Digital culture promises a business model that oscillates between blockbusters and niche products. This model leads to upheavals in the market that directly affect, for instance, the quality and availability of computer games. Just as we rely on cognitive models in our interactions with individuals, our understanding of complex social situations depends, accordingly, on models of a larger scale, based on longitudinal studies of our basic assumptions. In the field of finance, for instance, there are those who, trusting the rational expectation that markets will “naturally” correct themselves, seek “beta” in the wisdom of crowds, and there are reflexive behaviorists who, assuming that the world is in a constant state of flux or disequilibrium, look for “alpha” opportunities in the madness of the marketplace. Media studies is similarly divided between those who regard crowdsourcing as an impetus for transformation and progress (from the printing press to the mass media of the twentieth century and on to the highly differentiated media of networked cultures) and those who believe that the dispersion of the masses into ever more channels has produced a droning noise that threatens to smother whatever informational value or entertainment value might exist.1

In an effort to mediate between these seemingly irreconcilable positions, I have examined the planning cycles of the entertainment-software industry, although of course the disclaimer must be kept in mind that past performance is no guarantee of future results, whether we are dealing with finance or media studies. The question is what media studies may gain when speaking of the “new masses,” regardless of whether such insight can be expressed as formulaically as an alpha or beta deviation, in looking into to the business cycles of American computer-game developers.2 In this context, an interesting role is played by Metacritic and Gamerankings, two websites that aggregate criticism on the Internet.3 These sites have become so influential in the gaming industry that they are often used to determine the salaries and bonus packages of its managers and executives. Finally, the discussion will turn to synthetic worlds, for it is time for observers of new media to familiarize ourselves with trends in the entertainment software industry that have come to govern both the quality and the variety of the products being offered. While media studies typically approached the phenomenon of computer games in terms of media archaeology or of ideological critique, my intention here is to underscore the influence of management and of journalistic criticism over the hardware and software that are developed for this new mass medium.

I. The New Masses

Just as cars, radios, televisions, and record players enriched the experience of our daily lives, the same now goes for computer culture, and especially for computer games. It is an axiom of media economics that one observes a dual market: an exchange of wares on the one hand and the package and sale of attention on the other.4 Technical infrastructures enable and facilitate the development of creative content, which is then advertised and made available to consumers, but media consumption is not only the consumption of films or computer games; it is also the consumption of the time that someone is willing to devote to such things. In this regard, the supply and demand of production flows in the opposite direction of the supply and demand of attention. Information can be packaged and sold without being scarce or used up by that process; digital media bring a highly segmented and fragmented audience in contact with suppliers of content and distraction. While it is true that attention is limited, for media economics it is just as relevant that the consumption of a film, a CD, or a game does not use them up – they remain available for further circulation. At the same time, it is no less relevant to note that, both in terms of quantity and quality, consumer goods in the entertainment sector exhibit a growth rate that far exceeds the growth rate of attention itself. Whereas more movies, music, and software become available every year, the collective attention of the new masses does not increase proportionally. In the past, moreover, when it was difficult to start a new television station, radio station, or newspaper, the supply of such products was limited to a far greater extent by the barriers to market entry. A classic example of this is the vertical integration of Hollywood studios during the 1950s and 1960s.5 In the case of traditional media, their business has mostly been limited by the costs of infrastructure, operation costs, and production costs, whereas the price they have paid for attention has been relatively low. Today, attention is no longer so cheap. In 1982, television advertising in the United States occupied only six minutes of every hour of broadcasting, yet this number had doubled by 2001 and has continued to climb. The same trend applies to newspapers, magazines, and radio stations. For these investments in attention, it is relatively insignificant to distinguish between advertising time and advertising expenditures; as long the participants in these industries find it more advantageous to invest in advertising instead of in production quality or infrastructure, the budgets allocated for advertising will continue to grow. The implication of this for entertainment software is that attention is not a factor in the creation of quality, at least so long as it is cheaper to purchase attention than it is to develop a computer game. Although it would be nice if the market were such that quality and popularity went hand in hand, within the sphere of the mass media there is an inefficient relation between production quality and public success, and this inefficiency is modulated only by a sharp rise in advertising costs. A film can be marketed in the movie theater, on a DVD, on television, in foreign countries, as well as in the form of toys, books, posters, T-shirts, and so on. This explains why film producers depend on blockbusters and why the music industry relies so heavily on hit songs. During the past two decades, the computer-game industry has come closer and closer to adopting this model. As advertising costs continue to rise, however, less money is left for the production of quality; thus either the investment in production must rise overall, or quality will suffer.

The Internet age promised to provide greater flexibility for producing, distributing, and marketing media content in novel ways, but this hope raises questions about the reality and nature of the vaunted new masses. The assumption was that the new media landscape will no longer be dominated by production costs but rather by the costs of attention, because new technologies would lower the cost of production and distribution. Nonetheless it is doubtful that these new media conditions will lead to direct correlation between the popularity of a product and its quality. A detailed analysis of the computer-game market in the United States will reveal why this is the case.

The American computer-game industry has been boasting since 2007 that its earnings surpassed those of the film and music industries. The total market size of nearly nineteen billion dollars for 2007 was three times larger more than a decade earlier; half of that money was spent on entertainment software and the other half on hardware, with consoles and mobile devices accounting for larger share than computers and displays. Whereas the American music industry shrank by nearly ten percent between 2002 and 2007, and the film industry remained stagnant, game manufacturers grew by more than twenty-eight percent. Of course the price of a game is usually higher than that of a movie ticket or a CD. Of the sixty or so dollars charged for a new console game, a quarter represents the markup by the retailer, an additional fifteen to twenty percent pays for royalties or licenses (to comic-book authors, athletes, film studios, or other copyright holders), ten percent is set aside for refunds and theft, ten percent pays for amortized investments in software (in the Unreal Engine, Massive, or any other code that serves as the basis for multiple games), and another ten percent is often paid to hardware manufacturers such as Nintendo, Sony, or Microsoft for providing their interface. After packaging and distribution are taken into account, this means that approximately twenty percent of a game’s retail price goes to its development.

Over the past ten years, however, development costs have soared while retail prices have remained stable. Because of the demand for more complex graphics, larger game worlds, and more sophisticated interactions, the average investment in a new title has increased from approximately fifty thousand dollars for a sixteen-bit game to three million dollars for a sixty-four-bit game. In other words, it typically cost around two million dollars to develop a game for the original PlayStation, games for the Xbox could cost between three and seven million dollars and games for the Wii cost as much as twelve million; but for the PS3 and Xbox360, the cost of developing a game easily exceeded twenty million dollars, and several ambitious projects in fact made even that number seem rather low. At the same time, the development cycle for games has become longer and longer, and this puts further pressure on profit expectations. Thus it is no surprise that the fastest-growing sector on the internet has been so-called “massively multiplayer online role-playing games” (MMORPGs). For in addition to being able to reach a broader demographic, online games of this sort also generate additional revenue streams beyond the one-time sale of packaged software, through subscriptions, advertisements, and secondary markets.

Before we turn to online games, however, a few observations about the market for console games should be made. There is a basic distinction between game developers that produce their own hardware (such as Microsoft, Sony, or Nintendo) and those that are not bound to any single platform and thus enjoy the benefit of having multiple options at their disposal (such as Ubisoft, Activision Blizzard, Electronic Arts, Take Two, or THQ). Within the game industry, the software market share of hardware developers is relatively small. Nintendo’s sits around eleven percent, Sony’s around four percent, and that of Microsoft is little more than three percent. The remainder of the market is divided among the large hardware-independent developers, with Electronic Arts leading at twenty-four percent, followed by Activision Blizzard with twenty percent, Ubisoft hovering above six percent, and a long list of smaller companies behind them. Formerly influential brand names such as Disney or Sega hardly play a role any more, while up-and-comers such as Valve and Zynga have entered the fray. Although it can be advantageous for hardware-independent developers to offer a given game on multiple consoles (examples include Take Two’s Grand Theft Auto or Ubisoft’s Assassin’s Creed), over the past few years it has also been quite profitable to combine one’s own hardware and software. In this regard it should not be forgotten that both Sony and Microsoft have strategically lowered the price of their consoles in order to stimulate software sales, so much so that both the Xbox360 and PlayStation 3 have at times been selling below cost. During the last development cycle, a great many below-average games were sold for the PlayStation, but now Nintendo’s Wii and DS are the most obvious niches in the market without any direct correlation between game sales and game quality. In the past, Electronic Arts, Activision, and THQ were able to remain profitable not only by selling big hits like Call of Duty but also by selling a number of other products that were predominately targeted at children or casual players. In 2009, however, when Electronic Arts laid off a tenth of its employees in order to save 120 million dollars, the CEO explained that advance reviews in trade publications and on industry websites had begun to influence sales. Two websites, in particular, were identified as responsible for aggregating a game’s ratings and providing its overall score, a score that is in turn influencing sales, market shares, and stock prices: Metacritic (which also rates films and music) and Gamerankings.6

Unlike film and music criticism, which is at least informed by the total impression made by one work or another, game journalists usually have to file their reports about new products long before they can become completely familiar with the entertainment software under review (even if they have a full week to play a given new game). It is characteristic of the new media landscape for these two websites to aggregate and summarize game critiques from European, Japanese, Canadian, Australian, and American magazines in a single collective score, even if these magazines themselves do not use a grading scale. Metacritic and Gamerankings typically publish formulaic summaries of journalists’ first impressions about a game even before any sales data has been made available, and this is the source of their influence over the industry, over Wall Street, and over the general public itself.

When Activision’s game Spiderman 3 reached the stores on a Friday in May of 2007, for example, the tepid early reviews by industry journalists (50/100) reverberated in the form a five-percent loss in the company’s share price, followed by additional stock price drops in subsequent weeks. In August of the same year, by contrast, the game Bioshock received a nearly perfect score (96/100) in its advance reviews, and Take Two’s stock rose by twenty percent within a few days. According to Activision’s CEO Robert Kotick, a systematic study of 789 games for Sony’s PlayStation 2, conducted during the year 2005, had demonstrated a clear correlation between game ratings and sales numbers. Sales of an Activision game doubled for every five points it received over 80, and for this reason developers became convinced that ratings not only reflected the quality of games but their sales potential as well. “Everyone wants 85 or better,” as Jim Ward, the president of LucasArts, in 2007 told the Wall Street Journal.7 Since 2005, in fact, numerous developers including Activision and Take Two have incorporated these collective game ratings into their employment contracts. Even industry partners take such aggregate scores into account during their contract negotiations. Since 2004, for instance, Warner Brothers has directly relied on such ratings when calculating the licensing fees for games based on its movies: The better the game, the lower the fees, whereas more is charged for lower-quality games that risk tarnishing a film’s reputation.

It was not until early 2011 that the industry influence of Metacritic and Gamerankings began to wane. The turning point was when the game Homefront (a THQ product in which the North Korean army invades the United States in the year 2027) enjoyed respectable sales despite its low ratings. Early reviews of the game were negative enough to cause the company’s stock price to drop by more than twenty percent in a single day, and yet the sales numbers stubbornly went on to defy expectations.8 Metacritic had given Homefront an overall rating of 72. Although twelve of the twenty-eight reviews were positive (80/100 or better), most reviewers thought the game was a flop (the ratings ranged from 40 to 93). Nevertheless, Homefront spent nearly an entire month as a bestseller on Amazon and elsewhere and quickly turned a profit for THQ, which ultimately sold some two million copies of the game.9 Such occurrences are interesting for media studies because they complicate the conventional distinction between quantity and quality. Under the conditions of digital culture, media theorists have to concern themselves not only with matters of quality but with quantitative data as well.

It can be revealing to correlate aggregate ratings of games with sales data collected by the NPD Group, a market research company, and it casts a grim light on the role of criticism – in the entertainment software scene it no longer plays, it seems, the role arbiter of cultural merit. Whereas film criticism rarely reflects the taste of the general public, and music criticism usually has little patience for pop hits, in the world of computer games there is a direct and quantifiable correlation between a game’s reviews and its popularity. Games with the highest ratings (again, these scores are simply an average of the first impressions reported by a few dozen industry journalists) sell three times as well as games with collective ratings between 80 and 90. The latter, in turn, sell twice as well as games rated between 70 and 80, which themselves have twice the sales of games with ratings between 60 and 70. It is therefore no surprise that a mere four percent of all the games sold between 2000 and 2006 were responsible for a third of all revenues, and that this situation has become even more pronounced in recent years. By now, a fourth of the titles sold are responsible for eighty-six percent of the industry’s total earnings. Both UBS and Deutsche Bank conducted longitudinal studies (released in 2007 and 2010, respectively) that together analyzed more than 1,600 games. According to UBS’s findings, the fifty games with ratings over ninety, which constituted no more than three percent of all available games, outsold all the other games combined, while twenty-three percent of the games on the market had received an aggregate score of 60 or lower. UBS also compared the multi-year average ratings of products by Activision, Electronic Arts, Take Two, and THQ with those of the products made by Microsoft, Sony, and Nintendo. Between 1997 and 2006, the market share held by the hardware producers dropped from forty-four to seventeen percent, while that of the four large American third-party developers increased from sixteen to forty-seven percent. Whereas Activision did quite well with Call of Duty, Marvel Ultimate Alliance, Tony Hawk, and Guitar Hero, a number of its film adaptations – notably Spiderman 3, Shrek 3, and Transformers – performed rather poorly. During the same time period, Electronic Arts was able to rely primarily on Madden Football and Need for Speed, which respectively accounted for twenty percent and five percent of the company’s five-year revenues, while its own film adaptations, namely Harry Potter and Lord of the Rings, likewise proved to be less profitable (each accounted for a mere five percent of revenues). CEO John Riccitiello blamed Electronic Arts’ lost market share directly on aggregate game ratings: “We did not have any internally developed breakaway titles and none of EA’s internally developed titles reached a Metacritic rating of 90 or greater.”10 Take Two relied even more heavily on a single franchise, Grand Theft Auto, which earned more than a billion dollars between 2000 and 2005 (more than half its total earnings), while other offerings such as Midnight Club and Max Payne were worth only a tenth of that. For its part, THQ was mainly reliant on licensed games from Disney, Pixar, Nickelodeon, and World Wide Wrestling Entertainment. In the case of these games (think of SpongeBob SquarePants, Scooby Doo, Rugrats, Finding Nemo, and Power Rangers), the correlation between sales and reviews was less explicit. The overarching trend, however, has led the industry to change its ways and tilt increasingly to the winner-take-all model: A decade ago, the twenty-five best-selling games accounted for a quarter of the market, whereas now the same number of bestsellers makes up more than forty percent of the games sold.

In this precarious situation, it is no wonder that the industry went looking for a hedge – and they went online. Regardless of whether they belong to the latest 256-bit generation or to the old 128-bit platform, game consoles are increasingly faced with the challenge of having to compete with networked home computers, which can simultaneously be used as digital television sets, chat interfaces, and telephones (not to mention the convenience of shopping, banking, etc). Nintendo, Sony, and Microsoft have thus devoted more and more of their resources to networking technology, and console games are being touted more and more frequently for their networked gaming modes over WiiWare, the PlayStation Network, or Xbox Live. In its first six months, Activision’s Call of Duty: Modern Warfare 2 was responsible for 1.75 billion minutes of play on Xbox Live, and prompted twenty million transactions for additional game content.11 Xbox Live (available since November of 2002) connects more consoles than WiiWare (available since March of 2008) or the PlayStation Network (available since May of 2006), but the greatest variety of networked games is available on the Wii (the PlayStation Network has the fewest users). Whereas both the PlayStation Network and Xbox Live offer access to Facebook, Twitter, and Netflix, this is not true of the Wii (which only started offering Netflix recently). Game consoles, however, have been losing market share as well, and this is because entertainment software has been changing under the growing influence of web and mobile culture. Many game developers are looking for new markets; their interest is not only on producing casual games for Facebook or smart phones, which are inexpensive to create yet not that highly profitable, but also in developing large-scale online games with thousands or even millions of participants. The first titles of this sort were produced by newcomers like PopCap, Zynga, and Playdom, which tested new strategies for releasing their products, including making them available as iPhone or Android apps and on portals such as Pogo, Yahoo Games, Bebo, or WildTangent. Now, however, American developers are trying to learn from Asian MMORPGs and to lure, from among the players of popular adolescent titles such as Habbo Hotel, more and more subscribers to new online offerings. This, above all, is the great promise of the game market that the new masses are expected to fulfill.

II. Virtual World Economies

Four fifths of the Internet users in the United States work or play in synthetic worlds. The American market for computer games amounted to some nineteen billion dollars in 2007, as mentioned above, and this number is projected to reach thirty billion or more. One of the most noteworthy trends has been the strong growth of MMORPGs and synthetic worlds, which have millions of users in Asia and North America. At peak times, the game World of Warcraft alone has had between ten and twenty million players and earned the Vivendi company more than a billion dollars of annual revenue. Nintendo offers MapleStory on its DS platform; Square Enix’s games include Final Fantasy and Dragon Quest, while Electronic Arts is responsible for Ultima Online and Dark Age of Camelot. These worlds are distinguished by their great variety of simulated economic activities, from simple trades and auctions to opportunities for earning or losing “gold” or other items in the game. Because the division of labor is such a prevalent aspect of role-playing games, it comes as no surprise that this economic structure has led to a number of secondary markets. Everything from virtual swords and magic elixirs to entire characters and player accounts are available for sale on eBay, IGE, PlayerAuctions, ItemMania, and other websites. My first personal encounter with these markets occurred fifteen years ago when one of my neighbors was promoted from programmer to usability manager. One of his first tasks was to figure out how, precisely, his company’s software was being used. One product line was an inexpensive alternative to Photoshop, and as he found out it was being used by numerous hobbyists to design virtual armor, buildings, flags, and animals that could be incorporated into various game worlds. Many game developers count on the creativity of their subscribers and allow or encourage the use of skins, maps, loot, and other “user-generated content” in their online games. As Samuel Weber has commented, a distinction needs to be drawn between those who turn virtual goods into real currency (which itself is somewhat virtual) and those who dream of creating a different life through their consumption.12

Regardless of whether online game worlds are interpreted in anthropological terms as meeting a basic need or in historical terms as a critique of technologized modernity, they are popular because role-playing allows for different styles and motivations to coexist and interact. Players can meet others, explore virtual territory, solve puzzles, conquer, dominate, spy, cast spells, heal, and so on. Avatars enable embodied forms of communication to take place in a common space; they are less expensive than teleconferences or business trips; and they are equipped with faces and bodies chosen by their users, which suggests that, as far as communication is concerned, virtual worlds offer at least as much as the telephone. Interactivity with a steadily growing number of other participants entails that the quantity of entertainment will be translated into quality, and so long as the experience remains enjoyable, it can be assumed that synthetic worlds will continue to expand. There is no guarantee, however, that this will happen. For instance, there has been some controversy about the practice of selling virtual currency for actual money instead of earning such currency by laboring through a given game. When virtual goods (skins, coins, elixirs, property, etc.) can be bought and sold directly instead being acquired within a game, purists regard this not only as a degradation of the playing experience but also as an imbalance that reflects negatively on the game’s design. More than anything else, MMORPGs promise to provide escapist fantasy worlds of dungeons and dragons in which the pressure to buy and sell things, so persistent in everyday life, can temporarily be avoided. Yet this boundary is rapidly beginning to crumble as auctions and other secondary markets continue to skew the balance between the time that a player invests in a game and the abilities that can be acquired by other means.

Most of the efforts to secure the border between virtual worlds and “reality” have proved to be overly expensive and complicated. Moreover, many players go far beyond mere escapism and construct virtual existences for themselves in which virtual income and virtual life go hand in hand. This, in turn, raises the question of whether such income should be taxed and perhaps even the question of whether an “exodus” into virtual worlds is under way.13 As soon as such worlds develop a sophisticated division of labor and become socially differentiated, players will expect certain opportunities to acquire things in exchange for the time that they have invested elsewhere in the game, and fictitious game currencies will come to be exchanged, at a more or less regular rate, with national currencies. Experts have estimated that the value of such currency trading lies somewhere between one and two billion dollars per year. Just how games become economically productive is evident in the sweatshops, in places like China and Mexico, that employ cheap labor to level up virtual characters on behalf of interested customers. Their customer base includes not only casual gamers who lack the patience to cultivate their avatars for months on end, but also for instance documentary film makers who need a particular character for machinima, or marketing companies that seek to make profits in virtual worlds.14 Other customers might include game journalists who want to write about a virtual world without having to spend months of precious time in it; game developers who need to explore the worlds designed by their competitors; or players who occasionally want to acquire additional magic objects, weapons, gold, or avatars for their guilds.

All of this commerce has made synthetic worlds into a social, political, and economic reality that has yet to be fully understood or appreciated. In an interview with the University of Chicago Press, Edward Castronova made the following insightful comparisons:

In virtual worlds – or synthetic worlds, “virtual” having lost much of its meaning – only the icons around which human interactions flow are nonreal. The interactions themselves are as real as any we have outside synthetic worlds. When six soldiers take out a machine-gun nest at Fort Bragg, the machine gun is real and the teamwork is real. When the same six soldiers take out a dragon in a synthetic world, the dragon is not real but the teamwork is. In synthetic worlds, the things we trade may be fantastic, but the process and value of the trade is real.15

Along with other developers and researchers, Richard Bartle has traced such practices back to 1987, when they began to feature in multi-user dungeons and other online forums. As early as 1997, players of Ultima Online were able to purchase virtual goods on eBay, which was a fairly new site at the time. In 2001, Brock Pierce founded the Internet Gaming Exchange, which provided a means for customers to purchase virtual currencies and goods. The serious academic study of virtual economies began with Castronova’s analysis of the macroeconomic activity in EverQuest II, a game that had become available on Sony Online in 1999. According to his calculations, the average hourly wage on the fictional planet of Norrath ($3.42) was competitive with that in Bulgaria; moreover, Norrath’s per capita GNP of $2,226 exceeded that of China and India during the year under investigation.16 In 2003, Linden Lab inaugurated Second Life, a synthetic world that requires all of its players to purchase so-called Linden Dollars with real currency (by credit card or through PayPal). Since then there has hardly been a single MMORPG whose currency and avatars have been treated any differently. Avatars can be auctioned on eBay, while swords, horses, gold, and other virtual objects can likewise be sold for real money to the highest bidder.

If it is indeed true that almost half of all the players in synthetic worlds exchange money and engage in commerce, then it might be beneficial to give this market a closer look and to evaluate its consequences for our understanding of “games” – and for the bottom lines of the companies that operate them. The best observations thus far have come from commentators, such as Castronova and Julian Dibbell, who are participants themselves.17 For 750 Linden Dollars (about $2.25 at the time), Dibbell offered his work as an eBook that could be read by an avatar in Second Life, and he also offered it as a hardcopy for 6,250 Linden Dollars (approximately $18.75). Autographed editions of both versions were also available for auction, and all of this was done under the following motto: “On behalf of Julian Dibbell, ‘Julian Dibbell’ is selling *Play Money* for play money.” Castronova, for his part, found nearly twenty million dollars of commerce on eBay that was related to EverQuest II, and from this he concluded that there must be even more of such economic activity within the game itself. A good deal of discussion has surrounded the idea that “real-money trading” has degraded or distorted the concept of “games,” pointing in part to the influence of limited restrictions on trade, and partly to the very real existence of transaction costs. But players and developers do not have many options at their disposal: because every game has participants who insist on their right to conduct such business, companies have been forced to react in one way or another. There have been three general approaches to the issue of introducing trade and barter into the gaming worlds: prohibition, segmentation, and integration. A comparison of these approaches will allow us to speculate about how things might look if any of the competing companies had chosen to implement an alternative model. In each scenario, social, cultural, and economic values are transferred back and forth between synthetic worlds and the real market in a self-perpetuating loop.

Blizzard has expressed its appreciation for the continuous feedback provided by the large numbers of players of its games, though it has done so with some reluctance. Jeff Kaplan, one of the lead designers of World of Warcraft, pointedly cited the value of internet forums such as Wowhead,18 though he is quick to qualify his assessment: “The community agrees on nothing, but is never wrong; it pretends to agree with one another, but it is not good at problem solving; and it is not good at game design, but is a good tool to identify issues.”19 That said, the company has left the players of World of Warcraft with little doubt about its corporate position; both its End-User License Agreement and its Terms of Use are unambiguously draconian. Real-money trading is taboo, any players engaged in it will be banned from the game, and their WoW accounts will be closed. Linden Labs, however, provides stable and reliable tools not only for designing avatars and their environments but also for all sorts of commerce on Second Life, from selling virtual real estate to exchanging certain services for Linden Dollars. Sony in turn decided to set up two different types of servers for EverQuest, one that prohibits real-money trading entirely (because it undermines certain approaches to the game) and one in which Sony permits the trading style of play, and itself functions as the broker. In the latter case, real-money trading is not only permitted; it is also linked to something called Station Exchange, which allows Sony to take part in the secondary market. Several questions come to mind in light of these three models: Do transaction costs hinder the free flow of social, cultural, and economic capital? What methods are available for players and developers to reduce these costs, and to what extent can the effects of such methods be observed? Answers to these questions are suggested by the qualitative, quantitative, and anecdotal evidence furnished by the existing scholarship devoted to “ludocapitalism,” by interviews with players, and by my own experiences in EverQuest II, World of Warcraft, and Second Life between 2006 and 2011. While pursuing such answers, however, I was quickly confronted by another phenomenon associated with these games, namely with the use of secondary software in synthetic worlds. This phenomenon radicalizes the issues surrounding games and markets even further.

III. End Users

On October 25, 2006, three representatives from Blizzard arrived unannounced at Michael Donnelly’s home office in Phoenix, Arizona: a lawyer, a vice president from corporate parent Vivendi, and a private detective. Donnelly, the owner and founder of MDY Industries, is the creator of WoWGlider, a popular and controversial piece of automation software for World of Warcraft. Although he has maintained that his business in no way violates Blizzard’s rights, the company accused him of infringing upon their contracts, copyrights, trademarks, and of breaching the Digital Millennium Copyright Act as well. According to Blizzard, MDY Industries has detracted from the value of their game, from the nature of MMORPGs, and from the legitimacy of intellectual property itself.20 Donnelly’s appeal, which was (curiously enough) delivered to Vivendi’s offices on the same day, maintained that Blizzard’s representatives had threatened him. Interesting, too, is the fact that Blizzard’s counter-appeal was not filed until February 20, 2007; it demanded that MDY shut down WoWGlider’s website, surrender the source code to Blizzard, and indemnify Blizzard for accrued losses. The speed with which Donnelly filed his suit suggests that he and his lawyers had already been in contact with Blizzard and were ready to go to court. Though somewhat unusual, Blizzard’s four-month delay indicates that the two parties were then undergoing some negotiations. At any rate, there was no doubt that the case would have enormous implications for End-User License Agreements and Terms of Use. According to Blizzard, MDY’s violation of contract law, copyright law, trademarks, and the Digital Millennium Copyright Act led to unfair competition and unjustly earned profits. Equating contracts with intellectual property, the company argued that the experience of World of Warcraft’s legitimate players was under constant attack from impostors attempting to appropriate the game for their own unlawful purposes.

Blizzard’s draconian licensing is yet another indication of how, on the Internet, contracts matter much more than laws.21 The problem in this case is not one of copy protection (WoWGlider functions independently from World of Warcraft’s code), nor is it plausible that Blizzard loses any profits when people let “bots” play the game in their absence (their subscriptions continue to be paid regardless). It might seem as though the main issue of the dispute is whether Blizzard loses any money when players are “banned” from the game, but this is also unlikely. Committed fans of World of Warcraft simply return to the game with a new subscription, and in many cases they even use the same credit card to pay for it. If there are any losses at all, these seem to be suffered by the players themselves, who have to start over with the game; this is another reason, in fact, why it is enticing to use WoWGlider. I found one user of the service who proudly posted a list of all his banned accounts:

~~Glided~~
l-60 Undead (Holy) Priest *BANNED*
l-60 Undead (Frost) Mage *BANNED*
l-60 Undead (Shadow) Priest *BANNED*
l-60 Troll (Fire) Mage *BANNED*
l-60 Troll Shaman *BANNED*
l-42 Undead Rogue (Melee classes aren’t for me so I gave up on this one) *BANNED*
~~Gliding~~
l-46 Troll Hunter *BANNED*
l-49 Tauren Druid *BANNED*

A conservative estimate of how long it typically takes for five avatars to reach level sixty, and another three to reach level forty, suggests that this single player paid something like the following amount of money to Blizzard: $20 × 8 new accounts + a subscription of $15 × 12 months = $340 or $28.33 per month. That is twice the amount of what the average player pays. One study has estimated that it takes fifteen days of playing time to reach level sixty by normal means.22 Even if it is assumed that the use of WoWGlider does not considerably slow down this process, it would take at least four months to accomplish what this player did. Taking into account the interruptions brought about by being banned, one could realistically estimate that the process might require an entire year to complete. It must be supposed, then, that diehard fans represent a considerable source of income for Blizzard, not a loss. I myself know that among my students there were at least twenty-seven players who have used WoWGlider, and each one of them had three accounts (on average). In other words, each of them paid Blizzard an average of $250 per year, which, despite any interruptions, is a third more than the company could have expected from a typical player.

More generally, it is worth asking to what extent the use of WoWGlider even affects the gaming experience of others. Is Blizzard able to demonstrate how many of its customers stopped playing the game because of their distaste for real-money trading or automation software? How many players have actually cancelled their subscriptions for this reason? And do these players object to auctioning off their avatars to players who, in large part, have no problems at all with such transactions or with WoWGlider? The main reasons why people stop playing World of Warcraft probably involve time management, a loss of interest in the contents of the game, and problems with the guilds. So why did Blizzard really want to sue MDY? In short, it came down to Blizzard’s unwillingness to yield any influence over the nature of their game. The new style of play, supported by secondary markets and software, would lie outside of Blizzard’s authority. This was problematic for the company because as always the main concern of game design is with maintaining a balance between rules that govern play and the leeway a player has while playing. Designing games is a matter of both limiting risks and allowing players to discover whatever risks remain. The involvement of secondary markets and software causes a game to shift fundamentally; Blizzard’s concern was less about the content of World of Warcraft than it was about the playing conditions of the game. What the company’s stance has entailed for World of Warcraft, however, is an unstable economy, doubts about the value of the game’s concept, high regulatory costs, reduced liquidity, and an authoritarian approach to the game’s management. Interestingly, Blizzard has been more tolerant in Europe, where the price for gold on the World of Warcraft servers is only an eighth of what it is in the United States.23

Sony’s introduction of server-specific auctions through Station Exchange has provided players of EverQuest II with a means of buying and selling items as they please on particular servers, whereas the company’s other game servers do not permit such activity. This solution has both reduced Sony’s customer-service costs and brought in new fees. What this has meant for EverQuest II is a stable economy that ensures lower costs for players and steady profits for Sony Online. In 2007, the first year that this model was implemented, in-game trading of $1.87 million produced $274,083 in fees and commissions for Sony. At the same time, the segmentation of the game community reduced Sony’s administrative costs by about thirty percent.24 If we assume that this demand for real-money trading is proportional to that in World of Warcraft, then it is clear that, in 2007, the approximately 8.5 million players of that game could have been responsible for roughly eight million dollars in WoW trade volume. For Blizzard, this could have amounted to around $1.3 million in commissions. Moreover, the legendarily high operating costs of the game could have been lowered as well. Blizzard employs approximately 1,300 “game masters” to monitor the internal activity of the game (this is one my students’ favorite part-time jobs). By applying Sony’s model, Blizzard could have cut nearly four hundred of these employees (at a wage of ten dollars per hour), a move that might have saved the company around eight million dollars that year.

For Second Life, from the very beginning Linden Labs established a stable economy with clear property rights. Everything created in the game belongs to the player who created it, while Linden Labs control only the tools and manages the game. Given that Linden Dollars are necessary for playing the game, the model is specifically designed for exchanging currency (the exchange rate fluctuates, but only minimally). The result for Linden Labs is low operating costs, productive play, and a profitable market. When I asked Maty Roberts, general counsel for Linden Lab, about Blizzard’s experiences, he sidestepped the question, stating only that game developers usually do not share such information with one another.25

The spectrum of reactions to the productive or unproductive frictions between the planned economies of online games and the actual market economy of the software industry indicates that, even in the case of computer games, the masses have become a commodity. If the growth of the American computer-game market has by now come to be dominated less by packaged goods than by subscriptions and secondary transactions, this means that the focus of the industry has shifted from goods, services, and demographic niches to the much-anticipated “new masses.” In this regard it should be noted that the scene in Asia is already far ahead of that in the United States. There, a much larger portion of players is engaged with online games, and at least a fifth of all gamers deal with online goods, services, and currencies.26 On the one hand, then, there is the rational expectation that the economies of computer games will find a natural equilibrium between rigorous central planning and playful self-regulation, and on the other hand there is the assumption of reflexive behaviorists that inherently unstable game worlds will lose cohesion as more and more people participate in them. It is only in the tension between these two positions that the new mass media have been able to commodify the masses.