+ CONNECT SERIES 63739 KNOWLEDGE MAP OF THE VIRTUAL ECONOMY CONVERTING THE VIRTUAL ECONOMY INTO DEVELOPMENT POTENTIAL www.infoDev.org CONVERTING THE VIRTUAL ECONOMY INTO DEVELOPMENT POTENTIAL KNOWLEDGE MAP OF THE VIRTUAL ECONOMY AN info Dev PUBLICATION WRITTEN BY: Dr. Vili Lehdonvirta & Dr. Mirko Ernkvist April 2011 Information for Development Program www.infoDev.org ©2011 The International Bank for Reconstruction and Development/The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved The �ndings, interpretations and conclusions expressed herein are entirely those of the author(s) and do not necessarily reflect the view of infoDev, the Donors of infoDev, the International Bank for Reconstruction and Development/The World Bank and its af�liated organizations, the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to infoDev Communications & Publications Department; 2121 Pennsylvania Avenue, NW; Mailstop F 5P-503, Washington, D.C. 20433, USA; telephone: 202-458-4070; Internet: www.infodev.org; Email: info@infodev.org. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Cover design by infoDev Typesetting by The Word Express, Inc. To cite this publication: Lehdonvirta, Vili. & Ernkvist, Mirko, 2011. Converting the Virtual Economy into Development Potential: Knowledge Map of the Virtual Economy, Washington, DC; infoDev / World Bank. Available at http://www.infodev.org/publications Table of Contents Executive Summary xi 1. Introduction 1 1.1 Background 1 1.2 Structure and scope of this report 1 1.3 Methodology 2 1.3.1 Value chain analysis 2 1.3.2 Expert interviews 2 1.3.3 Market size estimates 2 2. Defining and Segmenting the Virtual Economy 5 2.1 From digital abundance to virtual scarcity 5 2.2 Key characteristics and differences from traditional digital content industries 6 2.3 Segmenting the virtual economy 7 3. Third-party Online Gaming Services 9 3.1 Demand and supply 9 3.2 Market size 10 3.2.1 Online game market size 10 3.2.2 Earlier estimates of third-party gaming service market size 11 3.2.3 Estimating the gaming service market through player surveys 12 3.3 Business models 14 3.3.1 Retailers 15 3.3.2 Producers 16 3.3.3 Game operators 17 3.3.4 Supporting activities 17 3.4 Regulatory framework and industrial policy 17 3.4.1 Negative externalities from trade of artificially scarce assets 17 3.4.2 Contractual and legal regulation of virtual goods trade 18 Table of Contents iii 3.5 Case study: Purchasing virtual currency for World of Warcraft 20 4. Microwork 23 4.1 Demand and supply 23 4.2 Market size 24 4.3 Business models 25 4.3.1 Transformers 25 4.3.2 Work aggregators 26 4.3.3 Infrastructure providers 27 4.4 Regulatory framework and industrial policy 27 4.5 Case study: Using human workers to optimize an online retail search engine 28 5. Other Segments of the Virtual Economy 31 5.1 Cherry blossoming 31 5.1.1 Case study: Improving the visibility of an online store 31 5.2 User-created virtual goods 32 5.3 Other two-sided marketplaces 33 6. Development Potential of the Virtual Economy 35 6.1 Third-party online gaming services 35 6.1.1 Worker demographics, skills, wages, and career development 35 6.1.2 Distribution of income 40 6.1.3 Costs and profitability 41 6.1.4 Number of people employed 41 6.1.5 Competition, entry barriers, and changing business conditions 42 6.1.6 Changes in business and market’s conditions over time 43 6.1.7 Upgrading strategies 44 6.2 Microwork 44 6.2.1 Worker demographics, skills, wages, and career development 44 6.2.2 Distribution of income 46 6.2.3 Competition, entry barriers, and changing business conditions 47 6.2.4 Upgrading strategies 47 6.3 Other segments 48 6.3.1 Cherry blossoming 48 6.3.2 Microcontent production and other two-sided marketplaces 48 iv Knowledge Map of the Virtual Economy 7. Conclusions 49 7.1 Market opportunities 49 7.2 Development potential 50 7.3 Scope for interventions 51 7.3.1 Third-party gaming services 51 7.3.2 Microwork 51 References 53 Appendix 1. Sources of uncertainty in the gaming services market estimate 57 Appendix 2. Examples of third-party websites selling virtual game currency 59 LIST OF TABLES Table 1. Segmenting the virtual economy 8 Table 2: Opportunities for third-party gaming services in different types of games 10 Table 3: The global online game market 11 Table 4: Percentage of players buying from the secondary market 13 Table 5: Average amount spent on the secondary market per year 14 Table 6: Number of paying online game players 15 Table 7: Global secondary market size 15 Table 8. Differences between crowdsourcing and microwork 24 Table 9. Number of participating workers and average hourly payout in four assignments during September 2010 30 Table 10: Geographic location of surveyed gaming studios 36 Table 11: Size and gender diversity of gaming studios 36 Table 12: Gaming studio workers’ prior occupations 37 Table 13: Average weekly working hours of gaming studio workers 39 Table 14: Hourly wages of gaming studio workers 39 Table 15. Monthly operating costs of a medium-sized gaming studio in suburban China 41 Table 16: Operational cost structure of surveyed gaming studios 42 Table 17: Estimate of spending on wages in the third- party gaming services industry in 2009 42 Table 18. Estimate of the number of game laborers employed in 2009 43 Table 19: Upgrading strategies for gaming studios 45 LIST OF FIGURES Figure 1. Three-layer model of ICTs and economy 6 Figure 2. Basic value chain in the gaming services industry 14 Figure 3: An order of World of Warcraft gold for $100 21 Figure 4. Basic value chain in the microwork industry 25 Figure 5. Distribution of CrowdFlower’s search labeling work by country in September 2010 29 Figure 6. Revenue shares in the Chinese powerleveling industry 40 Figure 7. Revenue shares in the Chinese gold farming industry 40 Figure 8. Typical revenue shares in the microwork industry 47 vi Knowledge Map of the Virtual Economy Acknowledgements The Knowledge Map of the Virtual Economy was Ville Miettinen, Founder and CEO, Microtask Ltd, prepared by Vili Lehdonvirta and Mirko Ernkvist Finland under the supervision of Tim Kelly and Lara Miho Nojima, Associate Professor of Economics, Srivastava at infoDev, a donor-funded agency of the Seikei University, Japan World Bank. Jared Psigoda and To Minh Thu Jared Psigoda, Founder and CEO, Niubility Co., provided vital research assistance. The report was Ltd, China made possible by funding from the UK Department Andrew Schneider, Co-founder and President, Live for International Development (DFID). Gamer, Inc., United States Unggi Yoon, Judge and Internet law scholar, Contributions from the following expert informants Republic of Korea are gratefully acknowledged: The report also benefited from comments expressed Lukas Biewald, Founder and CEO, CrowdFlower, by the participants of the infoDev Virtual Economy Inc., United States workshop held on December 13, 2010 as part of the Julian Dibbell, journalist, author, France ICTD 2010 (International Conference on Anthony Gilmore, Co-founder, Nameless Media & Information and Communication Technologies) Productions, Japan conference at Royal Holloway, University of Eyjólfur Guðmundsson, Lead Economist, CCP London. Thanks are also due to the peer reviewers Games, Iceland who helped develop the initial terms of reference Vaughn Hester, Data Analyst, CrowdFlower, Inc., and have guided the study: Robert Hawkins, Anat United States Lewin, and Carlo Rossotto of the ICT Sector Unit Jun Sok Huhh, Lecturer in Economics, Seoul of the World Bank, and Victor Mayer-Schöenberger National University, Republic of Korea of the Oxford Internet Institute (OII). The authors Leila Chirayath Janah, Founder and CEO, are grateful to CrowdFlower, Inc. and Vaughn Samasource, United States Hester for sharing data used in the microwork case Nicolas Leymonerie, Chairman, IGDA Vietnam, study. Vietnam Preface vii viii Knowledge Map of the Virtual Economy About the Authors Vili Lehdonvirta is a Researcher at the Helsinki infoDev is a global development financing program Institute for Information Technology, a joint coordinated by an expert Secretariat housed in the research institution of Aalto University and the Vice Presidency for Financial and Private Sector University of Helsinki, Finland. During 2010–2011, Development (FPD) of the World Bank. It helps he was a Visiting Scholar at the Interfaculty developing countries and their international partners Initiative in Information Studies, University of use innovation and ICTs as tools for poverty Tokyo. Dr. Lehdonvirta holds a PhD in Economic reduction and sustainable social and economic Sociology from Turku School of Economics and an development. infoDev acts as a neutral convener of MSc (Tech) from the Helsinki University of dialogue, and as a coordinator of joint action among Technology. He has published over a dozen peer- bilateral and multilateral donors—supporting global reviewed research papers on virtual goods, curren- sharing of information on ICT for development cies, and economies. (ICT4D), and helping to reduce duplication of efforts and investments. It helps developing coun- Mirko Ernkvist is a Researcher in the Department of tries and their international partners maximize the Economic History at the University of Gothenburg, contribution and impact of the private sector Sweden. During 2010–2011, he was a Visiting through direct support for ICT-enabled innovation, Scholar at the Interfaculty Initiative in Information networking between entrepreneurs, investors and Studies, University of Tokyo. Dr. Ernkvist holds a donors, and by sponsoring cutting-edge research. PhD in Economic History from the University of Gothenburg. His research is concerned with Note: All dollar amounts are in US dollars unless entrepreneurship, development, and innovation otherwise indicated. strategies in creative industries. x Knowledge Map of the Virtual Economy Executive Summary The widespread adoption of information and Development impact communication technologies (ICTs) in everyday life and commerce has given rise to new digital prob- The gross revenues of the third-party gaming services lems and challenges. Although information provided industry were approximately $3.0 billion in 2009, by networks is abundant, the human attention most of which was captured in the developing required to process it is limited. And although countries where these services were produced. In digital resources in principle are unlimited, many comparison, the global coffee market, on which online platforms have artificial scarcities built into many developing countries are highly dependent, them as part of their design. The demand for these was worth over $70 billion—but only $5.5 billion scarce resources, their supply, and the markets where was captured by the developing countries that the supply and the demand meet, constitute a produced the coffee beans. This suggests that the computer-mediated virtual economy. virtual economy can have a significant impact on local economies despite its modest size. It can also During the past decade, developing countries have support the organic development of local ICT rapidly followed developed countries in ICT infrastructure by providing revenue models that adoption, and in some areas, such as mobile maintain existing deployments and justify new payments, even surpassed them. However, develop- private investments. ing countries’ roles in the digital economy mostly have been limited to being users and consumers of The third-party gaming services industry is one technology. Creating new digital services and example of a virtual economic activity that creates technologies for the global market requires ad- value for the customer by overcoming artificial vanced skills and infrastructure, and continues to be scarcities. Another example is a marketing agency out of reach for most entrepreneurs in developing that pays users to inflate the number of fans a countries. particular brand has on a social networking system, such as Facebook, in order to boost that brand’s In contrast, entrepreneurial opportunities in the visibility in searches. A problem with services that virtual economy of digital scarcities have sometimes create value by overcoming artificial scarcities is that been successfully exploited, even with limited skills they degrade the value of the platform for other and infrastructure. A prominent example of this is users. As a result, their net social value can some- the third-party gaming services industry. An esti- times be negative. mated 100,000 young, low-skilled workers in countries such as China and Vietnam earn their A different set of virtual economic activities creates primary income by harvesting virtual resources and unambiguously positive value by helping customers providing player-for-hire services in popular online overcome natural scarcities that linger in the digital games such as World of Warcraft. The demand for world. A prominent example of this is the microwork these services comes from millions of wealthier industry. Thousands of men and women in countries players who have a serious interest in the game such as India and Kenya earn primary and supple- world and the social connections it facilitates, but mentary income by completing simple human lack the time (and patience) to reach far into the intelligence tasks, such as assessing whether two game alone. images are of the same product, or transcribing a fragment of handwritten text. The demand for these with higher value-added. The gaming services services comes from companies, such as Amazon, industry successfully underwent such a transition that rely extensively on digital technology in their during the last decade. All aspects of the third-party everyday business, yet find that computers are not gaming services value chain—from production to yet able to do all of the tasks required of them. marketing—can now apparently be carried out from countries such as China and the Philippines. As a In contrast to the third-party gaming services result, the industry may employ as many as tens of industry, the microwork industry is still a new and thousands of skilled customer service and manage- emerging phenomenon. Although microwork itself ment staff in these locations. requires no technological expertise, converting computational and business problems into micro- Planned donor or NGO-led interventions could tasks and thus making them addressable by micro- enhance the development impact of the virtual workers is a significant technological challenge. A economy. Interventions should focus on segments handful of start-up companies are currently working based on natural instead of artificial scarcities in on expanding the range of problems that can be order to ensure that their net social contribution is addressed by microworkers. The potential market positive. The most prominent current example is the size is difficult to estimate, but could reach billions microwork industry. It is recommended that of dollars per year in the next five years, as the development interventions focus on three aspects of technology matures. Others are working on easier microwork: enabling demand for microwork, user interfaces and more efficient distribution building capacities for the supply of microwork in methods to allow microwork to be offered to some developing countries, and helping producers in of the poorest and least educated workers in developing countries to upgrade their businesses to developing countries. increase the value generated. The latter two objec- tives could be addressed with the development of software tools that facilitate the conduct of micro- work. To be successful in least-developed countries, Scope for interventions such tools would have to be primarily mobile-based applications. Like traditional labor-intensive industries, activities in the virtual economy are organized into value Besides microwork, development interventions chains that include functions such as marketing and could help promote the development of new digital quality control besides the manual work itself. The networks and services that have potential to provide manual work is typically carried out in developing jobs in the virtual economy in the future. In the countries, while the customer-facing functions are same way that access to high speed Internet back- generally located in developed countries. In the long bone connections helped India develop its business run, wage competition is likely to limit income process outsourcing (BPO) industry in the 2000s, so development from the manual work. For a sustained the development of mobile broadband networks development impact, it may be necessary for (so-called 3G and 4G networks) could assist a wider microwork entrepreneurs in developing countries to range of developing countries to create jobs and find ways to upgrade their positions in the value generate wealth from the new opportunities that the chain, and to offer a more diverse range of services, virtual economy brings. xii Knowledge Map of the Virtual Economy Chapter 1 Introduction 1.1 Background industry, which is distinguished in this report as a separate concept from the more established notion of crowdsourcing, relevant research has simply not The widespread adoption of ICTs in everyday life been carried out yet due to the novelty of the has given rise to a massive new market for digital industry. In these areas, it was necessary to comple- goods and services. Addressing the business op- ment the report with primary research. The resulting portunities in this market has traditionally required novel data and analyses should prove useful in significant skills and infrastructure, putting them understanding the development potential of the out of reach for most people in the developing virtual economy. world. However, new marketplaces and value chains have emerged that provide digital earning opportu- nities for semi-skilled and unskilled workers with access to relatively basic digital infrastructure. Thousands of students and migrant workers make a 1.2 Structure and scope of this living in China by playing online games and selling report the resulting virtual assets to wealthier players (Heeks 2009). A growing number of crowdsourcing The report is structured as follows. The next section and “microwork� platforms employ unskilled introduces the theoretical notion of a “virtual workers in digital tasks ranging from pattern economy� and explains how it is distinct from other recognition to data input. This “virtual economy� of ICT-related economic activities. The following digital goods and services, and its potential for sections describe in detail the main areas of the economic and social development, are the subject of virtual economy, their economic impact, business this report. models and value chains. The two major areas of the existing virtual economy are identified as 1) third- This report is the first phase in an infoDev project to party gaming services and 2) microwork. This report identify and exploit development opportunities in will focus largely on these two distinct but concep- the virtual economy. It is a “knowledge mapping� tually related areas. Gaming services is an established exercise to understand “what we know and do not industry that provides a rich set of evidence for know� in a particular field. It draws on existing analysis, while microwork is an emerging industry literature and expert opinions to provide a birds-eye with apparently significant development potential. view that can be used as the basis for further Other existing activities within the virtual economy research and possible interventions. The report also are categorized as 3) marketing related paid-for benefited from discussions at the infoDev Virtual connections in social media (“cherry blossoming�), Economy workshop held as part of the ICTD 2010 and 4) user-created virtual goods in virtual environ- conference at Royal Holloway, University of ments. These are not covered in detail due to their London, on 13 December 2010.1 limited development potential, at least at present. Existing literature is critically lacking in some of the The sixth section analyzes the development potential key interest areas of this report. Western accounts of of the virtual economy. Development potential is Chinese “game laborers�, which are understood to here understood as the ability to provide income to be part of a multi-million dollar industry, are mostly local economies through employment and based on a handful of journalistic accounts (Nardi & Kow 2010). The same data is echoed from one 1 See “Finding development potential in the scarcity of the virtual publication to another. In the case of the microwork economy� at: http://www.infodev.org/en/Article.659.html. Introduction to the Knowledge Map 1 entrepreneurial opportunities. Both short-run recognizes the different bargaining power of opportunities and long-run income development are different actors in the transaction environment, and considered. Development potential also includes the provides a framework for understanding the ability to support the development of local ICT competitive and commercial viability of different infrastructure. In the final section, the report actors in the industry. This way, it acts as a powerful summarizes the key findings, identifies important framework for development programs and efforts to gaps in current knowledge, and sketches out the create entrepreneurial opportunities for poor people, scope for possible donor or NGO-led interventions and enables empirically grounded assessments of the towards maximizing the development potential of different barriers and challenges in the industry. the virtual economy. Secondly, it focuses on the concept of value added or captured, as opposed to simple gross revenues. The geographical scope of the report is, in principle, This way, it makes it possible to assess the develop- global, but in practice the majority of empirical ment impact of the industry separately at each evidence concerning the gaming services industry is country and locality involved, including at the from China, due to the dominance of actors from bottom of the economic pyramid. that country. Some care must thus be taken in drawing conclusions for less developed countries. In this report, we analyze the value chains of the 1.3.2 Expert interviews current microwork industry that span from Western The experts interviewed for this study are listed in countries to developing countries in Asia, Africa, the Acknowledgements. They consist of corporate and South America. managers involved with various areas of the virtual economy as well as scholars and journalists who have investigated virtual economic phenomena first-hand. An initial set of key experts was identified 1.3 Methodology from literature and a second set through referrals from the first set. Despite efforts to contact relevant The main analytical framework used in this report is experts in the developing world, the majority of the a variation of value chain analysis used in develop- informants represent developed country actors, ment studies, as described below. The main sources although with first-hand knowledge of activities in of data were the existing literature (cited where the developing world. The interviews were semi- appropriate) and expert interviews (described structured and focused on each informant’s areas of below). New primary data was also gathered expertise. The results are reported in a consolidated through a survey of Chinese gaming services form that integrates informants’ opinions with producers and from the corporate database of analysis without distinguishing between individual CrowdFlower, one of the companies that is active in informants, except when there is a particular reason. the microwork industry. These will be described This mode of presentation was necessary for later in their respective sections. conciseness. 1.3.1 Value chain analysis 1.3.3 Market size estimates A value chain analysis describes the activities that A major outcome of the knowledge mapping bring a product through the different phases of exercise is an assessment of the current market size production, distribution, and marketing to the final and future market potential of various branches of consumer. It also involves describing possible ways the virtual economy. There are two basic approaches to capture more value though different forms of to assessing the size of an industry or market: upgrading strategies. Intra-firm value chain analysis supplier side and buyer side. The former involves was first introduced by Porter (1985). The approach obtaining, aggregating, and extrapolating revenue was subsequently expanded also to supply chains. figures from suppliers operating in the market. The latter approach involves estimating the total There are some characteristic features of value chain consumption of the goods provided in the market. analysis that make it suitable for development If the buyer population can be defined and sampled studies (Kaplinsky & Morris 2001). First, it sufficiently, accurate estimation can be accomplished 2 Knowledge Map of the Virtual Economy with a survey study. In industries with complex 2. Calculating a new estimate using the buyer side value chains, or where suppliers and buyers are approach. Results from a number of credible difficult to define accurately, revenue figures from survey studies by the Korea Creative Content intermediaries or from complementary products Agency, the China Internet Network may be used as proxies for the purposes of estimat- Information Center, the International Data ing market size and trends. Corporation and others are used as the basis. By comparing the data from the developed Korean Economic activity in the virtual economy is highly market with that from the developing Chinese distributed. There are numerous suppliers and market, the report also attempts to account for consumers, and almost no systematically collected the difference in gaming services spending in data exists especially on the supply side. For developed versus developing countries. This has estimating the economic impact of the third-party become increasingly important following the gaming services industry, this report uses the rapid growth of domestic online game markets following methods: and the associated gaming services markets in developing countries. 1. Reviewing previous estimates and guesstimates. The report collects and critically assesses previous In the microwork area, this report presents analysts’ estimates from literature and informants. Some revenue estimates from two related markets, BPO industry informants are in an intermediary role and paid crowdsourcing, and assesses future market and thus able to estimate at least one fragment of potential based on this and other evidence. the market with some accuracy. Introduction 3 4 Knowledge Map of the Virtual Economy Chapter 2 De�ning and Segmenting the Virtual Economy 2.1 From digital abundance to Journalistic accounts suggest that thousands of people in countries such as China and Vietnam earn virtual scarcity their primary income this way (Heeks 2008). A growing number of paid crowdsourcing and The economic impact of the pervasive adoption of “microwork� platforms provide primary and ICTs in developed countries since mid-1990s can be supplementary income to unskilled workers in tasks characterized in terms of an eradication of scarcities ranging from pattern recognition to data input. (Shapiro & Varian 1999). Digital technology made it possible to duplicate and transmit various What is common to these new earning opportuni- information goods at near-zero marginal costs, ties is that they are characterized by the discovery eradicating scarcities in media and entertainment and development of digital scarcities that can be distribution. Computers and digital communication exploited without advanced skills. In the case of channels made it possible to automate common virtual goods, scarcity is artificially created and office tasks such as distributing memos, thus maintained by the publishers of online games and eradicating scarcities in clerical work. The resulting social networking sites for the purpose of making abundance has been a boon for many. Companies the goods desirable. Game laborers toil to harvest enjoy new efficiencies and better access to markets. these goods and sell them on others. In the case of Consumers enjoy unprecedented access to a massive microwork, natural scarcity remains in the supply of wealth of information and entertainment, and even labor for clerical work that could not be automated digitally connected consumers in developing because of the limitations of computing technology. countries benefit greatly. Companies in the microwork industry have invented ways to use technology to make this work At the same time, industries and individuals whose addressable by unskilled workers all over the world. economic contribution was based on overcoming These and other emerging digital scarcities that the old scarcities, such as music distributors, require time, effort, and comparatively few special- newspapers, and low-skilled white-collar workers, ized skills and resources to exploit are referred to in have seen their earning opportunities diminish. This this report as the “virtual economy�. “digital economy� mostly has created new earning opportunities for workers and entrepreneurs with Edward Castronova (2006a) first used the phrase advanced professional skills in disciplines such as virtual economy to refer to artificial economies technology and marketing (Florida 2002). They inside online games, especially when the artificially have been able to create and capture new scarcities scarce goods and currencies of those economies were in the economy. Unskilled workers and developing traded for real money. The phrase was subsequently countries have largely not been able to find business adopted in this meaning among game scholars and opportunities in the digital economy. in the game industry. As virtual currencies have started to be used in online services other than However, as the digital economy has grown, new games, such as social networking sites and crowd- services and platforms have started to give rise to a sourcing platforms, the term virtual economy has new set of economic opportunities that seems to started to see wider application. This report’s contradict the previous history. Individuals with no definition of the virtual economy builds on this formal training or qualifications are able to harvest meaning and further widens it by recognizing that virtual goods and currencies in online games and sell not only are virtual goods and currencies scare and their holdings to other players for real money. tradable within digital marketplaces, but so are many other intangible commodities, such as human effort. Figure 1. Three-layer model of ICTs and economy The relationship between the physical ICT infra- structure, the digital economy of services supported by the infrastructure, and the virtual economy that Virtual Economy labor emerges from the digital services, can be depicted as a three-layer model (Figure 1). Existing studies on ICT and development focus on the two bottom Digital Economy layers: how ICT infrastructure is produced and maintained in developing countries, how hardware manufacturing creates jobs, and how digital services ICT insfrastructure can be used to enhance productivity in sectors such as agriculture and trade (e.g., UNCTAD 2010). At the same time, the proliferation of digital services from e-commerce to social networking services in Source : Author’s elaboration developed as well as developing countries has given rise to new digital needs and problems. This demand, the supply that has arisen to meet it, and the markets where this demand and supply meet, may require substantial effort and investment, but together comprise the virtual economy. once that is done, the cost of creating additional copies by duplicating the original is negligible. The second difference is that from a consumer’s point of view, information goods are “experience goods�: that 2.2 Key characteristics and is, their value is derived from experiencing them and differences from traditional absorbing their content. digital content industries In contrast, the commodities of the virtual economy, also known as virtual goods, are similar to ordinary The following characteristics are typical of the goods. Their production can involve significant virtual economy: marginal costs. These costs may be due to natural scarcities, as is the case with microwork, where every ■ Centers around commodities that are digital yet individual task must be handcrafted. Although bits, scarce the “raw material� of these goods, are abundant, the ■ Demand arises from the increasing use of digital supply of human effort is scarce and imposes a services in business and leisure marginal cost of production. In other cases, significant ■ Supply is created through the expenditure of marginal costs of production arise from artificial human effort, and doing so requires relatively hurdles placed in the way of would-be producers by few specialized skills or resources the designers of the platforms. These hurdles may be necessary to safeguard the value of the goods. In the The virtual economy can be contrasted with the case of online games, virtual items and accessories that traditional digital content industries that produce in principle could be duplicated at no cost are made content for the digital economy. Traditional content unique and meaningful by requiring that significant includes such things as music, video, images, news effort be expended in order to obtain them. articles, and any other goods that can be represented in digital form. Economists refer to such goods as The value that consumers obtain from virtual goods information goods, because they differ from most is consummated in a range of ways (Lehdonvirta ordinary goods in two important ways (Shapiro & 2009a). Some aesthetically pleasing virtual goods Varian 1999). The first difference is that, from a might be consumed like information goods, by producer’s point of view, information goods involve experiencing them (Denegri-Knott & Molesworth high fixed costs but low marginal costs of produc- 2010). But a more important reason why consumers tion. Creating the first copy of an information good buy virtual goods is that the goods are built so as to 6 Knowledge Map of the Virtual Economy have tangible uses and functions in the games and 2.3 Segmenting the virtual digital environments where many people play out parts of their social lives today. Virtual goods are also economy used to signal social distinctions and bonds in the same way as material consumption commodities Commercially significant activities in today’s virtual (Martin 2008; Lehdonvirta 2009b; Lehdonvirta, economy can be roughly categorized into four Wilska & Johnson 2009). Thanks to artificial segments: scarcity, virtual goods are able to distinguish haves from have-nots in the digital environment—some- ■ Third-party online gaming services thing that information goods that can be infinitely ■ Microwork copied are not good at. In this sense, digital ■ “Cherry blossoming� consumers are often no less materialistic than ■ User-created virtual goods production material consumers: the only difference is that their material has become digital (Lehdonvirta 2010). In The third-party online gaming services segment business use, the value of virtual commodities such consists mainly of activities known as “gold farming� as microtasks is likewise functional rather than and “powerleveling�. Both are essentially services informational: they are cogs in a large machine. where an online game player hires someone else to play the game on their behalf. They do this in order Because of these differences, the value chains and to obtain the virtual rewards of the play without markets of the virtual economy are in principle having to spend the time and effort. In contrast, the fundamentally different from those of the traditional microwork segment consists mainly of services digital content industries. Traditional content catering to business clients. It involves breaking industry employs a small number of highly skilled insurmountable computational problems into producers, while the suppliers in the virtual simple human intelligence tasks or “microtasks� that economy use a large number of less skilled workers. can be distributed to and addressed by human Traditional digital content loses its value fast as its workers. novelty wears out, while virtual goods can be more valuable years after their creation than they were “Cherry blossoming� is a term used in this report to initially. It should be noted that the companies that refer to small marketing related digital tasks, such as produce the platforms on top of which virtual “liking� a brand’s Facebook page against a small pay. economies operate, such as online games and digital It resembles microwork in that it involves recruiting work exchanges, are themselves usually part of the large numbers of workers to complete small tasks for traditional content industries. a business client. However, unlike microwork, the tasks involve overcoming artificial scarcities created In practice, the distinction between the virtual by the designers of the platforms. In this aspect, economy and traditional content industries is not cherry blossoming is comparable to third-party always as clear. Digital music and film distributors online gaming services. The user-created virtual use digital rights/restrictions management technolo- goods segment consists of activities for producing gies (DRM) to impose artificial scarcity on media and selling user-generated virtual items, textures and files, which brings them conceptually closer to other artificially scarce virtual objects for virtual virtual goods (Lehdonvirta & Virtanen 2010). environments such as Second Life and Instant Online retailers adopt virtual currency based loyalty Messaging Virtual Universe (IMVU). Although the programs. Crowdsourcing-based content production resulting goods are artificially scarce to the buyers, models blur the boundaries between traditional web the real scarcity overcome by this activity is the content production and microwork. Many objects effort required to design the goods. of value may in the future find expression in scarce digital form, and be sourced and exchanged through The four segments of the virtual economy, their lightweight online interactions rather than through target groups, and scarcities are depicted in Table 1. the more rigid structures of the formal economy. In the following sections, the segments are analyzed However, this report focuses on today’s commer- in detail, focusing especially on the segments with cially significant areas that are distinct from significant development potential: third-party online traditional content production activities. gaming services and microwork. De�ning and Segmenting the Virtual Economy 7 TABLE 1. Segmenting the virtual economy Arti�cial scarcity Natural scarcity Consumer oriented Third-party online gaming User-generated virtual goods production services Business oriented Cherry blossoming Microwork Source: Author’s elaboration 8 Knowledge Map of the Virtual Economy Chapter 3 Third-party Online Gaming Services 3.1 Demand and supply measured in U.S. dollars or Korean Won could soon be observed for virtual goods ranging from charac- ters to gold nuggets (Lehdonvirta 2008). A major Online games have become a hugely popular form of object such as a castle could easily be worth hun- entertainment and social interaction. Hundreds of dreds of dollars. The biggest player-to-player trade millions of people around the world regularly play reported in the media was the 2007 sale of a online games. Among some players, there is a latent character in the online game World of Warcraft for demand to purchase advances in online games for real approximately $9,500 (Jimenez 2007). money. For example, in so-called massively-multi- player online games (MMOGs), players repeatedly Today’s surveys suggest that approximately one in kill hundreds of monsters in order to develop their four MMOG players engage in real-money purchases characters and obtain rare objects. This activity takes on a yearly basis (see the following section). The place in the context of a community of players, who trade no longer takes place between normal players: compete for, collaborate with, and compare each commercial suppliers have entered the market. In others’ virtual possessions. As a result, virtual goods this report, they are referred to as the third-party in the game obtain a social status value in the same gaming services industry, as they provide “gaming way as consumer goods do in physical environments. services� for a fee. They are third-party, because they Some players would rather buy those objects to enjoy are not affiliated with the game publishers. The their benefits than spend time and effort to obtain goods and services that they offer to players fall into them through their own play. Some rare objects may two broad categories (Gilmore 2009): not even be obtainable through gameplay any longer. This gave rise to a play-to-player “secondary market� ■ Virtual goods and currencies. Instead of where virtual game assets are traded for real money. spending time and effort to earn game currency themselves, players can purchase the currency Virtual currencies, items, and characters were first from the gaming services industry, which spends traded for real money in the early online games of the effort for them. the 1980s (Hunter 2006). The practice became ■ Powerleveling. This is a “player-for-hire� service widespread in the MMOGs launched in the late where a professional player takes control over a 1990s, such as Ultima Online, EverQuest, and normal players’ character for a few hours, days, Lineage (Castronova 2005; Huhh 2008). In these or even weeks, in order to build up the charac- games, normal gameplay involved hundreds of teristics of the character. Powerlevelers also sell thousands of players trading game items, accumu- “ready-made� characters. lated during months of play, for other game items. The designers intended the games to be like One industry informant suggests that virtual Monopoly: no real money would change hands. But currency sales account for more than three quarters around 1999, some players began to put their game of the market. Powerleveling and ready-made goods on auction at ecommerce sites such as eBay. characters account for most of the rest. Perhaps surprisingly, they soon received bids from other players. When an auction was completed, In most games, the publisher of the game does not payment was carried out using ordinary means, such endorse the secondary market for virtual goods and as check or money order. The two players then met services, and trading takes place on third-party up in the game and the seller handed the auctioned marketplaces. There are also some secondary markets object to the buyer. This way, an exchange value that are sanctioned by the publisher. In 2005, Sony Online Entertainment, one of the biggest Western 3.2 Market size online game publishers, launched a marketplace where game assets belonging to certain of its games The third-party gaming services market has to be can be traded for real money against a transaction understood in relation to the global market for fee (Robischon 2007). Other game operators have online games. This section first gives an overview of generally not followed Sony’s example. Live Gamer, the growth of online game market. The official a company that provides virtual commerce plat- online game market numbers presented in online forms for game publishers, operates Sony’s market- game market size estimate section do not take into places today. account the market for third-party gaming services served by gaming studios. This is followed by an More recently, many game publishers have begun to analysis of the size of the gaming studio market. The respond to the demand for virtual goods by selling analysis includes an overview of earlier estimates the goods to players themselves (known as the “item from the literature as well as a new estimate, payment� or “free-to-play� model). On one hand, presented in this report, that uses a new estimation this legitimizes the idea of trading virtual goods for methodology. real money. It has become the dominant revenue model for online games in Asia and increases the overall virtual goods market size (Lehdonvirta & 3.2.1 Online game market size Virtanen 2010). On the other hand, it also means Early data on the global online game market is that significant parts of the value added by the sparse, but one major industry analyst firm suggests third-party gaming services industry is being that the global market was around $1.45 billion by co-opted and taken over by the official game 2003 (DFC Intelligence 2008). In recent years, publishers. The “primary market� for virtual goods KOCCA in Korea has made efforts to analyze and competes directly with the suppliers in the second- aggregate different sources in order to come up with ary market. However, many games, including World a more reliable estimate. The sources include of Warcraft, the most popular online game globally, estimates from a large number of industry analysts have largely stayed away from this model. There and industry organizations in different countries continues to be significant demand for third-party (KOCCA 2010). KOCCA’s analysis indicates that gaming services, as shown in the following section. the global market for online games was $12.6 billion in 2009, up from $8.5 billion two years earlier The influence of game publishers’ revenue models (Table 3). In terms of geographic breakdown, the on the opportunity for third-party gaming service current market is dominated by East Asia with providers is summarized in Table 2. China as the biggest market accounting for 32% ($4 TABLE 2. Opportunities for third-party gaming services in different types of games Game publisher’s revenue model Opportunities for third-party gaming service companies Subscription based revenue model High. The inability of the of�cial game operator to support trade of most forms of virtual products and services creates a large, latent demand to be ful�lled by third-party gaming services. Value creating opportunities limited by different barriers to trade, including efforts by the games operator to curb it. Virtual goods sales based revenue model Low. The ability of the of�cial game operator to address the latent demand limits third-par- ty providers’ opportunities. The degree depends on the speci�c design and revenue model of the game, e.g., does it use a separate “earnable currency� and “buyable currency�. Some third-party services, such as powerleveling, may remain very valuable. Sanctioned marketplace High. The operator demands a relatively high transaction fee from trades conducted on the of�cial marketplace, but this is offset by better security and easier access to customers. Source: Author’s elaboration 10 Knowledge Map of the Virtual Economy TABLE 3. The global online game market (in millions) Others (primarily developing Year China Korea North America Europe Japan countries) Total 2007 $2,200 $1,700 $1,500 $1,600 $700 $800 $8,500 2008 $2,400 $2,600 $1,700 $2,000 $800 $900 $10,400 2009 $2,900 $4,000 $1,800 $2,000 $900 $1,000 $12,600 2010 $3,700 $5,000 $2,200 $2,500 $900 $1,100 $15,400 Forecast 2011 $4,500 $6,000 $2,500 $2,900 $1,000 $1,100 $18,000 Forecast 2012 $5,600 $7,200 $2,900 $3,300 $1,000 $1,200 $21,200 Forecast Source: KOCCA (2010) Notes: 1. The market �gures include both revenues derived from business models based on subscription fees for online games and revenues derived from the sales of virtual items and services by the game operator. 2. While the forecasts for the developed markets (Korea, North America, Europe, Japan) are reasonable, the authors believe that the future market growth potential for developing countries under the heading “others� is underestimated. The growth forecasts do not take into ac- count the accelerating growth rate that is likely to be seen in several developing East Asian countries. Several recent industry analyst reports also suggest more rapid growth in these countries (e.g., Niko Partners 2010, Strategy Analytics 2010, IDC 2010). billion) of the global market, followed by Korea theme that can be seen in several recent analysts’ with 23 % ($2.9 billion). reports, in which the rapid growth of developing countries in East Asia is highlighted as important During its early period, the industry had a high (Niko Partners 2010, Strategy Analytics 2010, IDC growth rate. Data from industry analysts indicate an 2010). An industry analyst has estimated that the annual growth rate of around 50% from 2003–2005 near-term opportunities for further rapid online (DFC Intelligence 2008). This period could be game market growth in East Asian developing identified as the phase of early adopters in the online countries are primarily seen in Indonesia, Malaysia, game industry. In recent years the industry has the Philippines, Thailand, and Vietnam (Niko entered a stage in which the early majority has Partners 2010). started to play online games, with a global annual growth of around 20%. This is also the growth rate forecast for the next few years (see Table 3). 3.2.2 Earlier estimates of third-party gaming service market size It should be noted that this change represents a Compared to the market size for online games, it is general pattern of the industry, and not the high more difficult to measure the size of the secondary heterogeneity seen between specific markets. market for gaming studios that is not measured by Looking closer at specific regions, a highly diverging industry organizations, government bodies, or growth pattern between developing and developed disclosed in company public filings. countries is evident. There is a dual global market structure in which several developing countries In order to estimate the market size, earlier estimates increasingly drive the growth of the global online have relied on two methods, 1) trade platform game market, while several of the Western markets transaction aggregation and 2) industry manager have considerably lower growth rates. This is also a guesstimates. What could described as a third Third-party Online Gaming Services 11 method has been to derive a market estimate based growing rapidly. Market research firm In-Stat on a mixed method that combines these two methods estimates that the primary market size reached $7.3 and earlier results in their aggregation. billion in 2010 (Reisinger 2010). A Korean govern- ment agency estimated in 2008 that the value of The first method was initially the dominating one secondary market trading might have exceeded one among researchers and analysts. The method tries to trillion Korean Won ($900 million), in Korea alone derive aggregate figures from a limited set of (Park 2010). This report is mainly interested in the transactions taking place at the dominating trade secondary market activity, i.e. the part of the virtual platforms for virtual items. economy of online games that is managed by actors other than the game publishers. Edward Castronova (2005, 2006a, 2006b) was the first economist to study the real-money trade Both of the two earlier methods of estimating the (RMT) of game assets. In 2001, he guesstimated the secondary gaming service market are less reliable in size of the RMT market to $5 million by measuring the current market environment. Industry manager the daily volume of RMT transactions on eBay for guesstimates have become less reliable in a market the dominant online game at the time, EverQuest environment that is fragmented without any (Castronova 2006a). In 2004, Castronova estimated dominating third-party intermediary company for the total volume of RMT transactions on the secondary market transactions. The trade platform dominating global trading platforms eBay and transaction aggregation has also become less Korean ItemBay at $100 million per year. reliable. The method had certain flaws from the onset, e.g., it did not take into account the part of As trade volumes increased, what started as a the secondary market trading that took place player-to-player phenomenon was soon recognized outside these platforms, such as direct player-to- as a business opportunity. Third-party intermediar- player transactions. More importantly, the method ies such as IGE emerged to buy valuable virtual has become increasingly uncertain as previous goods from players, repackage them, and resell estimates were made at a time when there were them to players for considerable profit (Dibbell only a few dominating retailers in the market. The 2006). They acted as market makers and made the current market is considerably more heterogeneous trade considerably more efficient and easier than it and difficult to grasp. It is characterized by a large had been before. As a major market platform, IGE number of trading platforms with shifting market also had information that researchers lacked. In share numbers. Players access gaming services 2004, IGE’s president guesstimated the size of the through a variety of channels: direct marketing, global secondary RMT market as $880 million per consumer-to-consumer (C2C) platforms, and year. However, he did not provide a detailed various intermediaries. Different geographical description of his method. The declining market markets are using different platforms. Developing share of IGE and a more fragmented market that countries such as China also have a thriving followed meant that this form of industry manager domestic secondary market for their own player guesstimating no longer provides the same base. reliability. In 2007, Lehtiniemi and Lehdonvirta estimated that 3.2.3 Estimating the gaming service the size of the global primary and secondary RMT market through player surveys market had reached $2.1 billion, based on an Because of the changes in market conditions detailed aggregation of different sources (Lehtiniemi & above, today, the best approach is to estimate the Lehdonvirta 2007). This figure was the first to market size from data derived directly from player include primary market activity as well as secondary surveys rather than intermediate retailers. A lack of market activity. Primary market activity refers to credible survey data was previously a barrier for this virtual goods sales directly from game publishers to method. However, the number of suitable surveys, players. Secondary market activity refers to sales by their reliability and geographic coverage has third parties. In the late 1990s the market consisted increased in the recent years. This report relies on a almost exclusively of secondary market activity, but selection of surveys conducted by governmental more recently, primary market activity has been agencies, market analysts and scholars. 12 Knowledge Map of the Virtual Economy Based on data on the number of players using survey sources. The surveys indicate that around one secondary markets and their annual spending in four online game players are buying secondary (secondary market average revenue per user, ARPU), market gaming services each year, with slightly it is possible to calculate an estimated size of the higher percentages in Korea and China than in the secondary market and hence the third-party gaming Western market. The results are remarkably consis- services industry. To improve accuracy, the analysis tent across surveys. takes into account the differences in secondary market spending habits in markets with different economic characteristics (e.g., developed versus Average spending on the secondary market developing countries). This method also has its Compared to the participation ratio, average disadvantages and uncertainties. Some of these are a spending on third-party gaming services is more result of the method itself; others are related to the difficult to get access to though surveys and also empirical data used. For example, while some more at risk of retrospective misreporting (players surveys only cover virtual currency transactions, are more likely to accurately recall if they have used others also include powerleveling and virtual item secondary market gaming services than to accurately sales. The various sources of uncertainty in this recall the exact amount they spent). A Korean method are summarized in Appendix 1. government agency survey provides the most reliable data on the subject, but the data had to be modified The estimate is calculated using the following somewhat in order to arrive at a number of average formula: spending. The survey data did not enable a calcula- tion of average spending directly. Instead, this Secondary market size in a region = (the number had to be derived from a calculation of a number of paying online game players in weighted average of consumer spending in different the region) x (share of players using the spans (share of players spending $0–4, $4–9, and so secondary market in a region) x (the on). Despite these transformations, the numbers average amount of money spent on RMT presented should be roughly accurate. per player per year in the region). Table 5 presents the estimated annual average spending for developed and developing countries. In Participation ratio order to arrive at an estimate of average spending in Participation ratio in the secondary market for developed countries, data from a 2009 Korean gaming services is presented in Table 4 together with governmental survey was used. Based on this data it TABLE 4. Percentage of players buying from the secondary market Percentage of players buying from the Region secondary market Source Note Korea 24.2% KOCCA White Paper on Korean Survey conducted 2009. N = 772. Games (2010) China 24.9% CNNIC (2009a) Survey conducted in 2009 North America, 22% Yee (2005) Europe and Japan Other regions 24% Author’s estimate The majority of other signi�cant regions for the secondary market are located in Asia. The participation ratio for these markets is estimated to be roughly the same as for major Asian markets (China, Korea). Source: Authors’ calculations based on sources listed Third-party Online Gaming Services 13 TABLE 5. Average amount spent on the secondary market per year Estimated annual average secondary market Region spending (among participating players) Source and notes Developed regions: Korea, $369 Estimate derived from Korean governmental Europe, North America, Japan survey (KOCCA 2010) Developing regions: China and $87.50 Estimate derived from Chinese governmen- other developing countries tal survey (CNNIC 2009a) Source: Authors’ calculations based on sources listed is estimated that an annual average spending is non-professional gamers, this figure is also the same around $369 for players using the secondary market. as the gross revenues of the third-party online This data was extrapolated for other developed gaming services industry. The data does not allow the markets (Europe, North America, Japan) with some share of non-professional sales to be estimated, but it minor modifications. For developing countries, is probably not significantly large. China is the single survey data from a 2008 CNNIC survey was used, biggest market for third-party gaming services, with which yielded an estimated average annual spending an estimated market of $1,510 million. Globally, of $87.50 per player using the secondary market. third-party gaming services revenues are around one The lower annual spending in developing countries fourth of the revenues of the game industry itself is an effect of the differences in purchasing strength (estimated at $12.6 billion in 2009). among developed and developing regions, some- thing that is also reflected in differences between average spending on online games in the official market among developed and developing regions 3.3 Business models (i.e. the official market for subscriptions and point cards for online games). Various kinds of companies are involved in the third-party gaming services industry, specializing in different functions. A distinct value chain can be Number of paying online gamers seen emerging from the relationships between these Based on surveys and industry analyst estimates, the companies, distinguishable into primary and number of paying online game players at the end of supporting activities according to the internal value 2009 is presented in Table 6. Because of the focus chain model. The primary activities are depicted in on paying online game players, the focus is more on Figure 2. These are discussed in more detail below. MMOG players and other more dedicated online game players. The estimates are based on data from governmental surveys and industry analysts. The numbers are assumed to reflect the situation in Figure 2. Basic value chain in the 2009. It is estimated that there were around 121 gaming services industry million online game players in 2009, of which 37 million were from Europe, North America, Japan, Game operator and Korea. Some 84 million online game players were from China and other developing markets. Producer The secondary RMT market size Retailer Based on the above figures, the total global secondary market size is estimated to have been approximately Customer $3.0 billion in 2009 (Table 7). Except for transac- tions that took place between purely Source : Author’s elaboration 14 Knowledge Map of the Virtual Economy TABLE 6. Number of paying online game players Number of paying online game Country/Region players (millions) Source Note Korea 7 IDC (2007) IDC estimated that there were 5.769 million paying online game players in 2006. This report made a con- servative estimate that as of 2009 this had increased to around 7 million players. Because of the focus on paying users, the number is not including a large num- ber of casual online game players that are not paying for their online games. China 69 CNNIC (2009b), Data based on CNNIC survey of MMOG players. IDC and CGPA (2004–2009) North America, Europe, 30 Think Equity Number based on aggregated data and estimates and Japan (2009), MMOG from industry analyst reports. charts (2010), DFC Intelligence (2008, 2010) Other developing markets 15 Niko Partners Conservative estimate based on aggregated data (2010), IDC from Niko Partners, IDC, and Pearl Research for major (2007, 2010) developing markets (e.g., Indonesia, Malaysia, the Philippines, Thailand, and Vietnam). The data has been adjusted to take into consideration only paying online game players Source: Authors’ calculations based on sources listed 3.3.1 Retailers through websites, search engine advertising, and Retailers are companies that market the products in-game advertising. They process orders, screen and services of the third-party gaming services fraudulent orders (customers not intending to pay), market to the consumers. They reach out to gamers manage logistics (stocks and delivery), manage TABLE 7. Global secondary market size (C). Annual (B). Share of spending on the Secondary market (A). Number of players using the secondary market size (A×B×C). Country/Region players (millions) secondary market (millions) (millions) Korea 7 0.24 $369 $620 Europe, North America, 30 0.22 $369 $578 Japan China 69 0.25 $87.50 $1,510 Other developing 15 0.24 $87.50 $315 countries Total global market 121 $3,023 Source: Author’s calculations based on Table 4, Table 5, and Table 6 Third-party Online Gaming Services 15 payments, and take care of customer relationships. usually lack endorsement from the game operator Some retailers have even built up goodwill among and are subsequently also shunned by the main- the customers that could be equated with brand stream gaming media, the retailers have to use value. Examples of retail websites are listed in various alternative and sometimes even innovative Appendix 2. techniques to reach their target group. Advertising in in-game chat channels is a typical technique, but The first retailers were located in the same countries as it significantly detracts from the gaming experi- as the initial customers, that is, the United States, ence, both gamers and operators detest it. One Korea, and other wealthy gaming nations. They had innovative but also somewhat distracting technique their own production capability or sourced the seen in World of Warcraft is arranging a large number products from local players. As the market grew, of dead bodies on the ground to spell the address of they increasingly outsourced production to low the retailer’s website. labor cost countries, especially China. Due to the lack of language and marketing skills, the Chinese Perhaps the most important advertising channel for producers were unable to market their goods directly retailers is search-engine advertising, such as to consumers. But gradually, actors in developing Google AdWords. The size of the market and the countries developed the necessary skills and took degree of competition is reflected in the very high over the retail sector. Even once-dominant U.S.- prices of search terms related to virtual goods trade. based retailer IGE has moved its customer service For example, advertisements on the search term functions to the Philippines. One driver behind this “wow gold� can cost as much as $6–8 per click. change has probably been the increasing legal Because of the high customer acquisition costs, pressure put on U.S.-based retailers by game retailers often take a loss on the first sale. They rely operators. on repeat purchases elicited through direct market- ing (instant messaging, in-game communications, Perhaps the most typical retailer today is a Chinese and email) for profits. That fact suggests that the company located in a regional capital. According to retailers should be highly focused on customer an informant in the industry, eight of the largest relationship management and they should consider Chinese retailers have estimated yearly revenues in the customer database their most valued asset. This excess of $10 million, with an additional 50–60 may also explain the apparent trend towards firms with over $1 million in revenues. They employ consolidation in the retail sector. The hundreds of university educated customer service and manage- retail websites one can find on the web are in ment staff with good language skills. They source practice operated by a much smaller number of their production from producers in rural and actual companies. suburban China (some sources also suggest South- East Asia). The sourcing relationships can be very fluid. One retailer can have a network of thousands 3.3.2 Producers of small gaming studios that the retailer contacts The popular term “gold farmer� refers to a game when needed through the Tencent QQ instant laborer who plays an online game in order to messaging system. Retailers also maintain extranet produce virtual currency that can be sold for real “buy sites� where orders listed by the retailer can be money. The first “gold farms�, offices where multiple fulfilled by any producer. farmers sit at rows of computers and earn wages by producing virtual currency, probably evolved from In the case of virtual currency, retailers do not gaming cafés. Although gold farmers have captured always deal directly with producers. There are some the attention of the Western gaming public, today wholesalers that act as intermediaries between they represent only a small part of the production producers and retailers. Their role seems to be to sector in the third-party gaming services industry. deal with fluctuations in demand. They may also simply be commodity speculators that are attempt- The main competitor to gold farms comes from ing to buy low and sell high as prices fluctuate. automated bot farms. Bot farms use arrays of computers that each run several instances of the Marketing the goods and services of the gaming game, each controlled by a program known as a bot. services industry is a significant challenge. Since they Perhaps only a tenth of the staff of a manual farm is 16 Knowledge Map of the Virtual Economy needed to monitor the bots. Another source of ■ The criminal hacker groups that steal currency competition are criminal hacker groups that break from accounts also sell these hacked accounts to into players’ and gold farmers’ game accounts, steal some manual and automated farms, to be used the currency, and sell it for real money to wholesal- in currency farming or in-game advertising. ers and retailers. One industry expert suggests that These so-called “black accounts� are sold for manual farms produce 30 percent of the virtual much less than new accounts purchased from currency sold by retailers, bot farms produce 50 the game operator, but may have a shorter percent, and hacker groups “produce� 20 percent by lifespan before they are detected and closed by stealing it from other players. the operator. In some games, generating account keys (“CD keys� or “product keys�) can also be In contrast to virtual currency, powerleveling used to create black accounts artificially. services are produced exclusively by manual ■ Online games usually have dedicated servers for producers. Current bot technology is not sophisti- different countries or regions. Farmers and cated enough to do powerleveling in the games that powerlevelers need to be able to play on the are popular at the moment. It is possible for manual servers where their customers are, not on the powerlevelers to use bot-like technologies to servers of their own country. Access to, for automate some parts of the process, however. example, U.S. game servers may be blocked from Hacker groups are not able to do powerleveling, Chinese IP addresses. To overcome this, the although they might offer used accounts for sale. producers need supporting infrastructure in the United States. Producers can be individuals or companies. Perhaps ■ Chinese individuals and companies cannot easily the most typical producer is a company in suburban accept electronic payments from abroad. Some China. Their employees are young males with financial intermediaries are necessary to organize disadvantaged backgrounds and no occupational the payment flows. skills. These “playborers� are examined in more detail in section 4. Because of the lack of language and business skills, the producers are not able to 3.4 Regulatory framework and effectively market their products directly to custom- ers, especially high-value foreign customers. Their industrial policy access to the market takes place mostly through the retailers. The third-party gaming services industry operates at the crossroads of a number of international ICT policy and regulation issues, many of which are still 3.3.3 Game operators evolving. These include consumer rights in digital The online game operators (also referred to as the services, online gambling, regulation of electronic game publishers) provide the platforms and contexts payment services, and taxation of virtual transac- where the virtual goods are produced, used and tions. This section focuses on the most direct consumed. Without an online game maintained by challenge to the industry: whether the business an operator, there is no demand for game currency model of harvesting and selling virtual goods for real or powerleveling. Producers thus have to buy the money is and should be legal at all. game from the operator and pay the monthly subscription fees. The game design and business choices of the operator have a large influence on the 3.4.1 Negative externalities from third-party services market. Most game operators do trade of arti�cially scarce assets not endorse the third-party services and even Real-money trade (RMT) of game items and actively oppose them. This is discussed in more currencies is highly controversial. On one hand, detail in section 3.4. trade is a social good, and RMT is particularly valuable to players for whom time is scarcer than money. On the other hand, RMT can cause a variety 3.3.4 Supporting activities of negative externalities to other players and game Besides these primary activities, there are also some publishers. These negative externalities are discussed supporting activities that deserve to be mentioned: by, among others, Castronova (2004, 2005) and Third-party Online Gaming Services 17 Lehdonvirta (2005). Some of the first negative “ethical� sources for their products, that is, manual externalities discussed were the following: farms instead of automated bots or goods that are outright stolen—although there does not seem to be 1. RMT breaks the fairness of the game. If any real way to verify this. powerful characters can be bought for real money, the playing field is no longer level, but favors wealthy players. This decreases players’ 3.4.2 Contractual and legal enjoyment. regulation of virtual goods trade 2. RMT can break the “achievement hierarchy� of As a consequence of the issues highlighted above, the game. Game characters and other virtual the legality of secondary market trading as well as possessions can be seen as visible evidence of the whole third-party gaming services industry is players’ achievements in the game. RMT breaks frequently questioned. This topic has been examined the link between possessions and achievement, by, among others, Fairfield (2005), Duranske thus lessening the possessions’ information value (2008), Lastowka (2010), and Lehdonvirta and for the player community. Virtanen (2010). The following summary draws especially on the latter. In recent years these objections have become somewhat moot, however, as many game operators Publishers of Ultima Online, the first game to have themselves started to sell virtual currencies and attract a significant secondary market, welcomed goods to their players. One way to rationalize this is secondary market trade. A few publishers permit it that not having real-money trade tilts the playing on strictly controlled marketplaces (for example, field in favor of time-rich players. In the operators’ non-U.S. players are not permitted to trade in eyes, the third-party industry thus becomes some- EverQuest 2). The great majority of publishers are thing akin to a competitor. But third-party gaming strictly against secondary market trade. Their Terms services producers are also accused of creating other of Service (TOS) and End-User License Agreements negative externalities: (EULA) prohibit secondary market trading. Players (consumers and professional players) have to click to 3. Game laborers can monopolize game content. accept these terms as a condition for entering the Professional players can be so effective in game. This is intended as a contractual prohibition harvesting scarce game content that they against trading activities. practically monopolize it, making it hard for ordinary players to get their hands on the most The contractual prohibitions are not a very effective valuable resources. deterrent in practice. Players and gaming service 4. Automated bot players run by gaming services providers who break the prohibition can be difficult producers can degrade other players’ social to detect. Even when detected, they face no harsher experience. punishment than the closure of their game accounts. 5. In-game advertising degrades the play experi- Third-party gaming services producers treat creating ence. new accounts as a normal business cost. Game 6. Secondary markets create incentives for cyber- publishers can try to seek stronger remedies from criminals and scammers. Virtual goods are courts, but recent cases in the United States and among the most sought-after commodities in the Europe suggest that courts may not always be general hacking scene (Krebs 2009). This forces willing to honor their clickable contracts and license game publishers to spend more on security and agreements, which are seen as highly one-sided. The increases their customer service costs (although gaming services industry goes on in a legal dark-grey one retort is that indeed any market where goods area. can be resold is an incentive for crime). In Korea, special purpose legislation has been These issues affect different games differently, but in created to attempt to curb the negative effects of many cases the net social value of secondary market virtual goods secondary markets. The Game trading is probably negative. Interestingly, some Industry Promotion Act (Act no. 7941, enacted on gaming services retailers have begun to advertise that April 28, 2006) was amended towards this end in they refrain from in-game advertising and only use 2007. The amendment makes it illegal to trade 18 Knowledge Map of the Virtual Economy virtual goods for real money if the goods are either In China and Vietnam, the legal status of the 1) used as an instrument in a game of chance, such third-party gaming services industry is ambiguous. as a virtual card game, or 2) obtained through In China, the Bank of China has issued several exploiting security holes, using automated bot communications pertaining to virtual currency characters, or other “abnormal� play. Those violating trade. Some communications suggest that secondary the law may be sentenced to a maximum of five market trade is considered illegal, while primary years in prison or a fine not exceeding 50 million market sales (publishers selling virtual currency to Korean won ($45,000). Trade pertaining to assets users) are permitted. The authors’ efforts to get obtained through normal gameplay was left regulators and policy makers to comment on virtual unregulated. goods trade for this report were unsuccessful. The online game industry itself is a regulatory hot potato In 2009, the Supreme Court of Korea affirmed the in these countries, being seen as an innovative rule by acquitting two virtual goods retailers from growth industry to be supported on one hand, and a charges on the basis that the virtual currency they sold social problem to be dealt with on the other (Box 1). was not shown to be obtained through any means Informants suggest that third-party gaming services other than normal gameplay. This is interesting from a companies have not seen regulation as a problem in development perspective, because it suggests that practice. This probably reflects the fact that they “manual� gaming services providers that employ remain small and inconspicuous companies, people to play games are on a better legal ground in operating in large part through the informal Korea than automated “bot farms� that use machines economy. for the same job. Both may still be in breach of contract with the game publishers, however. BOX 1. The political economy of the Vietnamese and Chinese online game market The success of third-party gaming services providers in a country is to some extent linked with the development and regulation of the domes- tic game publishing industry. The same infrastructure that enables the growth of the online game market is essential for the establishment of gaming studios. A growing domestic market also fosters new players and playing skills. The surveys and interviews conducted for this report suggest that most Chinese gaming studio workers have a background in online game play and are signi�cantly motivated by their interest in online games. This box will therefore briefly examine the political economy of online game regulation in a developing country context, taking Vietnam and China as case examples. In Vietnam, the online game market is in its infancy, but shares many of the factors underlying the growth of the Chinese market. Internet broadband penetration and market size in relation to total population are currently at the same level as in China 2–3 years ago. Further broadband penetration is supported by the 2010–2020 ten-year plan in ICT. Demographic conditions are very favorable in Vietnam, with a median population age of only 27.4 years. On the other hand, online games in Vietnam are currently under increasing regulatory pressure. It is still uncertain which direction the regulatory policy will take in the years to come. During the fall of 2010, regulators were effectively halting all new operating permits of online games to review the situation. Before this, regulation regarding the virtual economy of online games had been increasing. The Vietnamese regulatory situation resembles the online game industry’s early years in China (Ernkvist & Strom 2008). In the Chinese case, the process has thus far resulted in a set of stricter regulations, especially regarding online game playing among youth, internet café opera- tions, and aspects of games that could be regarded as related to gambling. At the same time, regulators have recognized the industry as rapidly growing and economically important. As the economic importance of the industry has increased, rivalry between government minis- tries over the jurisdiction over the industry and its various business permits has surfaced. For example, there has been open rivalry between the Ministry of Culture (MOC) and the General Administration of Press and Publication (GAPP) in this area (Wang & Murphy 2009). Due to a number of factors, the online game industry is vulnerable to these forms of political regulation. Because of the nature of the medium, it is often subject to the jurisdiction of ministries from a range of �elds, including those regulating cultural content, publishing, and Internet security. As a medium having a signi�cant influence on youths and young adults, youth organizations with relations to the ruling party (such as the Communist Youth League of China) are also often involved in efforts to influence content and regulation of the medium according to their objectives. The political interest is closely coupled with the economic interest. The current economic size and operational income of leading online game companies makes the industry economically attractive for governmental ministries in various ways. The extensive regulation in China, often expressed in vague terms, means that online game companies have incentives to use their connections to improve the speed and likeli- hood of regulatory acceptance (Ernkvist & Strom 2008). These economic incentives are especially signi�cant in the online game industry, because of its high sunk costs in the form of high development costs of online games. As a service that is closely tied to the rapid changes in technology and user demand, it is vital for an online game to receive timely regulatory approval in order not to fall behind competition. Third-party Online Gaming Services 19 3.5 Case study: Purchasing young and university educated. Some of them see the employment as an opportunity to learn about virtual currency for World of trading and developing IT skills. In this sense, it is a Warcraft first stepping-stone for a planned job in another industry for these employees. World of Warcraft (WoW), published by Blizzard Entertainment, is currently the leading global online The retailer has been rather successful and growing game with over 12 million active player accounts as its revenues on a yearly basis. It is focused on the of 3Q 2010. U.S. players pay a monthly subscrip- Western market and has built up a customer service tion fee of $16.99 to Blizzard. As a leisure activity, it center that can handle requests in English around is not a casual, irregular activity for most players. the clock. However, business has become more Many players spend a significant amount of time difficult during 2010 due to the lack of new content each week in WoW; surveys indicate an average in World of Warcraft and the game operators’ weekly playing time of 21 hours for this category of constant efforts to curb secondary market trading. online games. As a leisure activity, it also has many social dimensions with long-term online friendship David places his order for WoW gold though the and social bonds formed in the game. retailer’s website and pays through PayPal. PayPal takes a transaction fee of approximately two percent David is a fictional WoW player. He is 34 years old, from this amount (Figure 3). When the retailer has lives on the East Coast of the United States, has a received the order, their customer service staff busy job, and an above-average income. He has played conducts an anti-fraud check and clears it for World of Warcraft for six months and became part of delivery. The order is then forwarded to the retailer’s a guild that carries out quests in the game together logistics department, which checks if the virtual and spends time chatting. It has become something currency is in stock. The logistics department logs like a circle of friends for him. Recently, David had to into some game accounts and determines that gold work longer hours at the office. As a result, he could for this server is out of stock. It places an order on not progress in the game as fast as the other avatars in the company’s Chinese language buy site, promising his guild. This creates both social stress for David (not to pay $68 for the gold requested by David. performing as well as his peer group) as well as a practical impediment to play, because the game is The owner of a gaming studio in suburban designed in such a way that only avatars of approxi- Changsha responds to the request. A cybercafé mately the same level can play together. owner established the studio in the fall of 2008. It has 10 employees who usually work over 60 hours David is well aware that Blizzard disapproves of per week, playing WoW to earn virtual gold (most players buying virtual gold for real money. He also of them also play some WoW on their free time, knows that many players disapprove of it, perceiving using their own characters). As a small gaming it as a sort of cheating. Yet he feels that many of the studio, it has benefited from a network with five activities necessary to progress in the game are other gaming studios that collaborate to handle highly repetitive and work-like, and not at all demand spikes and other problems. Lately, Blizzard exciting. David has often noticed advertisements Entertainment closed a large number of the studio’s that offer the game currency for real money, or offer accounts citing a terms of service violation and to play his character up to a certain level for a fee. caused a spike in costs. The game laborers at the He decides to give it a try. David uses Google to studio are on a low monthly salary plus perfor- search for “world of warcraft gold�. A large number mance-based bonuses. The owner logs into his of search results and advertisements come up. He WoW account and delivers the gold to the account clicks on a site titled “WoW Goldmining� (real designated in the buy site. A 21-year-old immigrant example with a fictional name). worker from Western China, who earned approxi- mately $23 for the corresponding work, originally A gaming services retail company based in harvested the gold he delivers. Changsha, China operates the site. The retailer belongs to a handful of large retailers that each have As soon as the retailers’ logistics staff confirms the close to 500 employees. Most of the employees are delivery from the gaming studio, they deliver the 20 Knowledge Map of the Virtual Economy Figure 3. An order of World of Warcraft gold for $100 Player places an order for $100 worth of gold PayPal charges a $2 transaction fee Large Chinese retailer (500 employees) takes a $30 cut Small Chinese gaming studio (10 employees) earns $45 ex. wages Game laborer who produced the gold earns $23 Source : Author’s estimate based on survey and expert interviews. gold to David. There are two possible delivery consumable potions that allow him to make faster methods. The retailer and David can agree on a progress in the game. In the following months, the time and place in the game world where the retailer’s customer relationship management retailer’s logistics staff member will virtually meet department occasionally contacts David through David’s character and hand over the gold, or the e-mail, MSN instant messaging, and even voice gold can be delivered using the in-game mail chat programs, offering discounts on additional service. David chooses the latter. Once he gets the gold purchases and introductory prices on power- gold, he spends it on equipment repairs and some leveling services. Third-party Online Gaming Services 21 22 Knowledge Map of the Virtual Economy Chapter 4 Microwork 4.1 Demand and supply potential customers. For example, Amazon used crowdsourcing to identify duplicate product pages on its massive e-commerce site. It developed a Since the 1990s, the widespread adoption of ICTs website where people could look at product pages has made it easier for companies and public agencies and get paid a few cents for every duplicate page they to outsource tasks and business processes into correctly identified. Other tasks that companies different geographic locations. For example, outsource to anonymous Internet users include American hospitals use medical professionals in the market research, data input, data verification, Philippines to transcribe doctors’ statements into copywriting, graphic design, and even software text. The geographic distribution of tasks and development. This has given rise to a market for paid processes allows organizations to benefit from local crowdsourcing. A market study estimated that over differences in labor costs and skill specializations. In the past ten years, over one million workers have traditional Business Process Outsourcing (BPO), earned $1–2 billion via crowdsourced work alloca- clients contract with BPO companies that rely on tion (Frei 2009). The estimate is based on worker their employees to carry out the work. Since the headcounts and gross payment figures disclosed by early 2000s, a new model called “crowdsourcing� ten companies that facilitate crowdsourcing. The has emerged alongside the traditional BPO model. advantages that companies seek when using crowd- It entails outsourcing tasks traditionally performed sourcing instead of their own employees or tradi- by employees or contractors to a large group of tional outsourcing are cost savings and the ability to people (i.e. a crowd) through the Internet (Howe adjust rapidly to changes in the volume for work. 2008). Typically this is done by issuing an open call for contributions on a website. Unlike Amazon, most companies lack access to a large pool of potential workers. They seek workers Most early examples of crowdsourcing are extensions through task marketplaces: websites where companies of marketing campaigns. For example, Fiat asked or persons post requests and individuals looking for consumers to submit design improvements and work respond. One of the first task marketplaces create marketing material for Fiat Nuevo 500 was created when Amazon opened its crowdsourcing (Kleemann et al. 2008). The campaign resulted in website for use by other companies. The resulting 170,000 designs, 20,000 comments on specific service is called Amazon Mechanical Turk. features, and 1,000 suggestions for accessories. Companies in need of crowd workers can post While the campaign was a great success from a requests on the Mechanical Turk either manually or marketing point of view, it is not known whether through an application programming interface Fiat actually implemented any of the suggestions. (API). Today, there are probably over a hundred task Also, contributors were not paid for their efforts. marketplaces on the web. Most of them have been Many projects where firms solicit individuals for opened in the last few years. In China, task market- contributions are better classified as consumer places are known as witkey ( ) sites, and there co-production or co-marketing rather than as are dozens of Mandarin language ones on the web. crowdsourcing, because they center on tasks that would not be performed by an employee or a Task marketplaces can be roughly divided into two contractor. categories: those specializing in professional services and those catering for “microwork�. The former In other cases, companies approach the crowd as a group consists of sites such as Elance, Guru, and genuine workforce rather than as 99designs that facilitate large, complex tasks, such as Microwork 23 software development and graphic design projects. subset of it. Microworkers can be sought through The latter is exemplified by Amazon Mechanical task marketplaces, but they can also be sourced from Turk, which carries a large number of very simple elsewhere, including a company’s own existing staff, tasks that can be completed in minutes or seconds, or the players of an online game. The innovation in such as verifying whether two images are of the same microwork is in the transformation of information product or transcribing a line of handwritten text. work into micro-sized units, similarly to how Taylorism and scientific management transformed Purchasing professional services through a crowd- manufacturing work in the late 19th century. The sourcing marketplace is not necessarily very different resulting microtasks and tools can then be optimized from purchasing them from traditional contractors. for maximum productivity, as well as distributed to It involves preparing a call for bids, interacting with new and innovative labor sources, crowdsourcing the bidders, and choosing a winner. The marketplace platforms being one. The key conceptual differences usually facilitates payment transactions. In contrast, between crowdsourcing and microwork are sum- purchasing microwork can be very challenging. marized in Table 8. First, tasks such as transcribing audio interview recordings into text must first be transformed into a form that can be disseminated to multiple workers over the Internet. This typically involves breaking 4.2 Market size the work down to suitably sized “microtasks�, constructing a user interface that workers can use to The total revenues of the microwork startups inter- complete the microtasks, integrating quality viewed for this report did not exceed $10 million in assurance into the process, and recombining the 2010. This suggests that the total value of the whole completed microtasks into a final deliverable. While microwork market in 2010 was probably not more in theory there is a huge amount of repetitive than double-digit millions. Some leading technology everyday work in offices and homes around the companies, such as Amazon, eBay, and Intuit are world that could be outsourced to microworkers, the already using microwork as part of their business bottleneck is transforming it into a suitable form. processes. Companies in other industries, such as Second, it can be a logistic challenge to identify a insurance and document archival, have only recently sufficient number of suitably skilled workers to started some pilot projects. Since microwork is a complete an assignment within a desired timeframe relatively recent phenomenon, it is more pertinent to and budget. For tasks with strict deadlines, it may assess future market potential rather than current size. be necessary to tap several marketplaces and work sources simultaneously. To assess the future market potential of the micro- work industry, it is useful to note that microwork Consequently, microwork is today emerging as a services are partially a substitute for traditional separate concept from crowdsourcing, rather than a digital BPO outsourcing services, and also partially a TABLE 8. Differences between crowdsourcing and microwork Task size Source of workers Workers’ tools Skills Crowdsourcing From tiny tasks Open calls Workers may require Basic computing skills to (Howe 2008) (2–30s) to large external tools language skills and professional projects (days or skills weeks) Microwork Tiny tasks Open calls, staff mem- All tools and informa- Basic computing skills to lan- (2–30s) bers, subcontractors, tion embedded into guage skills BPO providers, online worker’s User Interface games (UI) Source: Author’s elaboration and Howe (2008) 24 Knowledge Map of the Virtual Economy subset of the paid crowdsourcing market. Analysts’ 4.3 Business models estimates of the size and development of these markets may therefore give some indications of the Various different kinds of companies are involved in potential of the microwork market. the microwork industry, specializing in different functions. Although the industry is still young, a The amount of money earned by crowdsourcing distinct value chain can be seen emerging from the workers has been estimated at $1–2 billion over the relationships between these companies (Figure 4). past 10 years and the yearly revenues of the crowd- sourcing platform operators at approximately $500 million in 2009 (Frei 2009). The crowdsourcing 4.3.1 Transformers market is said to have experienced rapid growth in Transformers are the ultimate link in the value chain recent years. before the client. Their role is to take the clients’ problems and transform them into forms that can be Global exports resulting from the offshoring of IT addressed by microworkers. Different transformers services and business processes were estimated at specialize in different industries and problems. They $92 billion–$96 billion in 2009 (UNCTAD 2010, know their potential clients well and understand the p. 49). Approximately 60 percent of this is informa- clients’ business processes. They are good at network- tion technology offshoring and 40 percent is BPO. ing and marketing. As the public face of the whole Much of the offshored work is software development value chain, they give presentations and interviews to and similar tasks that cannot be addressed through increase awareness of the industry’s services. They are microwork, but some of it is pure data input and located in countries of high labor cost, as this is labeling that that could easily be addressed by where their clients are most likely to be found. microworkers. The transformers are also highly proficient in Besides substituting existing crowdsourcing and technology and innovation, because reducing offshoring arrangements, microwork can also be business problems into microtasks that can be expected to expand the market, because tasks that distributed over the Internet is technically challeng- are currently not economically feasible at all can ing. As the industry matures, the technology for become economically feasible when transformed to transforming common tasks such as content microwork. Microwork platforms that offer moderation and audio transcription is likely to be on-demand application programming interfaces may commoditized, and thus become available to find new customers among the millions of small and organizations with no special technical skills. medium sized online businesses. However, expanding the market by finding ways to bring new problems into the scope of microwork Based on this simple review it is not possible to put will remain a technology intensive function. Some forward any exact figures as estimates of future notable transformers are CrowdFlower, which serves microwork market potential, but the size of the e-commerce companies, CastingWords, which neighboring markets as well as the potential to specializes in audio transcription, and Microtask, expand the market suggests that the microwork which specializes in form processing and archive market could be worth several billion dollars within digitization. Transformers have also been referred to the next five years, as the technology matures. as “work parsers�. Figure 4. Basic value chain in the microwork industry Infrastructure provider Microworker Aggregator Transformer Client Game operator Source : Author’s analysis. Microwork 25 4.3.2 Work aggregators free, and instead earn revenues by selling virtual Work aggregators provide transformers with the currency to the players. The players buy virtual actual workforce needed to complete the tasks. They goods and value added services inside the game are able to distribute microwork to thousands of using virtual currency. The billing solution that workers. Different aggregators provide different allows game publishers to offer these “microtransac- kinds of workforces. For some types of microwork, tions� to the players is usually provided by another demand occurs in peaks and tasks are time sensitive, company, a monetization services provider. which necessitates a large stand-by workforce. For other types of microwork, demand is more constant Monetization services providers typically allow over time, allowing it to be satisfied with a smaller players to pay for their purchases using a large force or a large force of low-intensity workers. variety of payment methods, ranging from credit Language skills and demographics of the workers are cards to mobile payments. This is necessary in order also important characteristics of the workforce, as is to facilitate payments from different demographics obviously price. and countries. Two monetization services providers, Gambit and TrialPay, have recently added the option Different aggregators have very different approaches of paying for a purchase with work. If a player to building a workforce. The most conventional selects this option, they are presented with a set of aggregators are traditional BPO companies and call microtasks. After completing the tasks, the player is centers that simply hire a large number of employees rewarded with an amount of virtual currency. The and provide them with computers. When employees monetization provider then pays the game publisher are not engaged in other work, they can be directed in conventional currency. This allows people who to complete microtasks. A more novel approach to cannot afford the prices or lack access to payment aggregating workers is crowdsourcing. methods to obtain virtual goods and make full use Crowdsourcing platforms, also known as work of the features of a game or a community. exchanges and task marketplaces, list available tasks on a website and allow any Internet user to complete A key capability that all work aggregators share is a task of their choice against compensation. The the ability to not only contact a large number of pioneering task marketplace is Amazon Mechanical people, but also to channel payments or other Turk. Other task marketplaces include LiveWork, compensation to the workers in a cost efficient way. ShortTask, and Clickworker. Channeling monetary payments to foreign countries and especially developing regions with little financial A hybrid between the crowdsourcing and traditional infrastructure is a considerable challenge. Amazon BPO models are aggregators that use an open call on Mechanical Turk only pays out to U.S. bank the Internet to recruit workers to a standing virtual accounts and in gift certificates redeemable at workforce. These “cloudworkers� use their own Amazon online stores. In contrast, U.S.-based computers and Internet connections to do the work, non-profit organization Samasource aims to deliver but do not get to choose the tasks that they work digital work to women, refugees, and youth living in on. Cloudworkers are typically used for telemarket- poverty in developing countries. It has built a ing and customer service tasks, but can also be used network of partner offices in India, Pakistan, Haiti, for microwork. Some companies that provide Uganda, and Kenya through which it distributes cloudworkers are LiveOps and West at Home. payments to workers. The partners are typically Internet cafés run by local entrepreneurs, who Some of the latest and most innovative work provide all the equipment and training that workers aggregators to enter the workforce market come need to start earning. from the online game industry. In essence, online gamers are being used as microworkers and compen- Txteagle is a for-profit aggregator that likewise sated in virtual game currency. To understand how targets developing countries. It provides workers this is possible, it is necessary to provide some with tasks that can be completed on a mobile background. Today’s online game market is ex- phone, and rewards the workers with airtime tremely competitive, which is pushing the monthly minutes through partnerships with local mobile subscription fees for games towards zero. The most operators. Through 220 different mobile operators it popular online games let players use the game for can accommodate workers in 80 countries, the 26 Knowledge Map of the Virtual Economy majority of which are developing nations. Since be applied. However, microwork is not unique in mobile phones are the primary means of telecom- this respect. Telecommuters and mobile staff munication in many developing regions, airtime members have become commonplace in developed minutes are a valuable form of compensation for countries already since the 1990s, and employment many potential microworkers. However, the range law is changing to facilitate them. of tasks that can be completed on a mobile phone that cannot display graphics is limited. Cherry identifies two other issues that present a much bigger challenge in interpreting microwork through employment regulation. One is the 4.3.3 Infrastructure providers question of whether microworkers are to be viewed After the microworkers themselves, a final link in as employees or mere independent contractors. the value chain is the infrastructure providers that Employment relationships typically impose much provide the workers with the hardware, connectivity, heavier responsibilities on the employer than and financial services necessary to carry out the contractor relationships. Not surprisingly, micro- work. These consist of such companies as Internet work aggregators such as Amazon Mechanical Turk cafés and mobile operators. Most part-time micro- typically declare that their workers are independent workers in developed countries probably use their contractors. own computers and home Internet connections to carry out the work, but in a developing country On one hand, it is hardly satisfactory that profes- setting, it is more common to use a mobile phone or sional microworkers are excluded from the protec- access the Internet from an Internet café. The tions afforded to other workers. It increases their availability of such facilities depends on local vulnerability and most likely distorts the job market. demand as well as the general infrastructure of the In a related decision, the U.S. Internal Revenue region (electricity and communication networks). If Service in 2008 ruled that manual greeters for the suitable infrastructure is not available or is too Electric Sheep Company in the virtual world Second expensive to use, earning through microwork is not Life were employees, rather than independent possible. If infrastructure is available, microwork can contractors. probably promote its further development by generating demand and paying customers. On the other hand, current crowdsourcing and task marketplace-type microwork aggregators could probably not operate at all if they had to form full employment relationships with each worker. Low 4.4 Regulatory framework and transaction costs are vital in facilitating low-value industrial policy transactions. The regulatory burden could include dealing not only with tax authorities, but also with Miriam Cherry (2009, 2010) provides thorough insurance companies and occupational health early analyses of the policy and regulatory issues services. The workers’ diverse domiciles further relating to crowdsourcing and microwork from the complicate the issue. This may be an area in which perspective of U.S. and international employment attention from policy makers is needed to facilitate regulation. Employment regulation in general is the growth of the market. concerned with such issues as minimum wages, discrimination and equal opportunity, workplace Another question identified by Cherry is where to safety and compensation for injuries, unionization, draw the line between work and leisure, or volun- and privacy. For example, should employees bear teering. The question is important, because volun- responsibility for repetitive strain injuries (RSI) teering is typically excluded from employment suffered by microworkers? Existing regulations are regulations, such as minimum wage requirements. mostly borne out of the needs of traditional factory Microwork and other productive digital activities are and office work settings, and as a result are some- sometimes ambiguous in this respect. Examples of times difficult to apply to work that is conducted ambiguity include completing microtasks to earn over the Internet with no relation to a physical virtual game currency, completing microtasks to workplace. The lack of a physical location may also support a third party, such as a venture or a charity cause uncertainty as to which country’s laws should (sometimes known as “crowdfunding�), or Microwork 27 contributing content to a commercial site with a peaks and the work needs to be completed fast in promise of a potential future payoff. order to maximize sales. Many site operators thus seek to outsource this work. One company that they The informants also pointed to some specific can turn to is CrowdFlower, a start-up founded in regulatory topics. In the area of worker discrimina- 2007 by Lukas Biewald and Chris Van Pelt. tion, our informants suggest that microwork is According to the company’s self-introduction, virtuous in the respect that employees do not “CrowdFlower uses crowdsourcing to harness a necessarily even know the gender, race, and other round-the-clock workforce that spans more than 70 personal characteristics of their workers. Neither countries, multiple languages, and can access up to employment decisions nor compensation can thus half a million workers to dispatch diverse tasks and be based on any other factor than measured provide near-real time answers.� According to productivity. The flipside of this is that employees Biewald, the company’s revenues for 2010 are “single may not be in the position to enforce regulations digit millions� and it has “contracts for tens of that pertain to these characteristics, such as rules millions next year�. CrowdFlower is located in San against the use of child labor. Microwork is obvi- Francisco, near many of its potential clients, and its ously far from the dirty and dangerous menial tasks executives frequently give presentations in technol- that child labor is usually associated with, so there is ogy industry events. hardly need for alarm. Some microtasks are not very different from the unproductive leisure activities When CrowdFlower gets an assignment from a that young people engage in online anyway. Still, client, it needs to transform the requested work into virtual child labor may be an issue that needs to be a form that can be carried out by microworkers. For addressed in the future, especially if the number of example, for one major online retail/auction site, workers sourced from online games and paid in CrowdFlower provides an API that the client uses to virtual currency continues to increase. transmit source data and retrieve results from CrowdFlower’s server automatically. It has created a UI that displays fragments of the source data (search queries and search results) in a web browser and 4.5 Case study: Using human provides buttons for labeling the fragments appro- workers to optimize an online priately. This UI is presented to the microworkers. retail search engine Workers click on the buttons to provide their responses, and are presented with a new task after Major online retail and auction sites such as every click. CrowdFlower’s server infers the best Amazon, eBay, Taobao, and Rakuten offer selections solution to each task from the workers’ responses, that consist of millions of different stock-keeping and passes the results to the client. A large part of units. Providing a search engine allowing customers CrowdFlower’s work is quality control. Workers who to find and purchase what they want from a large fail to respond correctly to at least 70 percent of selection is crucial to success for these sites. To make periodic screening questions are prevented from products show up in search results, it is often contributing further to the assignment in question. necessary to annotate them manually with relevant At the same time, CrowdFlower constantly opti- labels. Humans can also help improve the accuracy mizes the UI and instructions and examples are of search algorithms by labeling search results in a provided to the workers in order to enhance way that reflects how well they correspond with the accuracy and productivity. search query. These tasks are important, because inefficient search engines can cost thousands or During September 2010, the online retail client gave millions of dollars in missed sales and make it CrowdFlower four different search labeling assign- difficult for consumers to find what they need. Since ments. CrowdFlower broke these assignments down product inventories are constantly changing, this to 178,955 fragments. Each fragment was distrib- work has become a standard part of the supply chain uted to several different workers to facilitate for some online retailers. statistical inferences. This resulted in a total of 1,541,814 tasks to be completed by workers. Human labeling of products and search results is CrowdFlower specializes in the marketing, trans- time consuming, but the demand for it occurs in forming, and quality control functions, so it has no 28 Knowledge Map of the Virtual Economy direct contact with the actual microworkers. Instead, communities. They help online game publishers it uses various work aggregators as a labor source. In earn money from games that can be played for free September 2010, CrowdFlower used the services of on the Internet. When Gambit receives work six different work aggregators to satisfy the requests requests from CrowdFlower, it puts them in a of the retail client. One of these sources was Amazon queue. Every time a player playing one of Gambit’s Mechanical Turk. Another was Samasource, a partners’ games indicates that they need more virtual non-profit organization that delivers digital work to currency, Gambit displays one of the microtasks to women, refugees, and youth living in poverty in the player. After completing the task, the player is developing countries. Two others were Gambit and rewarded with virtual currency. The player then uses TrialPay, which assign microtasks to online gamers. the virtual currency to purchase new virtual items CrowdFlower chooses which labor sources to use for and services inside the game or online community. each assignment by assessing factors such as cost, Gambit compensates the game publisher in real language and other skill requirements, age restric- money for the virtual currency disbursed. tions, and work content that may be objectionable in certain cultures. Overall, the six aggregators provided the retail client with labor from approximately 6,000 workers in 50 Leila Chirayath Janah founded one of countries during September 2010. Over a third of CrowdFlower’s main work aggregators, Samasource, the work came from the United States. Other large in 2008. Samasource’s client-facing functions are contributors were India, Pakistan, and Philippines, based in San Francisco, and it provides not only which reflects the fact that this work required work aggregation but some microwork transforma- proficiency in English. Workers in Kenya provided tion services as well. It has provided services to approximately four percent of the work. The LinkedIn, Intuit, and the U.S. State Department. Its country distribution of the work is shown in typical tasks include content moderation and Figure 5. converting scanned documents into digital form. When Samasource receives work requests from The total amount of money disbursed by CrowdFlower, it distributes the requests (which are CrowdFlower for this work was $8,797.15, or essentially links to pages on CrowdFlower’s web approximately half a cent per microtask. This equals server) to its international network of delivery only approximately $0.70 per worker per month, centers. The delivery centers are simple offices that which reflects the fact that most workers completed contain multiple computers connected to the only a small number of tasks. Those who completed Internet. A worker who gets paid for each task they more tasks could earn significantly higher payouts. complete operates each computer. The centers are According to data provided by CrowdFlower, the monitored by Samasource staff, but owned and operated by local entrepreneurs. For instance, one delivery center is located in a refugee camp in Dadaab, Kenya. The workers there Figure 5. Distribution of are mostly refugees from the neighboring war-torn CrowdFlower’s search labeling state of Somalia. According to Samasource, training provided by Samasource staff allows even refugees work by country in September 2010 who have never seen a computer before to earn Other 11% money as microworkers. Many Somalis understand Kenya 4% United States 37% English and are thus able to take on search labeling tasks originating from CrowdFlower’s American Indonesia 4% clients. Lack of cultural knowledge nevertheless Canada 5% causes difficulties in search related tasks and limits Philippines 8% the quality of the contributions. Other two work aggregators that CrowdFlower taps Pakistan 12% India 19% into, Gambit and TrialPay, are companies that provide monetization services for online games and Source : CrowdFlower. Microwork 29 average payout in the four different assignments was between $0.15 and $1.15 per hour (Table 9). TABLE 9. Number of participating CrowdFlower’s payout includes both the work workers and average hourly aggregator’s revenue share as well as the worker’s payout in four assignments during compensation, so the actual amount of money that September 2010 reaches the worker can be much lower. In the case of Amazon Mechanical Turk, the aggregator takes approximately 10 percent of the payout and relays Average 90 percent to the worker. In the case of Samasource, Assignment Workers payout/hour the aggregator takes approximately 15–30 percent of 1 657 $1.15 the payout, 30 percent goes to the local infrastruc- 2 235 $0.15 ture provider (delivery center) and workers get approximately 40 percent. Based on these figures, it 3 5,230 $0.47 can be calculated that the workers in Kenya earned a total of something like $140 from these assignments 4 1,422 $0.41 during the month. Source: CrowdFlower Gambit and TrialPay give their workers none of the actual payout, but compensate them in virtual currency, so the value experienced by the worker is percent was directed towards virtual currency largely subjective and depends on the individual. Of payments and 72 percent towards national the $8797.15 disbursed by CrowdFlower, 28 currencies. 30 Knowledge Map of the Virtual Economy Chapter 5 Other Segments of the Virtual Economy 5.1 Cherry blossoming operating in this area are Subvert and Profit, Bulkfans and Usocial. Companies, brands, artists, and publishers are increasingly turning to the Internet to reach new This activity could be called “cherry blossoming� customers and fans. Worldwide spending on after the Japanese term for paid fans or spectators Internet advertising totaled $65 billion in 2008 (sakura). Since cherry blossoming involves users (IDC 2009). But the best kind of publicity is not recommending brands or products for money, it advertising. Comments, blog posts, and forum decreases the information value of recommendation entries by individual consumers greatly influence the engines and rating systems. It may not be illegal, but buying decisions of their peers. Even more influen- it certainly goes against the intentions of the tial are recommendation engines, rating sites, designers of the systems, and provides no added toplists, and similar systems on other websites that value to any of the other users. Thus, although aggregate consumer opinions to provide buying cherry blossoming is in many ways similar to advice to visitors. On social networking systems microwork, the crucial difference is that it is directed such as Facebook and Twitter, the number of towards overcoming artificial as opposed to natural connections, followers, or fans that a brand or scarcities. individual has works as an indicator of their popularity and desirability, and helps them appear more often in search results. In the digital world, 5.1.1 Case study: Improving the consumer endorsements such as these have direct visibility of an online store economic value. This is an actual case with fictional details added and company names changed. AirMobile is a new British In recent years, a market has emerged for various online store for mobile phone accessories. There are kinds of consumer endorsements. For example, an already numerous established online stores in the individual who owns a Facebook account can be accessories market, so consumers looking for phone paid $0.10 to become a fan of a particular artist. accessories online are unlikely to ever bump into The typical value chain of this activity is as follows. AirMobile. AirMobile uses search engine advertising A publisher or brand owner asks a social media to attract consumers, but since there are many marketing agency to improve the brands’ or artists’ advertisers in this market already, the cost of visibility. The marketing agency, either directly or attaching ads to relevant search terms is high: a through an intermediary, contacts a large number of single click on the phrase “phone accessories� costs people having accounts in the social media service(s) $3.50 on Google AdWords. Since AirMobile’s target in question. Often this contacting takes place group is mostly on Facebook, it could also buy some through the very same crowdsourcing marketplaces Facebook ads, which are slightly less expensive. But that are part of the microwork value chain. In other the problem with both types of ads is that their cases, an intermediary has built a standing reserve of effect is more or less temporary: after the advertising individuals ready to take assignments via email or budget is spent, the company is just as invisible other direct channels. Finally, the individuals online as it was when it started. complete the assignment and are paid by the agency, possibly through the intermediary or the task AirMobile learns that in order to start showing up in marketplace. The payment may also take place in Facebook searches automatically, it would have to credits or virtual currencies. Some intermediaries have a large number of people liking it. Facebook displays search results in the order of popularity. up the task. For larger projects, AZ Consulting This is similar to how Google orders search results could ask a programmer to write a script that checks based on the number of other pages linking to them. the completions automatically. Each of the 290 The usual way to attract a large number of likes and workers gets $0.10 deposited on their ShortTask links is to build one’s online reputation over time, account. Once they have earned $50 or more, they contributing interesting content, and possibly using can request the money to be paid out to their PayPal occasional advertising campaigns. AirMobile does account. not have the patience or skill to do this by itself, so it hires AZ Consulting, a small social media After two weeks, the consultant tells AirMobile via consulting agency, to improve its online visibility. email that the goal of the project has been met. AirMobile is now slightly more likely to show up in The objective of the social media visibility campaign searches, thanks to its fabricated popularity. is set at 250 Facebook likes. AirMobile will pay the consultant a lump sum, negotiated down to $3000, when the goal is met, and otherwise participate in no way in the execution. AirMobile assumes that over 5.2 User-created virtual goods the next several weeks, AZ Consulting will produce some interesting content for its Facebook page and In recent years there has been much discussion engage people in conversation. But the consultant is regarding new, easily accessible entrepreneurship busy with other projects and decides to use a opportunities relating to content production for shortcut. It visits the website of ShortTask.com, a new online marketplaces. The first prominent California-based crowdsourcing platform with over discussions related to 3D virtual worlds. More 125,000 registered workers. The consultant uses recently, authors have highlighted other so-called PayPal to deposit $50 on its ShortTask account, and two-sided marketplaces that bring together custom- announces the following task on the site: ers and independent producers. Examples of the latter include the Facebook application interface and Easy! Facebook Like iPhone App Store. Step 1. Login to your Facebook account (you Entrepreneurship opportunities in 3D virtual worlds must have one to complete this task) have mostly focused on the leading virtual world platform Second Life, and to a lesser degree, Project Step 2. Go to this Facebook page: http://www. Entropia. Both these virtual worlds provide a virtual facebook.com/AirMobileAccessories currency that is exchangeable into U.S. dollars, making it possible for its users to earn real world Step 3. Click on the “Like� button there. currency. In addition, Second Life facilitated an unprecedented degree of user generated 3D content Step 4. Copy and paste your Facebook and scripts that add functionality to the creations. username in the field below so that we can The initial growth of these two virtual worlds also confirm that you have completed the task. led to the announcement of a number of other platforms, several of them in developing countries The reward for the task is set at $0.10, and the task (e.g., HiPiHi in China). The virtual economy of is made available in 290 copies. Over the next few these platforms enables entrepreneurial ventures in days, thousands of individuals who regularly the creation of virtual goods for sale, development of complete tasks on Shortask.com for some additional virtual land and real estate, virtual tourism, and income see AZ Consulting’s task, and eventually all various other services. Second Life gained consider- 290 copies of the task are taken up by a worker. The able attention for its entrepreneurial opportunities workers are mostly from India, Bangladesh, and the in the business press and academic journals during Philippines. 276 likes appear on AirMobile’s 2006 and 2007. However, many such earlier studies Facebook page, so some workers did not complete were primarily based on individual examples and the task properly. Instead of manually going through focused on potential future opportunities (e.g., each username to find the culprits, AZ Consulting Ondrejka 2007). The fact that some of the underly- decides to just pay all of the 290 workers who took ing characterizing factors for real-world 32 Knowledge Map of the Virtual Economy entrepreneurship and economy were present in entrepreneurship in virtual world platforms is Second Life did not automatically mean that it unlikely to provide development opportunities and becomes a breeding ground for large-scale entrepre- employment in developing countries on a larger neurship opportunities (Evans 2007). More recent scale. data indicates that large-scale entrepreneurship opportunities on the platform have largely not been realized. 3D virtual worlds still struggle to break out from their niche market status; leading Second Life 5.3 Other two-sided had stagnant participation during Q2 2010.2 marketplaces A recent study surveyed entrepreneurship in both The global market for social networking services Second Life and Project Entropia (Kieger 2010). The (SNS) was estimated to be 830 million users in study’s method and relatively limited number of October 2009 (SK Communications 2010). participants mean that it cannot be seen as a general Facebook is the leading SNS platform with 500 representative study of entrepreneurship conditions million active users as of August 2010. Facebook is on these platforms. Nevertheless, it provides a better also an example of the convergence of technological understanding of some of the characterizing features platforms for SNS with 150 million users accessing of virtual world entrepreneurship. For both plat- the service from their mobile phones. Notable forms, the degree of experience of the specific virtual Chinese SNS platforms include Tencent QQ (68 world and perception of “fun� were relevant for million users in late 2009), Baidu Space (63 million launching a new venture in the virtual world. users), and Kaixin001 (25 million users). Other Entrepreneurs had examined the specific market notable SNSs from developing countries include segment they had entered in the virtual world and Russian Vkontakte with 23 million users and South defined a business strategy, indicating that entrepre- Korean Cyworld with 21 million users (SK neurship recognition is an important aspect of virtual Communications 2010). world entrepreneurship. As a result, the ability to rapidly respond to and adapt to these changes was an Facebook recently implemented a virtual currency in important entrepreneurial precondition for virtual its platform, offering new opportunities for indi- world entrepreneurship. There was a high degree of vidual entrepreneurs to profit from the microtrans- environmental uncertainty and turbulence for virtual actions of virtual items and services on the platform world entrepreneurship, a result of the changes in (Shafer 2010). The credit also offers new opportuni- regulation of the game imposed by its operator as ties for the non-profit sector and charity by enabling well as the rapidly changing economy, user bases and a new fundraising tool directly to users. Several large technology of the virtual world itself. SNS application developers come from developing countries, but they are predominantly mid-sized Second Life’s operator Linden Lab provides their developers from technology clusters in larger cities.3 own statistics of what they define as unique users The location advantages provided by these technol- with a positive monthly flow of dollars. According ogy clusters are likely a reflection of a combination to Linden, the positive flow was gained through the technological skills, business networks, and content creation or other services in the game. There economies of scale involved in the business of SNS were around 66,000 users with a positive flow in application development. August 2010, but the majority earned less than $50. The number of unique users with a positive flow of The global market for smartphone applications is more than $100 in a month was limited to 6,900 dominated by Apple’s mobile operating system (iOS) users. This number has been declining during 2010. The economic market for content creation and 2 Second Life reported 805,000 monthly repeat logins during the quarters. Of these accounts, around 493,000 participated in the games economy individual services in virtual worlds is still relatively through the economic transactions of the games currency. limited. Growth of the market for virtual world 3 China could serve as an example of this. Two of the top 10 Facebook ap- plication developers measured monthly active users (MAU) comes from Hong entrepreneurship has stagnated, raising questions Kong (Pencake Limited and 6Waves). For the SNSs RenRen.com and Kaixin. com from Mainland China, many of the dominating developers are from concerning the future potential of the market in the Beijing (Rekoo and Tong Chi-Star Technology) and Shanghai (Five Minutes and short and medium term. As a result, Kai Ying Network Technology). Other Segments of the Virtual Economy 33 with a cumulative number of more than 6.5 billion resource requirements compared to, for example, application downloads as of September 1, 2010.4 traditional video game development. In this sense With a software development kit that is relatively they could be seen as part of the virtual economy accessible in terms of technological skills and a that emerges on top of digital services. However, the standardized payment platform, a large number of startups that rose from obscurity to exploit this individual developers have become successful opportunity, such as Zynga and Playfish, have since entrepreneurs on the platform. Low development become large companies with valuations of hun- costs have been crucial and individual innovative dreds of millions to billions of dollars. The “viral developers with lower overhead costs have had a loops� in platforms like Facebook that allowed the competitive advantage over many larger companies products of these companies to reach millions of (Bloomberg 2010). A list of top 50 applications in users for free have since been largely closed and Asia reveals a large presence of developers from replaced with paid advertising. For these reasons, Indonesia, Malaysia, Philippines and India (e27 this market is now increasingly dominated by 2010). Several app developers from developed established players and starting to resemble tradi- countries are also outsourcing a large part of their tional digital content business. app development efforts in order to reduce costs. When two-sided marketplaces like Facebook and 4 Application development for other smartphone operating systems has iTunes App Store first emerged, they presented a also been growing rapidly, most notably Android. According to one estimate significant entrepreneurial opportunity for develop- from May 2010, Android had the second highest global market share of Smartphone operating systems (26%) after Apple’s iOS (40%) (AdMob Mobile ers, with low entry barriers and low skill and Metrics, http://metrics.admob.com/ (accessed March 7, 2011)). 34 Knowledge Map of the Virtual Economy Chapter 6 Development Potential of the Virtual Economy In the previous sections, the report introduced the In order to analyze the current sociology of work in main areas of the virtual economy and described the this field, the authors decided to attempt a survey different business models that are used in them. In study of gaming studio workers. The result is the each of the areas it was found that at least some part first published survey in this field. Respondents were of the economic activity physically takes place in recruited with the help of an industry expert in developing countries. This suggests that they have China, using the different communication channels potential implications for the development of local that intermediate retailer companies use to contact economies and ICT infrastructure. In this section, gaming studios. The questionnaire covered a wide the report examines this development potential of range of topics (Box 2). the virtual economy in more detail. For each of the areas, the report first examines what is known about The results must interpreted carefully due to the their current implications for individual workers: small sample size (N = 26) and the fact that the skills that they require, wages that they pay, and sampling method is not nationally representative in career development opportunities that they offer. China, let alone globally. Nevertheless, it provides The report then assesses the potential for income interesting indications in light of the lack of any growth and local economic development in the previous quantitative studies in the field. Future future by analyzing income distribution, entry efforts with a larger survey using more stringent barriers, and upgrading strategies in the value chains. methodology could provide more reliable results concerning this large shadow industry. 6.1 Third-party online gaming Geographic location Table 10 shows that most of the gaming studios in services our sample came from ChangSha, the capital of the Hunan province in China, and Beijing. Because our 6.1.1 Worker demographics, skills, survey was not drawn from a nationally representa- wages, and career development tive sample, it is difficult to draw far-reaching The sociology of work of gaming studios is relatively conclusions regarding locational aspects. However, it unknown. Some visits to gaming studios have been seems that most gaming studios are close to some of reported in the press and media, but conclusions have been drawn from a very small sample of gaming studio establishments. Despite the rapid change and heterogeneous market for gaming studio services, these few cases have become representative BOX 2. Sociology of work topics of the general view of gaming studios (Nardi and covered by the survey Ming Kow, 2010). Recent reports from the industry also indicate that the classical view of the sociology Geographic location of gaming studios Size of gaming studios of work of gaming studios is changing. Rather than Gender distribution a large number of workers playing online games on Prior activities before starting gaming studio work Skill formation from gaming studio work in past and future career their computers, the largest majority of gaming Perceived problems related to gaming studio work studios are now bot farms consisting of a large Relatives’ views on gaming studio work number of computers on which management Working hours Salary levels oversees automated bot programs. places some restrictions on the location. The expert TABLE 10. Geographic location of interviews also indicate that other factors such as surveyed gaming studios local “guanxi�, or informal networks of influence, have an important role in localization decisions. City Number ChangSha 12 Size of establishments and gender composition of workforce Beijing 9 The survey results suggest that gaming studios are ZheJiang 1 highly heterogeneous, both in terms of size and in terms of gender composition (Table 11). This makes Chongqing 1 it difficult to describe the size or gender composition HuaiHua 1 of a “typical� gaming studio. The survey had 12 responses for questions about the number of Nan Chang 1 employees and 9 responses for its gender composi- tion. The sizes of the gaming establishments in the ShangHai 1 survey were in the range of 10–200 workers. N=26 Source: Authors’ survey Established views of gaming studios as male dominated workplaces are not always correct. While three of the smaller gaming studios in the survey the major cities with few examples from rural areas. had an all-male staff, all gaming studios with 100 or Access to the necessary technological infrastructure more employees reported a more gender diverse TABLE 11. Size and gender diversity of gaming studios Number of female Percentage of male Number of workers Number of male workers workers workers 200 NA NA NA 200 150 50 75% 200 200 100 66% 150 100 50 66% 140 70 70 50% 100 60 40 60% 30 NA NA NA 20 20 0 100% 20 20 0 100% 15 15 0 100% 10 2 8 20% 10 NA NA NA N = 12 Source: Authors’ survey 36 Knowledge Map of the Virtual Economy TABLE 12. Gaming studio workers’ prior occupations Prior activity before starting gaming studio work Student Unemployed Working Number of respondents 1 6 9 Percentage of respondents 6% 38% 56% N = 16 Source: Authors’ survey workforce, with 50–75 percent male workers. There was also a smaller gaming studio with a female BOX 3. Perceived value of skills majority; it was reported that only 20 percent of derived from gaming studio work in workers were male. earlier and current life Have some of the skills that you have learned working at the Prior activities before starting to work at the gaming studio been helpful in some other work or activities you have been engaged in? gaming studio Yes: 7 Asked about their prior activities before starting as No: 8 gaming studio workers, over half of the respondents N = 15 Some responses: indicated that they had been working (56%), over “I develop relationships in my work.� one third had been unemployed (38%), and a small “yes, in the international business trade� “yes, it’s better to learn more� proportion had been students (Table 12). This indicates that gaming studios provide job opportu- Source: Authors’ survey nities for unemployed young adults and attract youth that are not in education. BOX 4. Perceived value of skills Skill formation and gaming studio work derived from gaming studio work One important question is the potential role of for future career gaming studio work in capacity formation towards other activities. When asked if the skills gained from In the future, do you perceive that the skills that you have learned gaming studio work had been helpful in their as a gaming studio worker will be helpful to you in other work or activities that you want to engage in? current or past activities, around half responded Yes: 9 positively (Box 3). In the detailed elaborations, one No: 6 N = 15 respondent cited developing relationships and Some responses: another cited skills in international business. When “yes, my mind was activated in my work� respondents were asked if they perceived the skills “yes, on-line game promotion� “yes, I can begin my own career in this industry.� they gained from gaming studio work as helpful in “yes, it can help develop my career� their future careers (Box 4), slightly fewer than two Source: Authors’ survey thirds responded positively. In the detailed elabora- tions, one respondent cited skills usable in online game promotion work, while another saw the skills they gained as useful in pursuing a career in within only pass their time in gaming studio work without the gaming services industry. future goals or plans. Taken together, these answers indicate that many gaming studio workers see their work as a source of Problems faced by gaming studio workers positive skill formation towards future careers. This The survey identified some potential problems and provides a different picture than that of workers who concerns that gaming studio workers face in their Development Potential of the Virtual Economy 37 work and asked them to grade them on a 0–4 point survey of 2005, indicating that the perception for scale with ‘0’ if the issue is currently not a problem, the need to control online games has increased ‘1’ if the issue is a small problem, ‘2’ for a moderate among the Chinese public over time (Liang 2007). problem, ‘3’ for a major problem, and ‘4’ for a Informants suggest that such views are a problem for problem that makes it impossible for them to gaming studios as well as the game industry itself, as continue their gaming work. they make it more difficult to recruit workers. Based on 15 responses, issues perceived as major Working hours problems included 1) working hours and 2) negative views from family and relatives. The earnings level Severity of problem: Major was also perceived as problematic, but to a lesser degree. Most workers saw other players’ views as a Working hours were identified as a problem by most small or moderate problem. of the respondents. From our survey and earlier studies, it is known that gaming studio workers Of these four factors, the earnings level is a problem work very long hours. One of the difficulties with that gaming studio work shares with many other playborer activity could also be that work and free jobs in developing countries. The other issues are, at time become mixed up, providing few options for least to some degree, specific to the sociology of relaxation and time off. This is an aspect of the mix work in gaming studios. between leisure activity and work that has not been thoroughly researched. Some accounts of field Negative views from family and relatives studies also indicate that workers consume alcoholic regarding gaming studio work drinks during working hours, which could be a way to attempt to deal with long working hours. Severity of problem: Major Considering that most of the playborers are young adults, the long working hours are a a potential More than a fourth of the respondents perceived barrier for family formation. negative views from family and relatives as a major problem for their work. While the sample is too The present survey asked specifically about the small to draw far-reaching conclusions, the finding weekly working hours of the respondents, using five suggests that further attention on this topic is response categories: 0–10 hours, 11–20 hours, merited. Online gaming is a new activity in many 21–40 hours, 41–60 hours, and 60 hours or more. East Asian developing countries, and online game The results are presented in Table 13. The largest play has a bad reputation among many parents. The group of workers (38%) worked more than 60 hours reasons are likely to be multifaceted. It is known a week, indicating that concern over overworking is from earlier game and media studies that various certainly warranted. Few worked 41–60 hours, forms of social fears often accompany a new which could be considered a fairly normal Chinese medium (Williams 2003). Online game playing is full-time working week with some overtime. A seen as a time consuming activity, and in East Asia it relatively large group was working only 11–20 hours is often consumed within the “third place� of the (31%). A possible explanation is that these people internet café—which is itself seen as a threat that use gaming studio work as part time job alongside takes resources and time away from studies and other activities. Future studies should try to study more “serious� activities (Golub & Lingley 2008, the overworkers and part-timers separately, as their Ernkvist & Strom 2008). Within authoritarian problems and conditions probably differ. states, there are also several biases against online gaming that seem to have other underlying reasons, Earnings level including the need to control new media and its content. The most reliable survey data on this topic Severity of problem: Moderate to Major suggests that around half (49%) of the Chinese public in 2007 identified online games as a form of The earnings level of gaming studio workers has content they thought should be controlled. The been a matter of dispute in earlier studies (Heeks number has gone up from 16% in the previous 2008). Earnings must obviously be set in relation to 38 Knowledge Map of the Virtual Economy TABLE 13. Average weekly working hours of gaming studio workers 60 hours or 0–10 hours 11–20 hours 21–40 hours 41–60 hours more Number of respondents 1 5 3 1 6 Percentage of respondents 6% 31% 19% 6% 38% N = 16 Source: Authors’ survey the hours worked. Many gaming studios seem to reports earning 4–10 RMB ($0.6–1.2) per hour. A have adopted a compensation system that combines few earn considerably more, in the range of 30–88 a fixed basic income with a performance based RMB ($4.6–$13.4) per hour. payment scheme. Salaries vary considerably depend- ing on the position of the gaming studio worker. Currently, the minimum wage for part-time work in Beijing is 11 RMB ($1.7) per hour. While it is The present survey asked specifically about the slightly less in some other parts of China, this still respondents’ hourly earnings level (Table 14). The implies that the earnings levels of the majority of results show considerable variance. The majority gaming studio workers are close to or below the TABLE 14. Hourly earnings of gaming studio workers Respondent number Hourly earnings in RMB, low to high Hourly earnings in dollars, low to high 1 ¥4 $0.6 2 ¥5 $0.8 3 ¥5 $0.8 4 ¥6 $0.9 5 ¥7 $1.1 6 ¥8 $1.2 7 ¥10 $1.5 8 ¥10 $1.5 9 ¥30 $4.6 10 ¥30 $4.6 11 ¥88 $13.4 Average wage ¥18 $2.7 N = 11 Source: Authors’ survey Note: One outlier that was regarded as unlikely was omitted. Development Potential of the Virtual Economy 39 minimum wage. This does not account for the fact 6.1.2 Distribution of income that gaming studio workers often enjoy benefits, It was estimated above that the total gross revenues such as lodging and food. Yet, the survey responses of the global gaming services industry were approxi- indicate that perceived problems concerning mately $3.0 billion in 2009. A key question from a playborers’ earnings levels are real and warrant development perspective is how this income is further attention. distributed between different links in the value chain. If most income is accrued by links outside the developing country or links that employ only highly Being viewed negatively by other online game skilled workers, then the development impact of the players industry is limited. In contrast, if a significant amount of income is accrued in poor areas or by Severity of problem: Small to Moderate links that employ less-educated workers, then the development impact is greater. In previous literature, negative views from other players are sometimes put forward as an adverse There are several ways to measure the income of a working condition for gaming studio workers link. One is to measure the value it adds to the (Nardi & Ming Kow 2010). The activities of gaming chain: the price of its outputs to the next link minus studio workers are seen in unfavorable light by many the price of its inputs from the previous link. Using other players in the online gaming world. This is this approach, this report can examine what share of probably due to the fact that in many cases gaming revenues obtained from the final customer is accrued studio workers disrupt the gaming experience of by each link. Figure 6 and Figure 7 show estimated other players, monopolize virtual resources, and are revenue shares of each link in the Chinese powerlev- responsible in producing a service that many players eling and gold farming industries based on informa- regard as cheating. The present survey suggests that tion obtained from our informants. These should these views are not a major problem for the playbor- ers themselves, however. Nardi and Ming Kow argue that the negative views Western players in particular are based not only on Figure 6. Revenue shares in the actual disruptions to gaming experience, but also on Chinese powerleveling industry a bias stemming from a tendency to contrast the “notion of clean/orderly/high-tech/high-culture Game operator <1% EuroAmerican societies through a depiction of their putative opposites constructed in the imaginary of the Chinese gold farmer� (2010). In other words, Producer 70% they connect the issue to cultural prejudice and racism. It is true that popular understanding of Retailer 30% playborer activity is based on very limited informa- tion. Studies such as the present survey help paint a Customer more realistic picture of playborer activity, in good and in bad. Source : Author’s elaboration. Figure 7. Revenue shares in the Chinese gold farming industry Game operator <1% Producer 60–70% Retailer 30% Customer Black accounts <1% Wholesaler 5–10% Source : Author’s elaboration. 40 Knowledge Map of the Virtual Economy only be taken as rough estimates. Gilmore (2009) Consequently, gaming studios operate with very low presented somewhat different estimates; the profit margins, certainly lower than retailers. Gaming proportions are likely to fluctuate according to studios also pay much lower wages to their workers changes in supply and demand. than retailers. If income distribution by profitability and wages is measured, then the development impact The figures suggest that the majority of revenues in of the industry seems more limited than what the the third-party gaming services industry are accrued raw revenue shares suggested. The retail companies by the producers. If the revenue share of the and their university-educated workers occupy the producers is 70 percent, then the total gross most profitable positions in the industry. revenues of the gaming studios were approximately $2.1 billion in 2009. In the powerleveling industry, The survey asked about the general operational costs the producers are gaming studios that employ game of gaming studios (Table 16). The question received laborers. In the gold farming industry, game laborers few responses (N = 3), probably because most carried out approximately one third of the produc- ordinary workers are not in a position to have such tion. This analysis suggests that the gaming studio information. Nevertheless, the data gives some industry has significant development impact by insight into what the operational costs might look directing revenues to functions that provide like for a gaming studio. Only small gaming studios employment to low-skilled workers. (10–20 workers) answered this question. Operational costs for larger establishments might be different. 6.1.3 Costs and pro�tability Workers’ wages are the largest operational cost for The revenue share analysis presented above provides the respondents, representing approximately one only a partial view of the income distribution in the third of total costs. Other operational costs includ- value chain. It does not take into account differences ing technological infrastructure, physical infrastruc- in the cost structures of different links. The infor- ture, and general administrative costs take up 15–20 mants suggest that although the gaming studios percent each. capture the largest share of gross revenues in the powerleveling industry, the gaming studios also have the highest operating costs. Table 15 illustrates the 6.1.4 Number of people employed approximate monthly operating costs of a typical The number of people employed by the industry is medium-sized gaming studio in suburban China. of direct development interest. Earlier employment TABLE 15. Monthly operating costs of a medium-sized gaming studio in suburban China Type of operating cost Cost in RMB Cost in dollars 50–80 m of�ce space and sleeping quarters, rent 2 ¥1,000–2,000 $150–300 Depreciation of 10 computers (purchase price 2,000–3,000 RMB ($303–455) ¥550–830 $80–125 each, economic lifetime 3 years) Internet connection ¥1,000–2,000 $150–300 Electricity ¥1,000–2,000 $150–300 Wages for 20 game laborers (800–2300 RMB ($121–349) per month each) ¥16,000–46,000 $2,400–7,000 Food for 20 workers ¥9,000–18,000 $1,350–2,700 Total ¥28,550–70,830 $4,280–10,725 Sources: Heeks (2008), Authors’ survey and expert interviews Development Potential of the Virtual Economy 41 TABLE 16. Operational cost structure of surveyed gaming studios Company 1 Company 2 Company 3 Type of cost (20 workers) (10 workers) (10 workers) Average Technological infrastructure (PCs, broadband, etc.): 20% 15% 15% 17% Physical infrastructure (land, housing etc.) 15% 15% 20% 17% Workers salaries 40% 20% 35% 32% General administrative and overhead cost 5% 15% 30% 17% N=3 Source: Authors’ survey estimates range from 100,000 to one million (Heeks 2008); however, the methods used to generate these TABLE 17. Estimate of spending on estimates are not documented and they are probably wages in the third-party gaming simply educated guesses. Using the figures presented services industry in 2009 in this report, it is possible to calculate a reasoned (millions) estimate, although the accuracy of the estimate is still limited by the significant uncertainties in the base figures. Gross revenues $3,023 Gross revenues from currency and items (85%) $2,570 Based on the figures presented in the sections above, it is possible to estimate how much of the gross Producers’ share of revenues (65%) $1,670 revenues of the gaming services industry are spent Manual gold farm revenues (30%) $501 on game laborers’ wages. The calculations presented in Table 17 indicate that approximately $300 Amount spent on game laborers’ wages (33%) $165 million were spent on such wages in 2009. By Bot farm revenues (50%) $835 dividing this number by one laborer’s yearly wages, it is possible to estimate how many full-time Amount spent on game laborers’ wages (3%) $25 equivalent game laborers were employed by the industry. Table 18 suggests that the number or game Hacker groups’ criminal revenues (20%) $334 laborers is approximately 100,000, using Chinese Gross revenues from powerleveling (15%) $453 wages. In addition, the retail and management functions of the industry directly employ perhaps a Producers’ share of revenues (70%) $317 few tens of thousands of skilled workers (customer Manual powerleveling (100%) $317 service staff, studio managers, IT support, and accounting). Amount spent on game laborers’ wages (33%) $105 Total spent on game laborers’ wages: $295 6.1.5 Competition, entry barriers, and Source: Authors’ analysis changing business conditions To understand why different links in the value chain differ in their profitability and wage-paying ability, entry barriers in the different sectors of the industry market and pushing prices downward. If a company are examined. Economic theory suggests that a manages to sustain significant above average company enjoying high profits should soon find its profitability over time, this suggests that there are profitability eroded by competitors entering the some kind of barriers that prevent new companies 42 Knowledge Map of the Virtual Economy TABLE 18. Estimate of the number of game laborers employed in 2009 Laborers Monthly wages Monthly wages Yearly wages Yearly wages employed (RMB) (dollars) (RMB) (dollars) full-time Upper limit wage estimate ¥2,300 $349 ¥27,600 $4,181 70,587 Lower limit wage estimate ¥800 $121 ¥9,600 $1,454 202,936 Average wage estimate ¥1,550 $235 ¥18,600 $2,818 104,741 Source: Authors’ analysis from entering the market. Typical entry barriers are limited competition and higher profitability. Their high capital requirements and rare skills and workers likewise command higher wages due to technologies that not everyone has access to. possessing special skills that not everyone in the job market has. From the available data, the report can identify at least the following barriers of entry to the retail In contrast, entry barriers to the production sector sector of the Chinese gaming studio industry: are low. A one-man game studio can be set up in a cyber café with rental equipment. Large retailers 1. Familiarity with online games, gaming services, maintain “buy sites� where anyone can become a and their associated terminology and culture. supplier to the retailer. The only skills required are 2. Communication and language skills. When gaming skills, and approximately 66 million Chinese selling to foreign customers, it is necessary to be played online games in 2009. Consequently, able to provide foreign-language customer competition in the production sector is fierce, and service through instant messaging and in some profits and wages are low. cases over voice communications. 3. Marketing skills. Typical advertising channels for acquiring new customers are search engine 6.1.6 Changes in business and keyword advertising and in-game advertising. Especially the former requires specialized skills. market conditions over time 4. Establishment of customer relationships. A Industry expert interviews and some recent reports Google AdWords advertisement under the gave the impression that the gaming studio business keyword “wow gold� can cost $6–8 per click for and market conditions had changed over time. The the retailer. Because of the high cost of advertis- survey asked respondent gaming studios to compare ing, retailers typically take a loss on the first sale their current business situation with the situation to a new customer. They rely on repeat sales to one year ago, requesting them to grade how different their existing customer base for profits. aspects had changed over time. 5. Ability to receive online payments from abroad. 6. Some business skills. For example, understand- In the survey, 12 gaming studios responded, giving a ing exchange rate risk. fairly good general picture. While most issues had 7. Some returns to scale may also be present, remained roughly the same compared with the because our informants suggest that large situation for one year ago, the survey indicated a retailers control the majority of the market. slight increase in the following areas: Returns to scale act as a barrier for small entrants. ■ Skills and education of available workers ■ Competition from other gaming studios The barriers listed above will certainly limit viable ■ The use of automated bot programs in the online entries to the retail sector. Consequently, compa- games nies that do possess the required resources enjoy ■ Access to financing Development Potential of the Virtual Economy 43 6.1.7 Upgrading strategies thus far seen widespread adoption and existing Upgrading strategies are ways in which actors in the efforts have been only moderately successful. value chain can increase the scope and value added of their activity, and thus potentially increase the Another upgrading strategy available to retailers is value captured. In line with the scope of this report, inter-chain upgrading, where accumulated logistics the focus of this section is primarily on upgrading and service skills are used to diversify into other strategies for gaming studios, but upgrading products and services. For example, many retailers strategies for other actors in the value chain are have started to sell game time codes (codes needed discussed. Different forms of upgrading strategies to play a game that charges for time played) besides have been identified in the literature; this report actual virtual goods. As new forms online entertain- uses a framework of six upgrading strategies in ment emerge, more such upgrading opportunities developing countries adapted from Mitchell, Coles might appear. and Kean (2009). In Table 19 these different upgrading strategies are described, with indications In the future, several additional upgrading opportu- of how they have been used among gaming studios, nities might become possible. The growth of online along with an estimate of the potential impact of games on mobile devices might enable new oppor- the upgrading strategy (small, moderate, and tunities for gaming studios with an organization and significant). All upgrading strategies are present to business model different from the current one. some extent among gaming studios. When it comes Online games have also become more integrated to impact, process upgrading seems to be the most with social network platforms such as Facebook. It is significant strategy, followed by horizontal possible that gaming services providers could coordination. functionally expand to provide services for other activities on the platform, such as services that are Besides these upgrading strategies available for related to how people handle their wider social gaming studios, there are also several upgrading network and optimize its potential. The integration opportunities available for other actors in the value of standardized virtual currencies on social network chain. A few of the more significant ones are worth sites (e.g., Facebook Credits) might also enable mentioning. gaming studios to offer products and services directly through the official payment channels of In recent years, some game publishers have assumed these networks. functions up the virtual goods value chain by starting to sell virtual goods directly to players. This has already had a negative influence on the value captured by gaming studios; our expert interviews 6.2 Microwork indicate that the opportunity for the secondary market to capture value in this market is signifi- 6.2.1 Worker demographics, skills, cantly reduced. The effect differs from game to wages, and career development game, but the influence of this publisher upgrading There are no comprehensive studies of the demo- strategy on the third-party industry is one of its graphics of microworkers, but a partial picture can largest downside risks in the years to come. be constructed by examining the demographics of the workers of individual work aggregators. Ipeirotis At the same time, an upgrading strategy available to (2008, 2010) has conducted surveys to examine the third-party retailers is to use their customer base and workers at Amazon Mechanical Turk, colloquially experience to integrate downwards into operating known as “turkers�. In 2008, 76 percent of the online games themselves. For example, Emperor respondents were from the United States and 8 Online is an online game launched in South Korea percent from India. In 2010, the proportion of U.S. in 2010 by IMI, a company behind Itemmania, the respondents has fallen to 47 percent and Indians country’s leading third-party virtual goods trading have increased to 34 percent. The remaining 19 platform. There are also several examples of game percent are from 66 different countries. This change point card distributors in China purchasing probably reflects the fact that in 2008, Amazon only licensing rights to online games and starting to channeled payments to U.S. bank accounts, whereas operate them. However, none of these cases have it recently also started to channel payments to India. 44 Knowledge Map of the Virtual Economy TABLE 19. Upgrading strategies for gaming studios Estimate of impact (small, Type of upgrading strategy Indications of use among gaming studios moderate, signi�cant) Horizontal coordination – Different Some reports of gaming studios collaborating with other Moderate to signi�cant. There forms of coordination and networking gaming studios in the production and sales of virtual are large potential upgrading horizontally in the value chain currencies and services. Seems to be more common bene�ts, but they are dif�cult to among smaller �rms. The horizontal coordination enables estimate with present data. them to have a better bargaining position with retailers and to reap some economies of scale. Horizontally integrated gaming studios can provide better volume and responsiveness, i.e. take up quality control functions from wholesalers and retailers, and thereby capture more value. Vertical coordination – Different forms of Some sources indicate that gaming studios have ex- Moderate coordination and vertical networking in tremely fluid relationships with wholesalers and retailers, the value chain while some indicate that there is also coordination and long-term vertical relationship building. Functional upgrading – Increase in the Few reports of gaming studios expanding into retail; Moderate scope of activities performed by actors some examples of gaming studios trying to skip the in the value chain intermediate step and sell directly to customers. Interviews indicate that this has been dif�cult to pursue as producers lack the language, logistics, and service resources of the larger retailers. Many gaming studios have expanded their range of offerings, e.g., from only virtual currency acquisition and sales to powerleveling services. However, survey indicates that most gaming studios are functionally specialized. Process upgrading – Production process The use of macros and bot programs is a form of process Signi�cant improvements and innovation that upgrading that has had a signi�cant impact on the pro- increases volume or reduces cost ducers’ productivity. However, this development has also had negative influence on employment opportunities in the gaming service sector. Research on the most effective playstyles and pathways is also common among produc- ers. Powerleveling services are less affected by process upgrading than virtual currency production. Product upgrading – Improvement in the Relatively few. Virtual currencies are rather standardized. Small product or service quality Powerleveling is a service with variable quality (time to achieve a certain level, reliability of service, etc.). Inter-chain upgrading – Use of skills On the company level, few reports of the use of skills and Moderate and experience developed in one value experience developed in gaming services to engage in chain to engage with another, more other electronic trade, including international trade pro�table value chain Several signs of upgrading at the individual level with some gaming studio workers aiming to use their skills to enter other business sectors Source: Authors’ elaboration Other aggregators reach workers in other countries. There are also interesting differences between the For example, Samasource currently has workers in countries. U.S. turkers are two-thirds female, while India, Pakistan, Haiti, Uganda, and Kenya, and Indians are two-thirds male. In India, the majority is plans to move to South Africa. single, while in the United States the majority is married. When asked about their motivations for Ipeirotis’s surveys suggest that both U.S. and Indian microwork, U.S. turkers are more likely than turkers are somewhat younger and poorer than the Indians to say that it is a way to kill time and obtain general population, yet better educated. The supplementary income, while more Indians say that education may simply reflect their younger age. it is their primary source of income. A typical U.S. Development Potential of the Virtual Economy 45 turker is thus perhaps a housewive killing time in the long term be detrimental for the workers’ online, while a typical Indian turker might be a psyche and motivation. Companies in the field are student or other bachelor relying on microwork for acutely aware of this issue. A basic solution is to their livelihood. make work less monotonous by alternating the types of tasks presented to workers. A more advanced idea The skill requirements in microwork tend to be low pursued by at least Microtask Ltd is to use game (Eagle 2009). At the most basic level, microworkers mechanics to turn sets of microtasks into more only need the ability to use a mouse, a touch screen, engaging constructions. If successful, this “gamifica- or other pointing device in order to, for example, tion� would turn work into something resembling an point to geographical features in an historical map exciting game. It could also introduce a social archive that is being digitized—an actual project element into microwork, through such constructions currently undertaken by Microtask Ltd. In prin- as high score lists, competitions, and collaborative ciple, not even literacy is necessary for such tasks, game modes. In some ways, this is what companies although in practice, the user interface may require such as Gambit and TrialPay are already doing by it. Some tasks completed over mobile phones channeling microtasks to actual online gamers. It require only the ability to speak and listen, such as seems that if the gaming element in the work is providing feedback to a speech recognition algo- exciting enough, no actual monetary payment is rithm that is being tested. In practice, such tasks necessary to get workers to contribute. How far this may be scarce, however. Assignments requiring is true is an unknown factor with important some level of proficiency in the English language are ramifications for the future of the industry. probably most common today. For example, CrowdFlower has dealt with assignments requiring English, French, German, Thai, Spanish, and 6.2.2 Distribution of income Portuguese. It was estimated above that the microwork industry has the potential to reach multi-billion dollar Microwork, like all business process outsourcing, revenues within the next few years. As with the aims to deliver value by exploiting local differences gaming services industry, it is important from a in labor costs and the availability of suitably skilled development perspective to assess how these labor. The emphasis in microwork tends to be on revenues are distributed among different links in the cost rather than skills. In principle, there could be a value chain. This is especially important in micro- microwork assignment that consists of specialized work, as the top links in the value chain are mostly tasks that only persons trained in a specific disci- located in developed countries. pline, such as law, can complete. Among popula- tions with rising education standards, such an Figure 8 shows the approximate revenue share application is not unimaginable. For now, however, captured by each link from revenues obtained from all the current microwork vendors focus on assign- a final customer. The estimates are based on ments with very low skill requirements. information obtained from the informants. At best, a microworker can hope to capture as much as 70 Like Taylorism and scientific management in the 19th percent of what the client pays. But in some cases, century, microwork is prone to criticisms relating to the microworker is not paid any money at all; de-skilling and commodification of work. Basic instead, the worker is compensated with virtual computer skills are a very valuable skill set to learn, currency in an online game. but beyond that, it is difficult to see how experience in extremely simple microtasks could develop skills As most of the companies in the industry are still that could be used in higher value added tasks. very young and growing, there are wide discrepan- Companies in the field are aware of this issue, and at cies in income distributions from one company and least Txteagle has promised to offer workers some project to another. As the industry matures, kind of a career development path. From a develop- competition can be expected to cause the distribu- ment perspective, this is a key issue to follow. tions to converge. The current percentages and cost structures are not very useful as indicators of what A somewhat related issue is that like repetitive such mature distribution might look like. A better manufacturing work, monotonous microwork may idea of the different sectors’ future outlook may be 46 Knowledge Map of the Virtual Economy Figure 8. Typical revenue shares in the microwork industry Infrastructure provider Microworker Aggregator Transformer Client <30% 0–70% 10–30% 20–60% Game operator 5–15% Source : Authors’ analysis. obtained by analyzing their competition and entry transformers is to create work that can be addressed barriers. by almost anyone. Basic ICT infrastructure and access to a work aggregator are the main external barriers. Suitable skills, discussed above, are the 6.2.3 Competition, entry barriers, and main internal barrier. Some niche types of work, changing business conditions such as tasks involving a language spoken by few Analyzing what barriers, if any, prevent new actors people, have higher entry barriers, but such jobs are from entering different sectors of the microwork also probably correspondingly fewer. industry allows us to assess how fierce competition is likely to be in those sectors. From a development Based on this tentative analysis, it seems that perspective, an ideal sector has sufficiently low transformers probably enjoy the biggest economic barriers to allow poor and uneducated people to rents in the industry, while workers are likely to face participate. However, if the barriers are too low, extreme price competition that pushes their wages competition may lead to a “race to the bottom�, down. In the next section, the report considers which in the worst case results in wages falling to a whether workers are able to improve their position level that no longer contribute to economic by using potential upgrading strategies. development. Becoming a transformer that accepts assignments 6.2.4 Upgrading strategies from clients and transforms them into microtasks Upgrading strategies are ways in which an actor in a involves several significant barriers. Transformers value chain can increase the scope and value added need to have significant technology and innovation of their activity, and thus improve their profitability. capabilities in order to solve the technical problem For example, an individual microworker could of transforming mundane tasks into microwork. invest in their own computer hardware to avoid This is probably the main bottleneck in the whole having to pay rent to a cybercáfe owner, thus industry. Moreover, transformers need business-to- capturing a larger segment of the value chain. Other business sales and marketing capabilities, which than this, however, any actual cases of upgrading on means that they are often located near their clients the part of microworkers are not known as of yet. in developed countries. Some speculations are offered below. Aggregators’ main entry barriers are access to An upgrading strategy that agricultural producers sufficient numbers of workers and access to payment have used and microworkers could perhaps benefit infrastructure that reaches the workers. Too small a from is horizontal coordination: forming a group or number of workers are not marketable to transform- collective that acts as a single agent in the market. In ers and clients. They also need a level of technologi- theory, this increases the producers’ bargaining cal proficiency to be able to distribute microtasks to power towards their buyers. Coordination could also their workers. Higher value-added aggregators also allow microworkers to take up some of the quality have skills in quality control. and volume assurance functions from their aggrega- tor, or even supplant their aggregator entirely. This Workers themselves have very few entry barriers. would again allow them to capture a larger segment Indeed, one of the main aims of microwork of the value chain. Development Potential of the Virtual Economy 47 In the future, the commodification of transformer 6.3.2 Microcontent production and technology may allow aggregators to expand into the other two-sided marketplaces transformer role and accept assignments directly Tools for designing virtual goods for online from clients. Transformers and ICT infrastructure hangouts and small games and applications for providers might seek to expand into the workers’ social networking sites and mobile phones are territory by integrating increasingly intelligent constantly getting easier to use. Operators of features into the technology, reducing the functions two-sided marketplaces such as the Second Life for which human intelligence is still needed. The XStreet and iPhone App Store are also competing in complete replacement of human workers still making their markets increasingly easy to enter and remains science fiction, however. earn revenues on for developers. Nevertheless, these earning opportunities still require advanced skills such as programming and graphic design to make use of, which severely limits their development 6.3 Other segments potential among the world’s poorest countries and populations. In this sense, virtual goods design and 6.3.1 Cherry blossoming application development remain part of the Online marketing is a large and continuously traditional digital economy rather than the “virtual growing market, so in principle, cherry blossoming economy� of scarcities exploitable almost with time could be involving an increasing number of people. and effort only. Indeed, most of the producers in But if manual gaming services producers are in the these industries are located in the Western world, in grey area, cherry blossomers are unambiguously contrast to the gaming services and microwork unethical. Moreover, since one Facebook or other industries. For these reasons, the development account holder can only recommend one brand potential of these areas is not analyzed further in once, and there are a limited number of brands, this report. cherry blossoming is not likely to provide significant income even to users in developing countries. For these reasons, its development potential is not analyzed further in this report. 48 Knowledge Map of the Virtual Economy Chapter 7 Conclusions 7.1 Market opportunities services sold by the game operators themselves. Market research firm In-Stat, estimates that the latter figure is $7.3 billion in 2010 (Reisinger 2010). The “digital economy� of online services and abundant digital contents that developed in the late 1990s has given rise to new kinds of entrepreneur- In 2007, Lehtiniemi and Lehdonvirta estimated the ship and earning opportunities. However, exploiting global virtual goods market size, including sales by these opportunities has required advanced skills and game operators (primary market) as well as sanc- technologies that place them outside the reach of tioned and unsanctioned third parties (secondary developing countries and poor and uneducated market), at approximately $2.1 billion. Compared people. More recently, a “virtual economy� of digital to this, it seems that the market opportunity for scarcities that do not require advanced skills to gaming services has grown significantly. The whole exploit has started to emerge from digital services. online gaming market has grown rapidly during this Basic ICT infrastructure and the expenditure of period, and gaming has expanded to new platforms time and effort rather than professional skills are such as social networking sites and mobile phones. often sufficient to turn these scarcities into income. Commercially active areas of the virtual economy At the same time, there are significant uncertainties today include various kinds of gaming services, in the future market opportunity for third-party microwork, and small marketing related tasks, such online gaming services. The most rapidly growing as “liking� a brand’s Facebook page for a small areas of digital gaming are online social games, payment. casual games, and mobile games. In-Stat suggests that these account for the majority of the primary It is difficult and not necessarily meaningful to market virtual goods revenues today. In contrast, the attempt to examine the economic impact of this third-party gaming services industry’s revenues so-called virtual economy as a whole. This report continue to be generated mostly from the MMOG focused on two concrete areas, gaming services and segment, and the genre’s megahit title, World of microwork. The choice of focus was based on Warcraft. It seems that the industry has so far been apparent commercial significance and development unable to establish any significant business in social potential. These two areas are in themselves large games and mobile games, although some trade is and diverse. Gaming services consists of such known to have taken place. Little is known about concrete activities as harvesting virtual items and the third-party gaming services industry’s R&D currencies in online games and acting as a hired efforts in this area. pilot for online game characters in place of their owner. Microwork ranges from simple data input The microwork industry is a much more recent and pattern recognition tasks to content moderation addition to the virtual economy. Task marketplaces and search engine evaluation. such as Amazon Mechanical Turk have existed for over five years, but the “transformer� companies that The gaming services industry’s revenues were convert client work into microtasks, and thus feed estimated in this report as $3.0 billion per year. This the markets with new work, have only existed for a figure is the amount of money spent by online few years. The number of dollars spent on micro- gamers on third-party, sanctioned, and more often work services is probably no more than double digit unsanctioned virtual goods and services globally. millions in 2010, although there is no clear-cut line The figure does not include the amount of money between microwork and more complex digitally consumers spent on virtual goods and value-added distributed work tasks. Based on market size figures from neighboring markets as well as microwork’s chains. The revenue share analysis indicates how potential to expand the existing markets by making much of the total revenues of the chains are previously unfeasible tasks economically feasible, captured by links in developing countries. Entry this report concluded that the market could be barrier analysis furthermore indicates how profitable worth billions of dollars within the next five years, as the links are likely to be. Second, the report looked the technology matures. Still, there is currently little at upgrading strategies, or possibilities for producers knowledge about how much work there is that will to move up in the value chain and increase their turn out to be addressable by unskilled microwork- scope and value added in the future. ers; for example, how much of traditional BPO involves simple data entry tasks. The findings suggest that both in the gaming services as well as the microwork industry, producers can Ultimately, the growth and sustainability of the capture as much as 70 percent of the revenues virtual economy’s market opportunities depends on obtained from end customers. Compared to the the sustainability of the scarcities on which they are global coffee industry, the virtual economy thus seems based. The findings suggest that the scarcity that to have a much better development impact. However, maintains the gaming services industry is threatened analyses of entry barriers suggest that it is very easy to from multiple directions, by automatization as well become a producer in these industries, indicating that as by the game operators themselves. Microwork is price competition is likely to be fierce. All this based on the natural scarcity of effort, and therefore suggests that even though the producers’ revenue seems to be on stronger footing. But in the long shares are high, their actual profit margins and run, repetitive tasks that need to be assigned to workers’ incomes are very low, and likely to remain so humans will grow fewer as artificial intelligence even as the markets expand. Market expansion will develops. The microwork industry must therefore provide earning opportunities for a larger number of continuously innovate to expand the scope of workers, but the incomes of individual workers might microwork to new sectors. In contrast, some game not grow, and in the worst case might even fall. This industry experts suggest that games will always need situation is sometimes known as “immiserising to have some degree of artificial scarcity designed growth� (Kaplinsky & Morris 2001). These conclu- into them, in order to provide challenge and reward. sions are rather tentative, however, and the situation Even if game publishers increasingly sell out parts of may be different in specific market niches. this scarcity to generate revenues, a hard core must always be retained for a game to feel like a game. The tentative analyses of potential upgrading strategies suggest that in both industries, producers can benefit from horizontal coordination, which in practice means that individual workers and produc- 7.2 Development potential tion companies form larger collectives. This can improve their bargaining position towards buyers, The market size figures presented in the previous and allow them to expand both upwards as well as section could be compared with, for example, the downwards in the value chain. One method of size of the global coffee market, which many downward expansion would be that producers invest developing countries are highly dependent on. The in their own computer infrastructure instead of coffee market is worth approximately $70 billion per renting it. Upward expansion entails taking up year. However, of the $70 billion, the developing additional value-adding functions, such as volume countries that produce the coffee beans capture only guarantees and quality control. For example, a large $5.5 billion (Osorio 2002). The market potential of and well-coordinated group of microworkers could the virtual economy likewise equals development become their own work aggregator, selling work potential only to the extent to which gains from the directly to transformers. However, our findings market are successfully captured in developing suggest that producers’ upward expansion ability is countries and eventually reach workers. The value limited by at least skills and education, particularly chain analyses presented in this report examined this language and business skills. issue from two perspectives. First, this report looked at the revenue shares and entry barriers of different The more mature gaming services industry shows a links in the gaming services and microwork value higher level of horizontal coordination between 50 Knowledge Map of the Virtual Economy producers than microwork. Instead of being However, in the virtual economy, scarcities can be individual contractors, game workers are often eradicated as well as entirely new ones created, as banded together in gaming studios of several dozen technology advances. This demands flexibility from workers, headed by an ex-worker. Notably, interme- intervention programs. diary roles between producers and customers have also been gradually taken over by educated customer service workers in China and the Philippines instead 7.3.1 Third-party gaming services of the West. What a decade ago was entirely a Third-party gaming services may seem like a good developed country phenomenon is now run from target for interventions, given that it is a large and developing countries. The microwork industry could successful industry. Gaming service production follow a similar pattern where higher value added could be introduced to least-developed countries. functions gradually move to developing countries. Ways to expand the market to social games, mobile games, and other platforms could be explored as a In summary, our value chain analyses suggest that means to further increase its economic development the most profitable positions in the virtual economy impact. However, the problem with services for are those closest to the customers. Producers are not overcoming artificial scarcities is that it can be very in a good position to increase their profits in the difficult to say when their net social value is positive future, although their business helps to fund the and when it is negative. Whilst providing value for development of local ICT infrastructure. A better their customers, in some cases they are causing prospect for producers can be found in upgrading negative externalities to other players and publishers. their business to encompass a larger scope of value This is reflected in the industry’s legal status. added. Not much is known about how this happens Sanctioned markets for game assets are limited. in practice in the virtual economy and what stumbling blocks it might involve. At the same time, one clearly positive thing about the gaming services industry is that it has activated thousands of young people from very modest backgrounds to create employment for themselves as 7.3 Scope for interventions digital entrepreneurs. Some of them have subse- quently branched out to other fields of ecommerce. One of the goals of development studies is to A more typical career path for these youth would identify ways in which social and economic develop- likely have been much less entrepreneurial and ment can be supported, enhanced, and accelerated. ambitious. If the circumstances and motivations that These can take the form of interventions to remove lead to this remarkable mobilization were under- obstacles and bottlenecks, disseminate knowledge, stood better, it could prove a powerful tool in efforts and develop capacities. This final section draws to bring about similar activation elsewhere. conclusions concerning the scope for possible Informants suggest that reasons for why Chinese interventions in the virtual economy. youth—as opposed to, for example, Indian youth— rose to become gaming services entrepreneurs The distinguishing feature of the virtual economy is should be sought in the strong online gaming that it is based around digital scarcities that require subculture in China. Indeed, cultural factors and time and effort as opposed to advanced skills and subcultures in particular have been implicated in resources to exploit commercially. This affords a studies on high-tech entrepreneurship before comparative advantage to developing countries with (Florida 2002). The instrumental role of cultural low labor costs. Within the virtual economy, products such as games in forming this culture is a different sectors have reached different stages in worthwhile research question. capturing the economic potential of these scarcities. One implication is that interventions should focus on sectors with market growth opportunities that 7.3.2 Microwork are still emerging, and hence retain the plasticity for When the question is about interventions that intervention programs to have a significant impact. directly leverage the virtual economy to bring about From this perspective, the microwork industry is development results, it is safest to limit the scope to currently the most promising area for intervention. sectors that are based on natural scarcities. In the Conclusions 51 microwork sector, interventions could be aimed at could help developing country actors capture a either the demand side (enabling more demand) or larger part of the microwork value chain for the supply side (how to address the demand from themselves, or 2) develop higher value-added developing countries). As the market grows, services to supply to the chain (e.g., business interventions could also be aimed at helping intelligence). One possible enabler in capturing a developing country actors capture a larger share of larger part of the value chain could be open source the value chain, as they have successfully done in the software that can be used to perform simple quality gaming services industry. control and transformation tasks. In the context of least-developed countries, such software should In the demand side, a handful of transformers are probably take the form of a mobile application. working on the problem of expanding the scope of digital problems that can be addressed through Besides the existing forms of microwork, future microwork. Their success will have a significant services for overcoming natural scarcities in the impact on the demand for simple microwork. These virtual economy could also be considered. There are are for-profit start-up companies such as Microtask currently few services aimed at individual consumers and CrowdFlower. They have been able to raise as opposed to corporations, but these could include, funding from private capital markets for this for example, services for organizing and labeling purpose. What obstacles remain and how this personal photo albums and other personal data process could possibly be accelerated could be collections, which are good growth markets because worked out with the practitioners. new digital devices are being constantly adopted. The main bottleneck is again the transformation In the supply side, there are various hurdles of step, which requires new technology. In the gaming varying height in connecting producers in develop- space, it is possible to imagine, for example, a ing countries to the global demand for microwork. publisher using digitally enabled blue-collar workers Depending on country, these include at least typical to replace poorly performing computer controlled factors such as funding, regulation, financial characters in order to improve the gaming experi- infrastructure, technical infrastructure, and training. ence. One intervention could simply be to open a The need for standardization might also be consid- dialogue between the development community and ered. These issues could be further explored with the digital services industry on ideas for future practitioners. An additional area to consider are services that provide real jobs in the virtual upgrading strategies: 1) what kinds of interventions economy. 52 Knowledge Map of the Virtual Economy References Blaszczynski A., Ladouceur R., Goulet A. and Act in Cyberspace. Alabama Law Review 60(5): Savard C. (2008) Differences in monthly versus 1077–1110. daily evaluations of money spent on gambling and calculation strategies. Journal of Gambling Issues 21: CNNIC (2009a). 2008 98–105. [Chinese Online Game Market Research Report 2008]. China Internet Network Information Bloomberg (2010) iPhone Game Apps Let Center: Beijing. Jumping Doodles Stomp ‘Street Fighter’. June 11. http://www.bloomberg.com/news/2010-06-11/ CNNIC (2009b). 2009 iphone-makes-game-market-a-bazaar-where-doodle- [Chinese Online Game Market Research jump-beats-fighter-iv-.html (Accessed March 15, Report 2009]. China Internet Network Information 2011) Center: Beijing. Castronova, E. (2004) The Right To Play. New York Denegri-Knott, J. and Molesworth, M. (2010) Law School Law Review 49: 185–210. Concepts and practices of digital virtual consump- tion. Consumption Markets & Culture 13(2): Castronova, E. (2005) Synthetic Worlds: The Business 109–132. and Culture of Online Games. University of Chicago Press: Chicago. DFC Intelligence (2008) The Online Game Market. DFC Intelligence: San Diego. Castronova, E. (2006a) Virtual Worlds: A First- Hand Account of Market and Society on the DFC Intelligence (2010). Virtual property and real Cyberian Frontier. In: K. Salen and E. Zimmerman money trade: A business and legal survey. DFC (eds), The Game Design Reader: A Rules of Play Intelligence: San Diego. Anthology, pp. 814–863. The MIT Press: Cambridge, MA. Dibbell, J. (2006) Play Money: Or, How I Quit My Day Job and Made Millions Trading Virtual Loot. Castronova, E. (2006b) A cost-benefit analysis of Basic Books: New York. real-money trade in the products of synthetic economies. Info 8(6): 51–68. Duranske, B. T. (2008) Virtual Law: Navigating the Legal Landscape of Virtual Worlds. American Bar Chan, D. (2006) Negotiating intra-Asian games Association: Chicago. networks: on cultural proximity, East Asian games design, and Chinese farmers. FibreCulture 8. http:// E27 (2010) The Top 50 Apps in Asia: Results, journal.fibreculture.org/issue8/issue8_chan.html September 2. http://www.e27.sg/2010/09/02/ (Accessed March 15, 2011) the-top-50-apps-in-asia-results/ (Accessed March 15, 2011) Cherry, M. A. (2010) A Taxonomy of Virtual Work. Georgia Law Review (forthcoming). http://ssrn.com/ Ernkvist, M., and Strom, P. (2008) Enmeshed in abstract=1649055 (Accessed March 15, 2011) Games with the Government: Governmental Policies and the Development of the Chinese Cherry, M. A. (2009) Working For (Virtually) Online Game Industry. Games and Culture 3: Minimum Wage: Applying the Fair Labor Standards 98–126. Evans, P. (2007) A Silicon Silicon Valley? Virtual Hunter, D. (2006). The early history of real money Innovation and Virtual Geography. Innovations: trades. Terra Nova, Jan 13. http://terranova.blogs.com/ Technology, Governance, Globalization 2(3): terra_nova/2006/01/the_early_histo.html (Accessed 55–61. March 15, 2011) Fairfield, J. (2005) Virtual property. Boston IDC and CGPA (2004–2009) China Game Industry University Law Review 85(4): 1047–1102. Report. IDC and China Game Publishers Association: Beijing. Florida, R. (2002) The Rise of the Creative Class. Basic Books: New York. IDC (2007) Asia/Pacific (Excluding Japan) Online Gaming 2006–2010 Forecast and Analysis. IDC Frei, B. (2009) Paid Crowdsourcing: Current State Group. & Progress toward Mainstream Business Use. Smartsheet. http://bit.ly/smartsheet_report (Accessed IDC (2010) Asia/Pacific (Excluding Japan) Online March 15, 2011) Gaming 2009–2013 Forecast and Analysis. IDC Group. Gilmore, A. (2009) China’s New Gold Farm. Journal of Virtual Worlds Research 2(4). http:// Ipeirotis, P. (2008) Mechanical Turk: The journals.tdl.org/jvwr/article/view/863 (Accessed Demographics. http://behind-the-enemy-lines.blogspot. March 15, 2011) com/2008/03/mechanical-turk-demographics.html (Accessed March 15, 2011) Golub, A., and Lingley, K. (2008) “Just like the Qing Empire�: Internet addiction, MMOGs and Ipeirotis, P. (2010) The New Demographics of moral crisis in contemporary China. Games and Mechanical Turk. http://behind-the-enemy-lines. Culture 3(1): 59–75. blogspot.com/2010/03/new-demographics-of-mechani- cal-turk.html (Accessed March 15, 2011) Heeks, R. (2008) Current Analysis and Future Research Agenda on ‘gold farming’: Real-World Jimenez, C. (2007) The high cost of playing Production in Developing Countries for the Virtual Warcraft. BBC News Online, Sep 24. http://news.bbc. Economies of Online Games. University of co.uk/2/hi/technology/7007026.stm (Accessed March Manchester Development Informatics Working Papers 15, 2011) no. 32. http://www.sed.manchester.ac.uk/idpm/ research/publications/wp/di/di_wp32.htm (Accessed Kaplinsky, R. and Morris, M. (2001) A Handbook March 15, 2011) for Value Chain Analysis. International Development Research Centre: Ottawa. Heeks, R. (2009) Understanding ‘Gold Farming’ and Real-Money Trading as the Intersection of Real Kieger, S. (2010) An Exploration of and Virtual Economies. Journal of Virtual Worlds Entrepreneurship in Massively Multiplayer Online Research 2(4). http://journals.tdl.org/jvwr/article/ Role-Playing Games: Second Life and Entropia view/868 (Accessed March 15, 2011) Universe. Journal of Virtual Worlds Research 2(4) http://journals.tdl.org/jvwr/article/view/643/623 Howe, J. (2008) Crowdsourcing: Why the Power of the (Accessed March 15, 2011) Crowd Is Driving the Future of Business. Crown Publishing Group: New York. Kleemann, F., Voß, G. G. and Rieder, K. (2008) Un(der)paid Innovators: The Commercial Huhh, J. S. (2005) Empirical Study on the Decline of Utilization of Consumer Work through Lineage 2 in Korea. http://papers.ssrn.com/sol3/papers. Crowdsourcing. Science, Technology & Innovation cfm?abstract_id=833847 (Accessed March 15, 2011) Studies 4(1): 5–26. Huhh, J. S. (2008). Culture and Business of PC KOCCA (2010). 2010 . Bangs in Korea. Games and Culture 3(1): 26–27. [White Paper on Korean Games]. Korea Creative Content Agency: Seoul. 54 Knowledge Map of the Virtual Economy Krebs, B. (2009) The Scrap Value of a Hacked PC. blog/how_big_is_the_rmt_market_anyw (Accessed The Washington Post Security Fix. http://voices. March 15, 2011) washingtonpost.com/securityfix/2009/05/the_scrap_ value_of_a_hacked_pc.html (Accessed March 15, Liang, G. (2007) Surveying Internet Usage and its 2011) Impact in Seven Chinese Cities. Chinese Academy for Social Sciences: Beijing. http://www.policyarchive.org/ Lastowka, G. (2010) Virtual Justice: The New Laws handle/10207/bitstreams/16013.pdf (Accessed March of Online Worlds. Yale University Press: New Haven. 15, 2011) Lehdonvirta, V. (2005) Real-Money Trade of Martin, J. (2008) Consuming Code: Use-Value, Virtual Assets: Ten Different User Perceptions. Exchange-Value, and the Role of Virtual Goods in Proceedings of Digital Arts and Culture (DAC 2005), Second Life. Journal of Virtual Worlds Research 1(2). Copenhagen, Denmark, December 1–3, pp. 52–58. https://journals.tdl.org/jvwr/article/view/300/262 http://www.hiit.fi/~vlehdonv/documents/Lehdonvirta- (Accessed March 15, 2011) 2005-RMT-Perceptions.pdf (Accessed March 15, 2011) Mitchell, J., Coles, C. and Keane, J. (2009) Upgrading along Value Chains: strategies for poverty Lehdonvirta, V. (2008) Real-Money Trade of Virtual reduction in Latin America. COPLA Briefing Paper, Assets: New Strategies for Virtual World Operators. DFID. http://www.odi.org.uk/resources/down- In: M. Ipe (ed.), Virtual Worlds, pp. 113–137. Icfai load/4626.pdf (Accessed March 15, 2011) University Press: Hyderabad. MMOGChart.com (2010). MMOGChart – Lehdonvirta, V. (2009a) Virtual Consumption. Charting the future of the MMOG industry. http:// Publications of the Turku School of Economics www.mmogchart.com/ (Accessed December 1, 2010) (A-11:2009), Turku. http://info.tse.fi/julkaisut/vk/ Ae11_2009.pdf (Accessed March 15, 2011) Nardi, B. and Ming Kow, Y. (2010). Digital imaginaries: how we know what we (think we) know Lehdonvirta, V. (2009b) Virtual Item Sales as a about Chinese gold farming. First Monday, 15(6–7). Revenue Model: Identifying Attributes that Drive http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/ Purchase Decisions. Electronic Commerce Research fm/article/view/3035/2566 (Accessed March 15, 9(1): 97–113. 2011) Lehdonvirta, V. (2010) Online spaces have material Niko Partners (2010) Niko Partners Predicts culture: goodbye to digital post-materialism and Southeast Asia Online Games Market to Reach $1.7 hello to virtual consumption. Media, Culture & Billion by 2014. Niko Partners press release, Nov 10. Society 32(5): 883–889. Ondrejka, C. (2007) Collapsing Geography: Second Lehdonvirta, V. and Virtanen, P. (2010) A New Life, Innovation, and the Future of National Power. Frontier in Digital Content Policy: Case Studies in Innovations: Technology, Governance, Globalization the Regulation of Virtual Goods and Artificial 2(3): 27–54. Scarcity. Policy & Internet 2(3), art. 2. http://www. psocommons.org/policyandinternet/vol2/iss3/art2/ Osorio, N. (2002) The Global Coffee Crisis: A Threat (Accessed March 15, 2011) to Sustainable Development. London: International Coffee Organization. http://www.ico.org/documents/ Lehdonvirta, V., Wilska, T.-A., Johnson, M. (2009) globalcrisise.pdf (Accessed March 15, 2011) Virtual Consumerism: Case Habbo Hotel. Information, Communication & Society 12(7): Park, S. (2010) Ruling to Boost Sale of ‘Cyber 1059–1079. Money’. The Korea Times, 10 January. http://www. koreatimes.co.kr/www/news/na- Lehtiniemi, T. and Lehdonvirta, V. (2007) How big tion/2010/01/116_58775.html (Accessed May 1, is the RMT market anyway? Virtual Economy 2010). Research Network, Mar 2. http://virtual-economy.org/ References 55 Porter, M. E. (1985) Competitive Advantage. The August 5. http://www.strategyanalytics.com/default.asp Free Press: New York. x?mod=ReportAbstractViewer&a0=5689 (Accessed March 15, 2011) Reisinger, D. (2010) Virtual goods revenue to hit $7.3 billion this year. CNET News, 15 November. ThinkEquity (2009) Think Entertainment: Gaming http://news.cnet.com/8301–13506_3-20022780-17. The Emergence Of Games As A Service. ThinkEquity html (Accessed March 15, 2011) LLC: London. Robischon, N. (2007) Station Exchange: Year One. UNCTAD (2010) Information Economy Report http://bit.ly/station_exchange (Accessed March 15, 2010: ICTs, Enterprises and Poverty Alleviation. 2011) Geneva: United Nations. SK Communications (2010) Social Networks. Williams, D. (2003) The video game lightning Presentation, January 28. rod. Information, Communication & Society, 6(4), 523–550. Shafer, I. (2010) Is this the dawn of the Facebook credit economy? Advertising Age, May 20. http:// Yee, N. (2005). The Daedalus Project, vol. 3–5 adage.com/digitalnext/article?article_id=143983 (10/19/2005). http://www.nickyee.com/daedalus/ (Accessed March 15, 2011) archives/pdf/3–5.pdf (Accessed March 15, 2011) Shapiro, C. and Varian, H. R. (1999) Information Yee, N. (2006) The labor of fun: how video games Rules: A Strategic Guide to the Network Economy. blur the boundaries of work and play. Games and Harvard Business School Press: Boston, MA. Culture, vol. 1, no. 1, pp. 68–71. Strategy Analytics (2010) The World of MMORPG: a Tale of Two Regions. Strategy Analytics Reports, 56 Knowledge Map of the Virtual Economy Appendix 1. Sources of uncertainty in the gaming services market estimate The player survey method of estimating the size 3. Different markets for gaming services depend- of the third-party gaming services market suffers ing on type of online game and its business model. from the following sources of uncertainty: Traditionally, MMOG games have more dedicated “hardcore� players than more casual online games 1. Misreporting the use of secondary market with a higher propensity to consume virtual goods channels. As a grey area activity, it is likely that and services. These casual online games have players could underreport the usage of secondary increased their share of the market over time as online market in surveys. Because the usage of secondary games have reached beyond the hardcore market. gaming services might have negative influences on their game accounts (e.g., account banning) if Also, the secondary market is larger in games where discovered by the game’s official operator, some all the gaming services are provided by third parties players might be reluctant to reveal that they have as opposed to the official operator. In so-called used the secondary gaming service market. The free-to-play or item payment based games, the third degree of underreporting might be different depend- party industry faces competition from the official ing on the survey design, survey body, and the region. game operator. This business model is especially common in Asia where it is dominant, but it has 2. Retrospective misreporting of secondary also been growing in importance in North America gaming service expenditures by players. Players and Europe. might retrospectively misreport their use of second- ary gaming in online games. In surveys, players are The result is what could be labeled a “mixed� online often asked to give retrospective accounts of their game market compared to a few years ago, with expenditures for the last month or last year. different business models and a more heterogeneous Hindsight biases and other cognitive factors might player base. The data in this report is focused mostly affect the accuracy of such accounts, creating effects on MMOG games. of either underreporting or over reporting. The exact effect of this in the online game market is unknown 4. Empirical reliability and lack of survey data and needs further study. Underreporting of expendi- regarding average user spending on the secondary tures compared to actual expenditures is most likely. market. The surevy data regarding the number of This assumption is based on the similarities of this paying players and the share of players using phenomena in the online game market compared to third-party gaming services are relatively reliable. gambling markets in which this effect has been However, currently there are few surveys yielding studied more extensively. It might be objected that reliable data regarding the amount of money users the gambling market might be different in some spend on the secondary market. This report used respects that render industry comparison of this governmental surveys from Korea and China and specific effect unreliable. Yet, several studies have extrapolated the spending data from Korea to other found similarities between the design of the virtual developing countries. It is possible that the Korean economy of online games and the mechanisms numbers are not generalizable to other developing underlying gambling. In the gambling literature, countries. Differences in methodology and design one of the more extensive studies of this subject between surveys have some implication on the found that monthly retrospective reporting was degree to which they can be compared. Future underreporting daily reporting of gambling expendi- surveys with more reliable and harmonious design tures that were almost 50% higher (Blaszczynski, and coverage of more geographic regions will Ladouceur, Goulet & Savard, 2008). decrease this source of uncertainty. 58 Knowledge Map of the Virtual Economy Appendix 2. Examples of third-party websites selling virtual game currency Company location (as stated Name of on the webpage, might be company Webpage address different in reality) Koala Credits http://www.koalacredits.com/ (accessed Mar. 4, 2011) Hong Kong Bank of Wow http://www.bankofwow.com/ (accessed Mar. 4, 2011) China IGE http://www.ige.com/ (accessed Mar. 4, 2011) United States Guy4Game http://www.guy4game.com/ (accessed Mar. 4, 2011) Canada, China, United States MOGS http://www.mogs.com/ (accessed Mar. 4, 2011) United States OffGamers http://www.offgamers.com/ (accessed Mar. 4, 2011) Malaysia Belrion http://www.belrion.com/ (accessed Mar. 4, 2011) Singapore, China, US, India EpicToon http://www.epictoon.com/ (accessed Mar. 4, 2011) ELYGAME http://www.elygame.com/ (accessed Mar. 4, 2011) China WeeGold http://www.weegold.com/ (accessed Mar. 4, 2011) China G4S http://www.game4shop.com/(accessed Mar. 4, 2011) United States gameSavor http://www.gamesavor.com/ (accessed Mar. 4, 2011) JPITEMS http://www.jpitems.com/ (accessed Mar. 4, 2011) United States GoldCEO https://www.goldceo.com/ (accessed Mar. 4, 2011) United States MyMMOshop http://www.mymmoshop.com/ (accessed Mar. 4, 2011) United States, Ireland, Australia Mmosale http://www.mmosale.com/ (accessed Mar. 4, 2011) Hong Kong OGPAL http://www.ogpal.com/ (accessed Mar. 4, 2011) United States D2pair http://www.d2pair.com/ (accessed Mar. 4, 2011) China GoLeveling http://www.goleveling.com/ (accessed Mar. 4, 2011) China Helper 4 Gamers http://www.helper4gamers.com/ (accessed Mar. 4, 2011) United States Source: Author’s analysis About infoDev infoDev is global development �nancing program among international development agencies, coordinated and served by an expert Secretariat housed at the World Bank Group, one of its key donors and founders. It acts as a neutral convener of dialogue, and as a coordinator of joint action among bilateral and multilateral donors—supporting global sharing of information on ICT for development (ICT4D), and helping to reduce duplication of efforts and investments. infoDev also forms partnerships with public and private sector organizations who are innovators in the �eld of ICT4D. infoDev is housed in the Finacial and Private Sector Development (FPD) Vice Presidency of the World Bank Group. For additional information about this study or more general information on infoDev, please visit www.infodev.org/publications. Information for Development Program www.infoDev.org THE WORLD BANK www.infoDev.org