. However, those respondents who had retreated to the indie space making small games for the smartphone market understandably reported the reverse. The reduced number of employees had to became generalists again and ‘[i]t’s going 360, we’re back to the old ad hoc days where creativity and innovation are more important than process and system’ (5A).

Hierarchy and Autonomy

Unsurprisingly, a more detailed division of labour is matched by a more tightly managed process. The pipeline or workflow is increasingly structured according to ‘milestones’ and organised through what one employee described as a ‘strictly hierarchical structure’. An animator referring to the start of a project commented that, ‘We’ve got a list so I just follow the list and do it – check, check, check’ (3B). Such an interpretation is confirmed by a lead programmer from another company, ‘Now there’s tools that monitor these things, software, you put a task in and then they can break a task up and mark it as completed’ (2C).

Each part of the project – technical/programming and creative/animation – has its own mini-hierarchy with a director at the apex. For those in the lower levels, it is particularly difficult to exercise influence over the creative process. Referring to a period when he was a coder, an employee commented that: ‘Sometime you see that stuff coming through but usually there’s just too many steps and too much set in stone already’ (6A). Above that, figures such as an overall production manager can veto outputs. Respondents spoke of a ‘very managed’ process, with increased monitoring from within and outside the project.

The latter is important because constraints on autonomy arise primarily from the industry context, notably the hierarchic powers of the external agent, the licence holder or publisher. There was frequent reference to ‘the chain’ of creative influence. Such powers and interactions concerning contract variation set the framework for external and internal management and creative tensions. As with other spheres of practice, it is the larger projects with bigger budget pressures that are more managed. In contrast, in the new small studios, it is ‘back to the garage’ and looser creative processes.

Working Hours, Work Intensity, Work and Play

The industry is notorious for its long hours, particularly at ‘crunch time’ near the end of projects. Respondents often had ‘horror stories’ to tell, including references to particularly stressful ‘death marches’ of round the clock working. But as with other issues the medium-term trend has been towards a regularisation of core hours, generally seven-and-a-half to nine with a small amount of flexibility for later starts and finishes. Comparing unfavourably with past periods where people were ‘working every hour under the sun and not getting anything out of it’, an employee said that: ‘It was one of the major selling points for me to come to this company that you get time in lieu’ (2C). Again the exception is provided by the more recent emergence of an indie space where small studios are trying to get a foothold in new casual gaming markets. Here respondents admit that ‘abuses’ with respect to time and rewards are common (4A/5A).

A number of respondents reported faster work pace and pressure to ‘do more with less’ (1B) and that ‘everyone is working more’ (3B). Enhanced work intensity is thus added to the traditional issue of ‘crunch time’ and excessive hours in the development cycle. However, this was not a universal view. Some interviewees said that there had always been comparable pressures. Others observed that higher expectations, better planning and equipment had resulted in enhanced competence and less crunch time. There was more consensus that the strongest pressure derived from larger projects where the costs had been pitched at too low a level due to the intense competition for value capture.

With respect to another dimension of time, there was remarkable uniformity across the interviewees. All but one was dismissive or contemptuous of the work as play/play at work notion: ‘That gets beaten out of you in the first week’ (1B). Some claimed that it was never true, a project manager adding that, ‘That’s the biggest rubbish, and I’ve heard that a few times. I’m sure people in chocolate factories don’t necessarily consume the product. It’s not Willy Wonka land’ (1C). Again highlighting the industry context, others said that larger firms and enhanced pressures on project completion ruled it out. The outcome was that the job was hard and they often felt ‘worn out’:

Like people always say, oh you work in the game industry – are you playing games all the time? I was like, oh man, you know how tired you are if you want to try and test something and keep playing? (3B)

Some even reported that they had become so bored with the repetitive aspects of the work that they had stopped playing games at home: ‘I don’t even play games now because I work with them every day’ (1E).

As for modders and fan programmers, no-one reported any involvement in their companies or projects. A number did recognise the role hobbyists played in larger firms and triple A games. In those circumstances they are used in the development process to test products and provide add-ons that extend the life of a game. Their value creation is expropriated via the publishers’ intellectual property (IP). However, as some of our respondents supported, the evidence is that fan programmers are often using modding as a means of entry into the industry.

Work Attachment and Identity

We certainly found considerable evidence of affective attachment, but it was neither undifferentiated, unqualified nor unconnected to the effort bargain. The issue of continuing attraction (or otherwise) to the industry was explicitly explored with respondents. The majority had positive perceptions – ‘I love my job’ (3A) ‘passion drives the industry’ (1B), ‘money is not important’ (3D). It was made clear that the attachment was not to the company but to the industry and to particular projects. For example a strong driver for some was the opportunity to create original IP or work on triple A games. Respondents recognised that passionate attachments to the work could lead to toleration of long hours and insecurity.

However, toleration was relative. Employees showed awareness of the changing terms underlying the effort bargain:

I think you start in a lovey dovey kind of way. I’d say that the grim reality comes in because it’s a business. So, there’s a lot less of staying up late working really, really hard just because you love it. It’s because you love it and you want to keep your job and also because you’re improving your portfolio or you’re improving your skill set or you’re doing something for yourself … I’m feeling less motivated because I’m getting paid less; people will adjust their performance depending on the price … So, I think once you rationalise how much you’re worth an hour, you start pulling back from just giving your all. (2B)

Even those who evinced passionate attachment understood that the conditions sustaining it were deteriorating, with clear consequences for perceptions of the effort bargain:

It’s less attractive. People are more inclined to be, well what are the benefits? How much super(annuation)? What time do I leave work? So that’s changed with these monstrously big projects. (5A)

I’d say weakened. It’s not as glamorous as I thought it would be. There’s a lot more boring stuff … There’s a lot of times when you are sitting at work and you don’t want to be there. Let’s face it – your eyes are sore, you’re tired, you want to go home and no amount of creative freedom is going to fix that. But I still love it. (6B)

A variety of issues were raised about declining attractiveness by respondents. These included work often becoming more specialised and even boring; that original IP was harder to secure given power relations in the restructured industry; continued poor people management and increased bureaucracy. Those who had been compelled or chosen to retreat into the ‘indie space’ had recovered some of their enthusiasm.

Discussion and Conclusions

The creative industries have been presented and promoted as exemplifying new trends in the post-industrial landscape, with changing boundaries of capital, management and labour; work and play; effort and attachment. A key contribution of the article is not merely to counter excessive theoretical claims, but to ground a more realistic and complex understanding of creative labour within the political economy of particular creative industries.

What we have tried to explain and illustrate is that the struggle to win work from buyers places considerable constraints on the exercise of creative capacities at work. This is a complex and uneven process. Creative workers are a undoubtedly a source of value in a labour-intensive industry and the availability and location of skilled labour at an appropriate cost is a key factor in locational or work allocation decisions (Izushi and Aoyama, 2006; Teipen, 2008). This may help to explain the uneven trends we have observed towards greater security and full-time contracts, at least for core workers. Wider evidence on the industry in the UK indicates that freelancers may constitute only 12.5 per cent of the ‘digital and entertainment media’ workforce, which is marginally lower than the figure for the whole UK economy (Skillset, 2011). Our interviews also indicate that as the workforce as well as the industry has matured, this greater stability is what many employees value. By contrast, the necessary association between project work, precarity and ‘freelance entrepreneurs’ (Yoon and Malecki, 2010: 241) has not been demonstrated with any consistent evidence.

What of issues of division of labour and hierarchy in the labour process? The uneven power struggle between capitals in the value chain and the subsequent cost pressures does appear to have led to tighter supervisory procedures and ‘multiple milestones’ (Deuze at al., 2007: 350). Workers are highly aware of and often chafing against the limits of creative autonomy, both with respect to their own work and the struggle to create and retain original IP. Yet it is telling that in our interviews, the most enthusiastic about their strategic and operational choices were those who had downsized to the indie space – small studios producing games for smartphones. Alongside constrained autonomy, there was an observable trend towards greater standardisation and specialisation. Other recent studies have reported a team-based process, but with an emergent hierarchy of skills, ‘production processes have shifted from one-person craftsmanship to an intricate division of labour’ (Izushi and Aoyama, 2006: 1846), ‘where workers fulfil narrowly defined roles within a hierarchical production pipeline’ (Deuze et al., 2007: 342); and are subject to ‘regular interventions from senior management and also by the persistence of an emphatically hierarchical division of decision-making’ (Hodgson and Briand, 2013: 309). However there is an important proviso with respect to both autonomy and specialisation. Changes in the division of labour are relative to the degree of discretion necessary to produce complex cultural products. The resultant scenario does not mirror a ‘closely monitored, standardised work processes of the assembly line’ (Schumacher, 2006: 151).

To move on to our final set of issues, theoretical claims on the nature of work identity, attachment and the (absence of an) effort bargain also rest, implicitly, on an exaggerated notion of autonomy. There is evidence of strongly aestheticised identities in our research and elsewhere. High levels of job satisfaction and ‘brand’ identity can co-exist with insecurity or generic, underpaid work, as in the case of employees in fashion or animation (Stahl, 2010). While affective labour is a theoretically over-burdened term, affective attachments are pertinent. Such attachments and relative satisfactions are likely to diminish the willingness of creative workers to engage in collective action or resistance. However, our evidence is that traditional attachments are perceived by many to be increasingly in conflict with industry dynamics and work relations.

Just as workers in call centres incorporate perceptions of emotional labour into their understanding of the effort bargain, so workers in creative industries may do the same with affective attachments. They are neither – to refer back to earlier terminology – victims of capital or passionate play slaves, but rather purposeful agents, prepared to use the extent of scarcity of skills in the labour market to move on to other companies, sectors (such as IT) or smaller operations (see Martin and Deuze, 2009). Moreover, while the actions may be largely individual the understandings are collective, formed not merely through workplace interaction, but in social forums such as the Digital Labourers Federation in Australia (http://www.dlf.org.au/), which was referred to by many of our respondents. What we are seeing is not a static ‘identity theory of value’ (Arvidsson et al., 2010), but a dynamic interplay of identity and interests in particular contexts. Our evidence supports the understanding that games workers have a sense of themselves as creative artists, but they know they work in businesses with particular commercial logics. Among employees, the mere mention of the view that they had the time or inclination to play while at work drew instant contempt.

There are some potential objections to the conclusions we have drawn based on the nature of the evidence of types of firms and labour. It is fair to say that industries such as games in Australia are at the margins of the value chain, therefore they may be less typical of the industry in other contexts in North America and Europe. However, even in countries such as the UK, where the industry is more central, the pattern of expansion, consolidation under the control of publishers and post-GFC fragmentation is similar (NESTA, 2011). There is also the issue of types of workers. Given our access via firms, there was a bias towards full-time and permanent employees. It is certainly possible that interviewing freelancers would have told a partly different story. However, a large number of our respondents had previously been freelance and were explicitly asked to reflect on broad changes in the industry.

Given that our argument rests in part on the specificity of particular creative industries and labour, questions may also be raised about the typicality of the trends identified. There is certainly evidence from other such industries confirming, for example, trade-offs between creative autonomy and job security among jazz musicians (Umney and Kretsos, 2014). However, we recognise potential limitations. Games, visual effects (VFX) and animation are a cluster of technically oriented industries with some similar characteristics and trajectories. Even within that cluster there are differences. Considerable prominence has been attached to the ‘free labour’ provided by modders and other types of participant consumers. That undoubtedly exists, but is confined to certain segments of ‘virtual labour’ in the games and animation industries (Holts, 2013; Izushi and Aoymata, 2006: 1849; Postigo, 2007). If the scope is broadened to the full range of creative industries, the heterogeneity is considerably greater. Advertising, theatre and traditional media are different from the high tech industries and from each other. This may explain some of the differences in evidence across studies.

In conclusion and returning to Christopherson’s (2008) earlier plea, we have sought to reconnect changes in creative labour to changes in the wider production and political economy environments. Researchers should not read off characteristics of such labour from its internal qualities, including its project basis. Banks (2010b) argues that while autonomy is a condition of profitable artistic production due to the need for perceived novelty and creativity, it is always ‘negotiated’ within particular commercial and managerial constraints. This is a reasonable observation, though our emphasis has been less on the art–commerce tension and more on the positions and practices of particular firms and creative labourers. The starting point of our argument is that the most creative players – the development studios – are caught in a highly unequal struggle to capture value in the production chain. The medium-term trend, as reported by managers and employees alike, has been towards greater substantive subordination of creative producers to media giants. This is a common trend in audio-visual industries. The majority of the costs and risks of production are borne by myriad relatively smaller, and more ‘flexible’/expendable firms’ (Fitzgerald, 2014).

Industries such as digital games, VFX, animation and new media are relatively new. They have experienced periods of rapid expansion and contraction during which the context in which creative labour is mobilised and managed has undergone significant change. We are now entering a period in games, in which that dynamic is changing again as some firms seek to capture value in the smartphone market, outside of the control of publishers. But it would be naïve to believe that this heralds a new ‘back to the garage era’ of autonomous production. The gateway to new opportunities is firmly defined by Apple and Google (Parker et al., 2014). Though research is always conducted in particular places and at particular times, we need to look at trends over time before attributing paradigm breaking characteristics to ‘exemplary’ industries. The hype currently attached to games and other creative industries is eerily familiar to the early stages of the internet industry and web-based work (see Mayer-Ahuja and Wolf, 2007). History may repeat itself for industries, but it would be better if academics avoided the same outcome.