‘Creativity Loves Constraints’: The Paradox of Google’s Twenty Percent Time


Google office, Zurich

by Abe Walker


This article takes as its central object of study Google’s innovative time off program, colloquially known as ‘20 Percent Time’. This program represents a radical departure from conventional approaches to organizing the workday, and is quickly gaining traction in the technology sector and beyond. Under the directive of management, Google programmers devote 20% of their working time to independent projects of their own choosing. While this program offers the appearance of freedom, it is shown that 20 Percent Time actually has the opposite effect, intensifying managerial control and heightening exploitation. The paper situates 20 Percent Time in relation to existing literature on workplace control, especially labour process theory and the critical management studies tradition, as well as Derrida’s notion of the gift. It is shown that while 20 Percent Time is a crucial component of Google’s attempts to control their employees, it is fraught with contradictions that may contribute to its undoing. The paper concludes by posing some possibilities for resistance within and against 20 Percent Time.


Theories of workplace control typically have little to say about freedom. The workplace is often understood as a totalizing environment, saturated with obvious and subtle forms of coercion, so the struggle for freedom is best confined to realm of leisure, or more typically, left off the agenda entirely. Yet emergent neo-normative control theory posits that freedom is now the defining element of the contemporary workplace. Under neonormative control, firms encourage self-expression, embrace behaviors that would ordinarily be considered deviant, and permit employees a high degree of discretion over the structure and content of their labour (Fleming, 2009; Fleming and Sturdy, 2009; Cederström and Grassman, 2008). Companies like Google have created a work environment that resembles a playground more than a prison camp. But neo-normative theorists are quick to point out that this freedom is deceptive, for it operates as a cover for the intensification of exploitation, all while making employees believe they are truly free. In this essay, I offer a partial challenge to neo-normative theory by showing that workplace freedom has a tendency to extricate itself from the contractual social relation and reinvent itself in ways that suggest a more pure form of gift, almost in the sense of Derrida (1994). I conclude by arguing for what I call a strategy of amplification: The intentional misuse of freedom to generate more freedom.

The boundary between work and play is growing increasingly blurry under post-Fordism (see Kavanagh; Goggin, in this issue). Undeniably, the workday is being restructured in new ways that mimic the exigencies of the market. Nowhere is this more obvious than at Google Inc. which consistently ranks near the top of Fortune Magazine’s annual ‘Best Companies to Work For’ listing (Fortune, 2010), and has become the darling of the business press in part because of its unique approach to integrating ‘work’ and ‘play’. The centerpiece of the much-touted ‘Google model’ is a provision that allows its engineers to spend 20% of their workday on projects of their own choosing. Colloquially called ‘20 Percent Time’ (formally ‘Innovative Time Off’), this approach has won the accolades of management experts and popular media commentators. While the idea of 20 Percent Time is not new (the electronics and adhesives manufacturer 3M has had a similar program in place for decades), Google’s policy is the first to gain such widespread attention (3M, 2010).

20 Percent Time is only one component of a unique and often quixotic managerial philosophy some commentators have labeled ‘The Google Way’. In an interview for the trade publication Fast Company, Google VP of search products and user experiences Marissa Mayer lists nine ‘principles of innovation’ that guide Google’s management policies. While the majority center around Google’s openness and willingness to experiment (e.g. #2: Ideas Come from Everywhere; #4: Morph Projects, Don’t Kill Them), others emphasize the repressive aspect of workplace control – especially #8: Creativity Loves Constraints. On this point, Mayer explains, ‘This is one of my favorites. People think of creativity as this sort of unbridled thing, but engineers thrive on constraints. They love to think their way out of that little box…’ (Salter, 2008) This reflects Google’s general approach toward innovation: Creativity is encouraged, but only within certain predetermined and fairly rigid confines.

This idea is reconfirmed in axiom #3: A License to Pursue Your Dreams. Describing the company’s 20% rule, Mayer says, ‘We let engineers spend 20% of their time working on whatever they want, and we trust that they’ll build interesting things’. This is a loaded statement that merits some analysis. The two clauses in the sentence (‘we let…’; ‘we trust…’) explicitly frame 20 Percent Time as a kind of social contract (i.e. we allow employees a certain degree of freedom, but in return they are obligated to use it appropriately). Of course, the phrase ‘interesting things’ is deliberately vague, and Mayer does not specify the basis by which an idea might be judged interesting or uninteresting. So far from absolute freedom, Google grants its employees a provisional freedom which can apparently be recalled should their projects be deemed uninteresting. Managerial consultant Bernard Girard goes as far as to suggest employees experience this ‘freedom’ as a form of debt: ‘I owe something to the company because I’m given the free to invent and develop my own ideas’ (Girard, 2009: 67). Later in the same text, he notes, ‘what Google gives with one hand, it recovers with the other’, further highlighting the contractual nature of the arrangement (Ibid.: 66).

In the business and popular press, Google’s 20 Percent Time is typically framed as an example of corporate benevolence, a gesture of goodwill, or a gift bestowed on workers by enlightened managers. But Google’s gift is not a one-sided act of charity. Instead, as company profiler Bernard Girard (2009) correctly notes, the 20% rule is consistent with the anthropological concept of potlatch, for it establishes a debt that is laden with the expectation of reciprocity, but can never be repaid in full. In this sense, it creates what Mauss called a ‘gift economy’. However, in Derrida’s (1994) deconstruction of Mauss’s notion of the gift, he argues that the genuine gift is aneconomic, that is, its logic is beyond – though perhaps not wholly separate from – the economic cycle. If the Maussian gift obeys a circular logic, the Derridean gift is open-ended and nonreciprocal. But Derrida goes even further, suggesting that since the mere acknowledgement of the gift would imply a sort of reciprocity, the gift must go unacknowledged by both the giver and the receiver. The pure gift cannot even speak its own name; it must not be recognized as a gift at all. The gift that does not embody all these characteristics descends rapidly into mere market exchange. As I will show, 20 Percent Time meets none of Derrida’s criteria for gift giving, because it is bound up in a contractual relationship between employer and employee. As Girard (2009: 66) notes, ‘Google’s assumption is that the 20% free time it gives its employees will be returned in information, innovation, and increased loyalty’. Google’s largesse is a carefully calculated managerial technique which has little to do with generosity. Later, I will suggest that 20 Percent Time can only begin to approach the status of the Derridean gift if workers are willing to break their end of the bargain, and in doing so free 20 Percent Time from the logic of economic exchange. But first, I will examine Google’s 20 Percent Time policy in relation to the existent literature on worker control.

Theoretical background

Controlling or disciplining employees is a critical feature of all economic enterprises, and scholars of work have studied and theorized about workplace discipline for many decades. Much of the HRM literature assumes that managers have complete discretion over the organization of their workplace. Employers can locate their firm anywhere on a spectrum between authoritarianism and benevolence based their own personal proclivities. From this perspective, firms ‘choose’ how to organize their workplaces. Employers can embrace a certain kind of control regime based on their values and predilections, and then develop practices and policies to facilitate that choice. This strategic approach to workplace management has a long history, from the Human Relations School to Alvin Gouldner’s (1954) description of the shift from punitive to bureaucratic control in one factory, to more recent theories that emphasize ‘corporate culture’ (Deal and Kennedy, 1982). One of the most popular versions of this school of thought is Michael Porter’s (1980) concept of ‘high road’ and ‘low road’ employers. Porter distinguishes between high-road firms that offer ‘well-paid jobs’ that are ‘secure and satisfying’, versus low road firms whose jobs are ‘insecure and unsatisfying’, and he argues that companies can pick their strategy. Once chosen, however, a control strategy has important consequences for the kinds of people that are hired, the ways they are trained, and the systems of rewards they face. In press materials and official company documents, Google carefully cultivates its public image as a ‘high-road’ firm, and strives to create the impression that it does so by choice.

The work of Amitai Etzioni (1975) offers a variation on this theme. Etzioni identifies three control mechanisms intended to achieve a compliant workforce. Coercive compliance involves the implicit threat of physical sanctions to control organizational participants, and typically suffers from high levels of alienation. Normative compliance builds a strong sense of ‘belonging’ permeating all levels of an organization, based on selective hiring and socialization of employees into a common culture or value-system. Etzioni (1975) describes normative compliance as a system that uses shared values – buttressed by rituals, slogans and symbols – to encourage the active compliance of organizational participants. This approach requires ‘that workers and their superiors share a basic set of values’ and depends initially on selection of employees to ensure like-minded people, followed by training and socialization of employees into company norms and values (Etzioni, 1975: 36). Expressive leadership and symbolic rewards also play important roles in such a system: ‘attainment of culture goals such as the creation, application, or transmission of values requires the development of identification with the organizational representatives’ (Etzioni, 1975: 65). In a third remunerative type of control, organizational elites maintain control over lower participants by manipulating material resources, either providing material rewards for good behavior or withholding them to encourage compliance. According to Etzioni, organizations adopt one of these control mechanisms depending on their pre-defined goals, with profit-orientated organizations generally adopting remunerative strategies, cultural and religious organizations using normative techniques, and organizations concerned with order resorting to coercion. As I will show, Google cloaks itself in normative garb, but the fit is imperfect.

A separate school of thought holds that firms develop new forms of worker control in response to historical contingencies. For example, Stephen Barley and Gideon Kunda (1992) claim that American capitalism undergoes periodic shifts back and forth between disciplinarian and value-centric workplaces, which they label ‘rational’ and ‘normative’ respectively, and they identify forces that push toward one extreme and then the other. Early American capitalism was defined by the Fordist model of production, characterized by the assembly line, Taylorism, and strict authoritarian system of control. After WWII, Fordism gave way to human resource management, a discipline dedicated to producing a compliant workforce through less repressive means. Popular rebellions against workplace control in the 60s and early 70s were ‘answered’ in the form of participation schemes, ranging from the team concept to quality circles, which purported to give workers a measure of control in exchange for their tacit consent (Parker and Slaughter, 1988). More recently, as the dream of workplace control fades, boredom and bureaucracy have become the chief enemies, epitomized by the Dilbert comic strip and the Mike Judge film Office Space. Management responded with a variety of techniques designed to make the workplace more pleasurable and ‘democratic’.

The failure of normative control

Google exhibits some characteristic of ‘classic’ normative control, but as I will show later, differs from the model in some key respects. In a discussion that resonates with Etzioni, Slavoj Žižek (2006) describes Google as a ‘liberal communist’ regime, with more than a touch of irony. The tech companies IBM, Intel, and eBay are also liberal communists, as are Microsoft CEO Bill Gates, venture capitalist George Soros, and the prominent neoliberal spokesman Thomas Friedman. Žižek argues that liberal communists attempt to reconcile the advanced capitalist logic of the Davos-based World Economic Forum with the anti-capitalist spirit of the World Social Forum in Porto Alegre (an amalgamation he dubs ‘Porto Davos’.) Under this rubric, the relentless pursuit of profit is allegedly rendered compatible with the Leftist dreams of freedom, compassion, and justice. Though Žižek does not share Etzioni’s language, he is referring essentially to the normative organization.

In an article for the French popular magazine Technikart, Olivier Malnuit posits what Žižek describes as the ‘10 Commandments’ of liberal communism. Most are familiar tenets of benevolent capitalism, but Rule 6 is in a class of its own and deserves special mention: ‘You shall not work: have no fixed 9 to 5 job, but engage in smart, dynamic, flexible communication’ (cited in Žižek , 2006: 10). Of course, Malnuit is engaging in a rhetorical ploy: Work has not been abolished – in fact by most measures global work hours are increasing. But if work as such has been made invisible for some, it has been replaced with new intensive modes of interpersonal engagement. This is a daunting prospect. While there is a long history of scholarship in the anti-work tradition (Black, 1985; Aronowitz and DiFazio, 1994; Russell, 1932), none of these authors suspected their dream would someday be realized via a severely attenuated version of the Habermassian communicative ideal. Malnuit may as well be describing Google. Advances in programming have eliminated much of the repetitive, mind-numbing work that used to preoccupy software engineers – especially entering large chunks of code. So most of what Google’s senior programmers ‘do’ is not implementation, but imagining and conceptualizing new programming possibilities – in short, the sharing of ideas.

Following Žižek , Google’s largesse might be positioned within the trajectory of corporate paternalism. Just as Andrew Carnegie built libraries and museums in steel towns and Henry Ford provided discounted (but substandard) housing for his autoworkers, Google’s Sergey Brin offers gourmet snacks and high-end espresso at ‘micro-kitchens’ arrayed across the shopfloor. (An informal rule specifies that no Googler is allowed to be more than 100 meters from a micro-kitchen at any time, as Wakefield (2008) reports.) Other perks at Google include 25 days of vacation per annum, tuition reimbursement at $12,000/yr, fully paid maternal leave for up to 18 weeks, and a variety of others that pale by continental European standards but are considered extremely generous in the United States.

But Google’s ‘liberal communist’ regime may have a downside. As Andrew Ross (2004) has argued, the ‘humane workplace’ often operates as a cover for the intensification of exploitation. Ross’s detailed ethnographic study of the early dot-com start-up Razorfish reveals that beneath the camaraderie, tech workers work longer hours for less pay than they would otherwise. This is undoubtedly true at Google – the employee cafeteria at Google’s New York office serves a full dinner 4 nights a week, with the expectation that many Googlers will continue working up to and beyond dinnertime, and many of the company’s offices offer on-site laundry rooms, haircuts, and gyms – virtually eliminating the need to ever go home. Yet at Razorfish, the spirit of ‘Porto Davos’ proved unsustainable over time, and the company narrowly avoided collapse. In Ross’s narrative, the bursting of the dot-com bubble serves as a literary turning point, after which the company lays off the majority of the workforce, largely abandons new-age management techniques, and reverts to a more traditional control model.

In describing his three forms of worker control, Etzioni (1975: 7) cautions that employing multiple control strategies simultaneously proceeds at the expense of internal coherence: The multiple control mechanisms will ‘tend to neutralize each other’. The fusion of economic goals and normative compliance results in an ineffective type, characterized by ‘wasted means, psychological and social tension, lack of coordination …. [and] a strain toward an effective type’ (ibid.: 87, emphasis added). He suggests that organizations have a tendency over time to shift under the weight of their internal contradictions, to eliminate inconsistencies, and eventually move toward implementing a model with only one kind of control.

This was Razorfish’s fate. By its own admission, Razorfish strayed from its original humanistic ideals and moved toward a more traditional business model. Over time, the need for the organization to make profits and add shareholder value trumped its commitment to ‘good jobs’. To the extent that the company still portrayed itself as a socially responsible employer, it was deceiving customers, shareholders, and its allies in the business community. In this regard, Razorfish’s shifting system seems to be an exemplar of Etzioni’s theory of compliance and control. Following Etzioni, Ross predicts that humane firms will sink back into traditional model.

The rise of neo-normative control

But Google seems to have fared far better than Razorfish. Emerging after the initial wave of dot-com boom and bust, Google has enjoyed steady growth since its inception, which the global economic downturn briefly slowed but did not stop. If some normative firms like Razorfish drifted back toward regulation and repression, others moved forward, modifying the normative model ways that initially seemed to placate even the harshest critics. From the ashes of normative control emerged the neo-normative firm. In contrast to the normative organization, which suppresses individual identity in the name of a collective and internally-consistent corporate culture, the neo-normative organization exhorts the employee to express their identity and ‘just be themselves’. As Fleming and Sturdy (Fleming 2009; Fleming and Sturdy, 2009) have argued, neonormative control is in large part a reaction against stultifying normative control, which employees tend to regard as artificial or false. If the normative organization silences individual expression beneath a veneer of sameness, the neo-normative organization embraces the ‘authentic’ self with open arms. But Fleming and Sturdy quickly shift to a more cynical register, noting that the injunction to ‘just be yourself’ – even to the point of welcoming absurd and unruly behavior that would normally be barred in the workplace – is itself a cultural mandate: ‘Despite the enchanting vocabulary of “freedom” and “emancipation” on which the neo-normative control culture draws, such “possibilities” do not actually liberate the employee, but rather create new forms of identity control’ (Fleming and Sturdy, 2009: 158). Under the guise of reducing managerial control, the neo-normative firm merely foists managerial responsibility onto the employee.

Google is frequently cited as an archetypal example of the neo-normative firm (Fleming, 2009; Cederström and Grassman, 2008). Like other neo-normative organizations, Google’s model relies on a heady and unstable mix of freedom and control. While Etzioni’s models of organization involve overt system of control, manifest in different ways, neo-normative management seems to portend what Fleming calls the ‘control-free’ organization, or ‘liberation management’. Permitting employees to be themselves is presumed to result in higher motivation and productivity levels. So liberation management offers a version of freedom that is never completely removed from the shadow of control. 20 Percent Time is a prime example of this tension, for the policy offers a measure of freedom, but one that is constantly subject to modulation. Employees are required to document 20 Percent Time projects, and the company has dedicated at least one full-time employee (the unfortunately named ‘Director of Other’) to overseeing its 20% policy (Yen 2008). Indeed, Google takes an intense interest in ensuring that Googlers use their 20 Percent Time effectively. As Iyer and Davenport (2008: 8) noted in a piece for Harvard Business Review, ‘These percentages … are closely managed’. Similarly, in an editorial for the New York Times, Google software engineer Bharat Mediratta explains that employees are free to form grouplets to collaborate on 20% initiatives, but concedes that these grouplets often ‘need guidance to make sure they are aligned with the company interest’ (Medirara and Beck, 2007). There is a clear sense here that ‘the company interest’ (the profit motive) ultimately trumps employee initiative. Even more glaringly, in a letter to shareholders introducing the policy, company founders Sergey Brin and Larry Page explain the policy in terms that highlight its restrictive nature: ‘We encourage our employees, in addition to their regular projects, to spend 20% of their time working on what they think will most benefit Google’ (Google, 2004b, emphasis added). This description contrasts sharply with Mayer’s previously mentioned open-ended depiction of the program (programmers work on ‘whatever they want’). The fact that senior Google executives cannot reach an agreement on the proper framing of the policy reveals its complexity: It is fraught with contradictions.

It should be noted that new forms of workplace control do not necessarily eclipse old forms. Instead, control regimes are often cumulative; a neo-normative model does not preclude normative or even regulatory and disciplinary strategies. For example, at the peak of the economic slump, one report indicated that Google had begun ‘firing slackers’ (Carlson, 2008). While this did not appear to be a widespread phenomenon, it lent credence to the idea that Google’s unique mix of control and freedom is a classic case of the iron fist beneath a velvet glove. Following this logic, freedom serves as a cover for more traditional forms of workplace control. But by most journalistic and popular accounts, Google exhibits very few signs of traditional repression. Even when Google does restrict employee freedom, it typically does so in a way that is consistent with its overall culture of permissiveness. Google has perfected the system of ‘positive’ reinforcement (encouragement) to the point that ‘negative’ reinforcement (discipline) is largely held in reserve. For example, Google’s system of accountability demands that employees submit brief weekly reports documenting their use of 20 Percent Time and share their reports through peer review (Girard, 2009: 66). So even the evaluative aspect of the program is decentered. All things equal, the neo-normative firm displays a preference for permissive strategies, without disallowing other possibilities.

While 20 Percent Time is just one component of a larger system of neo-normative control, it merits special treatment because it claims to be different and distinct from the workday proper. That is, by setting aside 20% – and only 20% – of the workweek for ‘freedom’, Google tacitly acknowledges that the remaining 80% is not (as) free. If taken to its logical conclusion, neo-normative theory would suggest that the entire workday should be employee-directed and ‘liberated’ from managerial oversight. Fleming (2009) offers a number of examples of firms that take the ‘liberation management’ model much further than Google, some even extending a variation of 20 Percent Time to the entire workday. The ‘limit case’ would be a firm with no control structure whatsoever – manifest or concealed – and that relies entirely on employees’ self-regulatory impulses, thus virtually eliminating the need for management. Google’s unwillingness to go this far suggests something different is at work. A close examination of 20 Percent Time will allow us to position emergent theories of neo-normative control in relation to contemporary debates about worktime.

Time and measure

According to Google’s official narrative, there is a seamless link between its need for innovation and its employees’ creative impulses. In promotional materials, the company boasts that the policy offers its engineers the ‘free[dom] to work on what they’re really passionate about’ (Google, 2010). Yet a more cynical observer might describe the dynamic at work as recuperation. In this sense, Innovative Time Off is a cynical response to well-documented ‘problem’ of stolen time, self-reduction, and ‘skiving’ among tech workers (Schoneboom, 2007). Rather than penalizing its tech workers for wasting time on independent pursuits, Google reintegrates and revalorizes its employees’ anti-work tendencies. This fact is compounded by the reality that virtually all of Google’s full-time (non-contracted) engineers are salaried employees without set schedules. So the very notion of ‘paid time off’ is far less meaningful than it would be for a waged worker (whose income is solely dependent on the number of hours worked). While Iyer and Davenport (2008: 7) express concern that ‘As people spend more time experimenting, productivity in operational areas may suffer’, they fail to recognize that work at Google is not a zero-sum game. In the most pessimistic scenario, 20 Percent Time might simply be a tool for generating absolute surplus value by lengthening the engineers’ workday.

Here we see what Gibson Burrell has called ‘the differánce between non-work and work’ (cited in Korczynski, Hodson, and Edwards, 2006: 172). 20 Percent Time is presented as leisure, but is already imbued with that which is deferred – i.e. work. As Burrell notes, ‘Whatever is submerged, whatever is deferred, whatever is relegated to the other side of the line, then comes to play a role in constituting that which is left inside’ (Ibid: 171). Google engineers are performing leisure which is always-already interpellated by its opposite. Google’s model may offer a preview of a capitalist ‘endgame’ in which work – disguised as play – becomes all-consuming. On the one hand, 20 Percent Time represents the formal subsumption of non-work activity (‘recreational’ programming) by the Googleplex. On the other, 20 Percent Time represents a radical reorganization of the labour process itself (i.e. real subsumption) through the reappropriation of the collective knowledge of the general intellect. As capital strives to revalorize the creativity that would otherwise be lost to leisure, it simultaneously restructures the labour process in response to the refusal of work. Of course, as Carlo Vercellone (2007) has noted, real and formal subsumption are not mutually exclusive categories, but here they operate in tandem with striking results.

In 20 Percent Time, there is a distant echo of the early shorter hours movement, which demanded: ‘8 hours for work, 8 hours for sleep, 8 hours for what we will’. The problem is that shorter hours movement of the early twentieth century conceived of leisure as a bounded domain – completely distinct from work, and in an antagonistic relation to work (Hunnicutt, 1988). This may have had meaning 100 years ago, but it no longer makes much sense, especially for tech employees, for whom the work-place is essentially unbounded by time and space. One disgruntled former employee notes in a blog post, ‘my experience was that the people who spent all their time at Google were the ones that ended up on the sexier projects or in charge of things. It was frowned upon to leave right after dinner’ (Arrington, 2009). The same employee writes, ‘mgmt always stressed on “Putting some Extra Effort” – in other words “Spending some extra hours”’ (Ibid.). In the most pessimistic scenario, 20 Percent Time might simply be a tool for generating absolute surplus value by lengthening the engineers’ workday.

Examined in this historical context, Google’s 20 Percent Time initially appears as nothing more than a variation on time-honored traditions. The idea that periodic breaks boost employee productivity has long been management doxa. Indeed, early advocates for shorter hours in the United States claimed that a shorter working day would boost productivity by allowing employees time to recuperate, then work faster (Hunnicutt, 1988). Likewise, defenders of recess for schoolchildren claim that the remaining hours will be more worthwhile if the students have a chance to run around (International Play Association, 2010). An American company known as MetroNaps charges businessmen $14 for a 20-minute nap in a special chair, arguing, ‘Good quality sleep improves health, safety, productivity and ultimately the bottom line’ (Metronaps, 2010). There is widespread consensus, supported by scientific research, that time off improves worker satisfaction.

But innovative time off, as opposed to traditional time off, may represent a new, more advanced stage of worker control, because there is an explicit expectation that the time off will itself be productive. Unlike Google’s famed on-site masseuses and free catered meals, this is no standard employee perk. It has already become abundantly clear that Google has the capacity to re-monetize this ‘lost’ time. A Google vice-president recently claimed that half of new Google ‘products’ emerge from its 20 Percent Time, citing Gmail, the social networking service Orkut, and the ‘reality browser’ Google Goggles (Miraclemart, 2006). A post on official company blog titled ‘Google’s 20 Percent Time in Action’ glowingly describes a simple keyboard shortcut for Google’s RSS reader application that originated in 20 Percent Time Time (K, 2006). So even though there is no formal requirement that Googlers spend their 20 Percent Time on marketable projects, the internal company culture clearly encourages the type of work that might eventually pay off – and Google is eager to brag about those projects that do. With only a hint of exaggeration, Eric Schmidt has written ‘Virtually everything new seems to come from the 20% of their time engineers here are expected to spend on side projects. They certainly don’t come out of the management team’ (Google Operating System, 2005).

The demand that these employees ‘innovate’ in their time off is somewhat ironic given that innovation is already their job description. One Googler describes 20 Percent Time as an obligation that compounds, rather than replaces, his ordinary responsibilities: ‘This isn’t a matter of doing something in your spare time, but more of actively making time for it’ (Iyer and Davenport, 2008: 64). Since Googlers are paid on a salary basis, the amount of time that might be dedicated to a particular task is theoretically unlimited. Likewise, the length of the workday is limited only by the corporeal body’s inherent demand for rest and sustenance. Most people would be incapable of working more than a 14-hour day, although anecdotal accounts indicate that some Googlers approach this threshold. With multiple tasks all on limitless timetables, the notion of ‘free’ or ‘spare’ time is almost meaningless. So taking innovative time off either means pushing tasks aside to make room, or extending the workday by 20%. Later, the same employee notes, ‘Heck, I don’t have a good 20% project yet and I need one. If I don’t come up with something I’m sure it could negatively impact my review.’ (Ibid: 64). Googlers face intense pressure to design ‘good’ projects, and lack of initiative could potentially be a black mark on one’s record. In the company’s own words, it ‘requires engineers to spend a day a week on projects that interest them, unrelated to their day jobs’ (Google, 2004a, emphasis added). The somewhat awkward use of ‘requires’ (rather than ‘permits’ or ‘encourages’) lays bare the company’s intentions.

The Google model

To be fair, Google’s model has achieved some measure of success: The firm ranks high on employee satisfaction surveys, and many Google employees express nearly unbridled affection for their employer. Herbert Marcuse (1955) argued that capitalism can never satisfy humankind’s erotic needs, because Eros (leisure and pleasure) is always subordinated to alienated labour. But capital has sought to resolve this contradiction, by making work both less alienating and more erotic. Moreover, experiments with what Marcuse would call ‘non-alienated’ labour, whether under the Spanish Popular Front (Seidman, 1991) or through worker-owned cooperatives (Kasmir, 1996) have rarely lived up to their idealistic promise. At Google, Eros and alienated labour have been successfully integrated, with striking results. A job announcement webpage proclaims, ‘We love our employees, and we want them to know it’ (Google, 2010). A Google employee writing for the official company blog opens with the line ‘I love my job’ (K, 2006). A popular company profile coos ‘if you enable people to follow their passions, they’ll as much as work for free’ (Jarvis, 2009: 239). Another employee tells ABC News, ‘When you’re passionate about something and it’s an idea you believe in, you’re bound to work harder on it’ (Hayes 2008). While love is a common theme in new management rhetoric (Andersen and Born, 2008) the explicit linking of love with the intensification of work is perhaps a singular phenomenon. Love discourse is typically employed in ways that mask its relationship to speed-up and make it appear ‘unconditional’; at Google, the compulsory function of love requires no veiling.

If we accept Hardt and Negri’s (2000) contention that ‘cognitive labour’ is at the cutting edge of the global work regime under late capitalism, Google would seems to be on the cutting edge of the cutting edge, given its leading role within the tech industry. As Fleming (2009: 106) notes, ‘the Google approach has become the pin-up case for the advantages gained when the employees can “just be themselves” within the confines of the for-profit firm’. And within Google, engineers and programmers are the most prized employees. But if Google engineers are the avant-garde of the working class, the rest of industry is catching up quickly.

As Google gains notoriety, there has been a spate of books on the company’s unique management practices. Jeff Jarvis’ (2009) book What Would Google Do? is perhaps the most popular, ranked among the top 50 bestselling ‘company profiles’ on Amazon a full 18 months after its original publication. (The title is a play on the ‘What Would Jesus Do?’ (WWJD) slogan which appears on bumper stickers and t-shirts in the United States). Jarvis’s book is aimed at the managerial strata, and his explicit recommendation is that they should model their companies after Google. It is no accident that Google substitutes for Divine Providence in this formulation. True to form, Jarvis treats his subject with a near-religious reverence, carefully avoiding opportunities to be critical. In the final section of the book, entitled ‘If Google Ruled the World’ Jarvis imagines what the world would be like if Google took over private industries and public services. Under subheadings like ‘St. Google’s Hospital’, ‘Google Power & Light’, and ‘The First Bank of Google’, Jarvis (2009) describes a Google-ized world with giddy enthusiasm. Far from a nightmare scenario, Jarvis takes the position that a world ruled by Google would represent a vast improvement over the current mix of public and private control. With Google’s recent entry into the telecom industry, many of these scenarios don’t seem particularly far-fetched.

In their Harvard Business Review piece, Davenport and Iyer speak fondly of Google’s 20 Percent Time. For these authors, the only problem is that the 20% requirement has not been extended to enough workers. They recommend that this discretionary time should be expanded to include ‘nontechnical’ and ‘nonmanagerial’ employees (Iver and Davenport, 2008: 8). (They might have added ‘contracted’ and ‘hourly’ employees to this list, as these workers are these workers typically do not receive a 20% benefit, particularly when they work offsite.) 20 Percent Time is increasingly viewed as a model that can be extrapolated to other workplaces. ABC News claims that adopting 20 Percent Time at a ‘could make corporate America more productive’ (Hayes, 2008). The technology blog Web Worker Daily, which primarily caters to freelance and contracted workers who make up the bulk of online content providers, urges its readers to participate in their own self-management and self-regulation and ‘mix up the workweek’ by ‘setting your own 20 Percent Time’. But Web Worker Daily breaks with Google in one important way. At the end of the short article, the author makes a recommendation that would be anathema at Google. The author instructs readers to ‘mark the end of your workday. Only by setting actual work hours can we draw a clear line between work and the rest of our preoccupations’ (Roque, 2010). The irony, of course, is that few if any freelance web workers self-impose an arbitrary schedule – like Googlers, their work hours vary widely depending on the availability of assignments, other obligations, and personal preferences. A commenter on same blog helpfully points out ‘Highly productive employees will usually manage to fit this 20 Percent Time into their day – regardless of whether it is condoned by the company or not’. If these signs are taken seriously, 20 Percent Time may soon become far more pervasive.


In summary, 20 Percent Time is presented as pure freedom, but it is experienced as provisional freedom. If neo-normative management control styles purport to correct the stifling impulse toward conformity of normative control, they may only intensify it, all under the guise of doing the opposite. At minimum, 20 Percent Time appears to reduce if not eliminate the problem of slacking. At a different company that implemented 20 Percent (inspired by Google’s example), a manager gloated, ‘People are way more efficient about 20 Percent Time than regular work time. They say, “I’m not going to [expletive]ing do anything like read newsfeeds or do Facebook”’ (Pink, 2011: 95). In an echo of Foucault, it seems a self-disciplinary regime is far more efficient that either an overt regulatory regime (Taylorism) or an overt cultural mechanism (normative control). One Google employee explicitly compares 20 Percent Time to selfmanagement: ‘People are productive when they are working on things they see as important or they have invented, or are working on something they are passionate about. It is like they are the CEO of their own little company’ (Vise and Malseed, 2009: 132). A CEO would presumably take much more interest in the fate of their incipient enterprise, than would a disinterested employee. But it may be that offering employees a taste of freedom in the form of 20 Percent Time only increases their yearning for real freedom.

Alternatively, some workers may actually be willing to sacrifice perks and quasifreedom for good pay, even if coupled with a more conventional disciplinary structure. Cederström and Grassman (2009) suggest that employees may prefer a model where management’s intentions – no matter how pernicious – are laid bare to one that seems modeled on an elaborate lie. One disgruntled Google employee seems to confirm this view, writing on an internet message board:

As a full-time employee I prefer a good salary to gradually evaporating fringe benefits… My twenty-two year old greedy magpie self was wholly drawn in by the idea of having sashimi anytime I wanted without paying a dime. But as nice as it is having a cushy 401K and unlimited sick days, I was not willing to sacrifice my personal happiness and career fulfillment, not even for all the free kombucha I could drink. (Arrington, 2009)

On the balance, Google employees might choose to forgo incentives and benefits like 20 Percent Time in exchange for traditional bread-and-butter demands like higher pay and shorter hours.

The remainder of the article will sketch out some tentative possibilities for resistance within and against 20 Percent Time. Neo-normative theorists have relatively little to say about resistance, but like all theories of workplace control, their model begins from the assumption that resistance is a constant and persistent feature of any capitalist enterprise. If outright resistance is rarely visible in the neo-normative workplace, it is only because these firms employ practices that are normally coded as non-work – self-expression, play, ‘fun’, freedom itself – and recode them as work or otherwise reintegrate them into the workday. Yet even as the high-tech sector continues to expand, programmers and engineers are overwhelmingly non-unionized in the United States. According to the Bureau of Labor Statistics, unions represent only 2% of workers in ‘computer systems design’ and ‘software publishing’ (the job titles that most closely reflect the work performed by Google programmers), compared to 9% of workers throughout private industry (BLS, 2010). In the United States, the most successful effort to date to organize programmers is the WashTech union, sponsored and financially supported by the Communication Workers of American (CWA). WashTech operates primarily on the model of a ‘workers center’, avoiding formal contracts in favor of advocacy work and legislative campaigns (WashTech, 2010). While their initial base was Microsoft’s ‘permatemps’ (long-term ‘temporary’ workers), they have since expanded to workers at 90 unique firms. However, their reach does not extend to San Francisco Bay Area, where the highest concentration of tech firms is now located, nor does it include many full-time non-contracted workers (the subset of Google’s workers who are granted 20 Percent Time). Indeed, Google employees have never seriously entertained the prospect of unionization, nor has any US union attempted to organize the company.

In his famous ‘Postscript on control societies’, Gilles Deleuze forecasts an ominous future for trade unions:

tied to the whole of their history of struggle against the disciplines or within the spaces of enclosure, will they be able to adapt themselves or will they give way to new forms of resistance against the societies of control? (Deleuze, 1995: 182)

One thing is certain: To the extent that unions are predicated on the ‘work-place’ and the ‘work-day’ as known quantities, the old trade unionism has no resonance in the technology industries. Indeed, much of Marxist theory rests on the assumption that workers experience exploitation as individuals, and that the rate of exploitation (as a unit of time) can be calculated on a per-worker basis. Yet this calculation breaks down when dealing with cybertime and posthuman bodies (Clough et al., 2007). It may be the case, therefore, that old forms of resistance are obsolete. For Fleming and Sturdy (2009), the preferred mode of resistance against the simulated individuation of neonormative control is via a return to the collective. A solidaristic response might be ideal, but at present, traditional forms of collective action do not seem to have much resonance in the high-tech sector. The answer may lie not in collapsing the individual back into the collective, but by extending the logic of self-expression to its natural conclusion. At present, Google only allows transgressions which do not fundamentally call the corporate culture into question. In a culture that nurtures transgressive behavior, the only bad subject is the one who exceeds the limits of ‘acceptable’ transgression.

How would Googlers refuse to return the ‘gift’? How might Google employees take full advantage of their employer’s largesse, and struggle against the becoming-productive of their playtime by pursuing projects that cannot easily be re-monetized (or perhaps even generate negative value)? Whether through coercion or through persuasion, most Google workers seem to use their 20 Percent Time in ways that are productive – in both the Foucauldian sense (generating new subjectivities) and the corporate lingo sense (increasing output). But this need not be the case. Two possibilities exist. Googlers might resist 20 Percent Time directly by fighting for ‘real’ breaks that are not governed by the logic of productivity. In other words, they might try to recover ‘authentic’ leisure in a workplace that is disguised as a site of leisure. This strategy would draw upon the traditional demands of the shorter hours movements: A shorter workday and workweek, longer lunches, additional paid holidays, more vacations. This is an appealing project, but another possibility exists.

Rather than oppose 20 Percent Time outright, Google’s demand for innovation might be extended to its natural conclusion. Googlers might take very seriously their employers mandate that they innovative during their time off. I am referring here to Negri’s (2005: 270) understanding of innovation: ‘The workers’ overturning of the totality of the reproductive conditions of capital’. Googlers might engage in sabotage, using innovative time off in ways that are destructive. Here sabotage refers not necessarily to destroying machinery, but to the concept as theorized by Elizabeth Gurley Flynn (1916): The conscious withdrawal of efficiency. In order to expose neo-normative control’s latent repressive function, Googlers might pursue projects that cannot easily be re-monetized (or perhaps even generate negative value). Recommended activities include sleeping, drinking on the job, writing poetry, releasing proprietary software to the public under an open-source license, or, the ultimate insult, blogging about creative uses of 20 Percent Time. Googlers might use 20 Percent Time as a staging ground for unrestrained hedonistic excess. Again, the point is not to reclaim some ‘authentic’ desire (versus the fake desire that Google attempts to cultivate), but to collectively carry the logic of innovation to its natural conclusion – in other words, to test the limits of their employers’ ‘benevolent communist’ regime. The point is not that these activities are non-recuperable – surely they are – but in short term, they represent acts of employee sabotage in ways that designing a new email client does not. In other words, the question is not how to reclaim leisure as a ‘pure’ and unfettered space beyond the reach of capital (if this were even possible!) but how to disrupt the cycle of capitalist valorization with the understanding that leisure as such may no longer exist.


Iyer and Davenport (2008: 4) provide a visual model of Google’s 20 Percent Time. The authors explain that Google’s innovative success is a direct result of carefully-cultivated dynamic relationships with employees, consumers, and third-party stakeholders. Their diagram describes ‘Google’s Innovation Ecosystem’ – a naturalistic metaphor that evokes both cybernetic models of information flows and ecological cycles. Their diagram has at least an elective affinity with elementary school-style visualizations of the Water Cycle. Consumers, Innovators, Advertisers, and Content Providers all appear within concentric circles, with a mysterious entity dubbed ‘The Google Platform’ in the center. The Google Platform refers not to the corporation per se, but to Google as an ethereal concept. Double-headed arrows indicate feedback loops between the stakeholders and The Google Platform. Crucially, the system is precisely calibrated such that equilibrium is constantly maintained. The model is a creation of the authors, not Google, and the authors give no clear indication that Google executives’ own selfconcept aligns with their own. But in a sense, whether or not Google actively understands itself as an ‘ecosystem’ is beside the point. For Iyer and Davenport, the appeal of this model is that innovation appears as a fully enclosed system – all parts contribute the whole, and there is no connection to the outside. Their diagram exists on a two-dimension plane, and is apparently autopoetic. In a faint echo of Derrida’s sacrificial non-gift, the diagram is a circle, an enclosure, a reciprocal market relation.

But this is not the only possible model. The Google Ecosystem is predicated on the active complicity of all parties. By reneging on their end of the ‘deal’ and refusing to sign off on the implicit social contract, Google employees can break the cycle of reciprocity. This will require reimagining 20 Percent Time not as a market relation, but as a ‘gift’ that need not be returned. Indeed, this is a necessary (if insufficient) condition for freedom.

When critical labour process scholars discuss the postmodern workplace, a tone of haughty disdain often colors their writing, as if to suggest that the class conscious programmer ought not derive any enjoyment from his breakroom foosball table. A symptomatic reading of neo-normative texts reveals a certain anxiety that ‘illusionary’ freedoms might ultimately prove seductive. At times one even detects a trace of nostalgic longing for the despotic workplace where compliance is guaranteed by the abusive foreman, not the sashimi buffet. The prescription is obvious: Peel away the layers of window dressing, and the neo-normative firm reveals its true nature; raw, naked exploitation. For these thinkers, the permissive or benevolent employer is an elaborate scam that must be unmasked in order to build an effective political project. But in their rush to condemn the ‘false’ freedom of the modern workplace, neonormative theorists ignore the possibility that this freedom – whatever its limitations – might also constitute a zone of vulnerability. Neo-normative workers might be better served by a strategy of oblique resistance than head-on combat: A subject-position of ‘familiar foreignness’.

Recasting 20 Percent Time as a gift will require moving beyond the level of analysis provided by the critical labour process tradition. Etzioni (2009), Žižek (2006), and Cederström and Grassman (2008) all have a common project – to cut beneath the deception and expose the humane workplace as an elaborate lie. They each argue, in somewhat different registers, that humanistic ideals cannot be realized in the context of a private enterprise that is legally obligated to maximize shareholder dividends. Permissive styles of managerial control must be condemned because they mask the underlying truth of exploitation. But Deleuze offers another way of thinking. In his commentary on Nietzsche, Deleuze calls for a Dionysian affirmation of life as against the ascetic negativity of the truth-seekers. He conjures ‘A thought that would affirm life instead of a knowledge that is opposed to life’. Under this logic, ‘[Life and thought] would go in the same direction … smashing restrictions … in a burst of unparalleled creativity’ (Deleuze 2006: 101). It is this creativity that 20 Percent Time renders possible.

As I have argued, the neo-normative corporation grants freedom provisionally as part of an implicit social contract. The unstated expectation is that employees will respond with loyalty, diligence, and above all gratitude, or risk being stripped of their freedom (and employment). Though it presents itself as a gift, neo-normative freedom is recallable, contingent, conditional, and deeply subsumed within a contractual logic. But a closer examination of the Derridean gift will reveal the impossibility of the gift within a system of neo-normative control.

Recall that for Derrida, contractualism stands in direct contrast to the logic of the gift. To be sure, this not to say that the gift is ignorant of its contractual Other. However, as Derrida (1994: 7) insists, ‘[the gift] must keep a relation of foreignness to the circle, a relation without relation of familiar foreignness’. The pure gift must be recognized by neither the doner nor the donee, lest it revert back to the logic of market exchange. Under the reciprocal terms of a contract, the event of the anti-gift forms into a circular pattern – recognition/gratitude/obligation/counter-gift – thereby annulling the very possibility of the gift itself. As evinced throughout this essay, there can be no question that Google is self-consciously aware of its 20 Percent Time and carefully manages an image that hinges largely on its supposed generosity. But in the pure gift scenario, the donee is already and immediately free from the debt obligation that the contractual antigift necessarily entails. Thus, it may be that Google’s ‘love’ for its employees is of the unrequited variety. Corporate paternalism finds its chief adversary in the figure of the spoiled child: Showered with gifts yet ungrateful and infinitely demanding.

At the moment when the cycle of reciprocity is broken, it is clear Google has already mis-recognized its gift by discounting its unpredictable nature. Google may ‘present’ its gift as means of encouraging innovation, increasing productivity, and winning the sympathy of its employees, but the gift knows nothing of the cold, calculating logic of measurement; it cannot be parceled out in neat 20% segments, for ‘the gift is excessive in advance, is a priori exaggerated’, it is an ‘infinite obligation’ (Derrida, 1994: 38). If there is a cyclical logic at work here, it is only anti-economy of blowback, a set of unintended consequences that cannot be contained. There is a volatile lawlessness and ungovernable nature to the gift that makes a mockery of management’s best laid plans: ‘The gift and the event obey nothing, except perhaps principles of disorder, that is, principles without principles’ (Ibid.: 123). This is a strategy of amplification: the intentional misuse of freedom to generate more freedom – less a ‘strategy of refusal’ than a perverse affirmation.

By granting 20 Percent Time, Google is demanding that its employees innovate – but only within predefined boundaries. Yet Google has given rise to an excess that eludes capture – a creativity that abhors constraints.

All photographs of Google offices by Pixel Pro Photography

Piece originally published at Ephemera | Creative Commons License


3M (2010) ‘Employee Engagement’ [].

Andersen, N.A. and A. Born. (2008) ‘The employee in the sign of love’, Culture and Organization 14(4): 325-343.

Aronowitz, S. and W. DiFazio (1994) The jobless future: Sci-tech and the dogma of work. Minneapolis:University of Minnesota Press.

Arrington, M. (2009) ‘Why Google employees quit’, Tech Crunch [].

Barley, S.R. and G. Kunda (1992) ‘Design and devotion: Surges of rational and normative ideologies in managerial discourse’, Administrative Science Quarterly, 37(3): 363-399.

Black, B. (1985) The abolition of work [].

Bureau of Labor Statistics (2010) ‘Computer system design and related services’ [].

Carlson, N. (2008) ‘Google finally starts firing slackers’, Business Insider [ google-finally-starts-firing-slackers].

Cederström, C. and R. Grassman (2008) ‘The masochistic reflexive turn’, ephemera, 8(1): 41-57.

Clough, P.T., G. Goldberg, R. Schiff, A. Weeks, and C. Willse. (2007) ‘Notes towards a theory of affectitself’, ephemera, 7(1): 60-77.

Deal, T. and A. Kennedy (1982) Corporate cultures. New York: Perseus.

Derrida, J. (1994) Given time: I. Counterfeit money. Chicago: University of Chicago Press.

Deleuze, G. (1995) ‘Postscript on control societies’, Negotiations. New York: Columbia University Press.

Deleuze, G. (2006) Nietzsche and philosophy. New York: Columbia University Press.

Etzioni, A. (1975) A comparative analysis of complex organizations. New York: The Free Press.

Fleming, P. (2009) Authenticity and the cultural politics of work: New forms of informal control. New York: Oxford University Press.

Fleming, P. and A. Sturdy (2009) ‘“Just be yourself!” – Towards neo-normative control in organisations’, Employee Relations 31(6): 569-583.

Flynn, E. G. (1916) Sabotage: the workers withdrawal of industrial efficiency [


Fortune (2010) ‘100 best companies to work for 2010’ [].

Girard, B. (2009) The Google way: How one company is revolutionizing management. San Francisco: No Starch Press.

Google (2004a) ‘Google gets the message, launches Gmail’ [].

Google (2004b) ‘2004 founders IPO letter’ [].

Google (2010) ‘The engineer’s life at Google’ [].

Google Operating System (2005) ‘Google CEO Eric Schmidt: The 70 percent solution’ [].

Gouldner, A. (1954) Patterns of industrial bureaucracy. Glencoe, IL: Free Press.

Hardt, M. and A. Negri (2000) Empire. Cambridge, MA: Harvard University Press.

Hayes, E. (2008) ‘Google’s 20 percent time time factor’ ABC News [].

Hunnicutt, B.K. (1988) Work without end: Abandoning shorter hours for the right to work. Philadelphia: Temple University Press.

International Play Association. (2010) ‘IPA/USA declaration’ [].

Iyer, B. and T.H. Davenport. (2008) ‘Reverse engineering Google’s innovation machine’ Harvard Business Review, 86(4): 58-68.

Jarvis, J. (2009) What would Google do? New York: HarperBusiness.

K., A. (2006) ‘Official Google blog: Google’s “20 percent time” in action’, The official Google blog [].

Kasmir, S. (1996) The myth of Mondragon: Cooperatives, politics and working class life in a Basque town. Albany, NY: SUNY Press.

Korczynski, M., R. Hodson, and P.K. Edwards (2006) Social theory at work. New York: Oxford University Press.

Marcuse, H. (1955) Eros and civilization. New York: Beacon Press.

Medirara, B. and J. Beck. (2007) ‘The Google way: Give engineers room – New York Times’, New York Times [].

Metronaps (2010) ‘About Us’ [].

Miraclemart (2006) ‘Marissa Mayer at Stanford University’ [ =soYKFWqVVzg].

Negri, A. (2005) Books for burning: Between civil war and democracy in 1970s Italy. New York: Verso.

Parker, M. and J. Slaughter (1988) Choosing sides: Unions and the team concept. Boston: South End Press.

Pink, D. (2011) Drive: The surprising truth about what motivates us. New York: Penguin.

Porter, M. (1980) Competitive strategy. New York: The Free Press.

Roque, C. (2010) ‘Mix up the workweek by setting your own ‘20-percent time’ posted 3-12-10 at 9:00 to Web Worker Daily [].

Ross, A. (2004) No-collar: The humane workplace and its hidden costs. Philadelphia: Temple University Press.

Russell, B. (1932) In praise of idleness [].

Salter, C. (2008) ‘Marissa Mayer’s 9 principles of innovation’ Fast Company [].

Schoneboom, A. (2007) ‘Diary of a working boy: Creative resistance among anonymous work bloggers’,

Ethnography, 8(4): 403-423.

Seidman, M. (1991) Workers against work: Labor in Paris and Barcelona during the Popular Fronts.

Berkeley, CA: University of California Press.

Vercellone, C. (2007) ‘From formal subsumption to general intellect’, Historical Materialism 15: 13-36.

Vise, D. and M. Malseed (2008) The Google story. New York: Random House.

Wakefield, J. (2008) ‘Google your way to a wacky office’ BBC News [].

WashTech. (2010) ‘FAQ’ [].

Yen, Y.W. (2008) ‘Help Wanted: Google – Director of other’, Money [].

Žižek , S. (2006) ‘Nobody has to be vile’, London Review of Books, 28(7): 10.

About the Author:

Abe Walker is a doctoral candidate in sociology at the CUNY Graduate Center, an adjunct instructor at Queens College (CUNY) and a rank-and-file agitator within AFT Local 2334.