From Game to Play Curiosity, Innovation & Entrepreneurship in Online Communities

Research Collection

Doctoral Thesis

From game to play curiosity, innovation & entrepreneurship in online communities

Author(s): Jäger, Peter Michael

Publication Date: 2010

Permanent Link: https://doi.org/10.3929/ethz-a-006473550

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library

Diss. ETH N° 19 371

FROM GAME TO PLAY

CURIOSITY, INNOVATION & ENTREPRENEURSHIP IN ONLINE COMMUNITIES

A dissertation submitted to ETH Zurich for the degree of Doctor of Sciences

presented by Peter Michael Jäger Dipl.-Ing. University of Stuttgart

born March 24, 1979 citizen of Germany

Accepted on the recommendation of Prof. Dr. Georg F. von Krogh Prof. Dr. Elgar Fleisch

2010

For my parents, Simon and all those friends that made Zurich my new home

ACKNOWLEDGEMENTS

Completing a work like this does not only involve time and passion but to a significant extent the support and companionship of friends and family. Working with colleagues who are at the same time friends makes the experience even more pleasurable. My first thanks go to Prof. Georg von Krogh who made all this possible. While constantly encouraging and motivating ambitious research, he also exerted the gentle pressure that is needed to eventually finalize a dissertation and achieve the goal set. Throughout my time at his chair I always got advice from a friend and never the order of a superior that I highly appreciate. I sure hope this friendship will endure the many years to come. I also would like to express my thanks to Prof. Elgar Fleisch who immediately accepted my request to be my co-referee. His support helped very much to successfully finish and compile the dissertation towards the end. Stefan Häfliger played a prominent role during my time at ETH since he supported and coached my work from the very first day in his always friendly, sympathetic and unobtrusive nature. Stefan and I coauthored three of the four included papers. We travelled to many conferences and discussed our research in most extravagant places all over the world making this collaboration one of the most excit- ing I ever had. His way of balancing work and fun showed me that outstanding research is more than a profession – it’s a lifetime commitment. Whoever had the joy of conducting research knows that success is not granted at all times. When such times come, it is very important to have several other good reasons to return to the office. I always had those reasons – my friends at the Chair of Strategic Management and Innovation. The team culture we had was something we were always admired for. Here I want to say thank you to all my team mates: Sebastian, Martin, Zeynep, Matthias, Renato, Andreas, Jan, Alban, Lise, Fotini, Helena, Pasqual, Lu- kas and especially our secretary Hilde who helped to keep me out of trouble more than once. My special thanks go to my very dear friend Simon who I was fortunate to meet during my first lecture I gave at ETH. His unconstrained mind and his casual nature always helped me to cheer up for another round of scientific work while at the same time he encouraged me with his investigative but de- termined way of discovering the big, wide world in everyday life. Among all the people I met during my time in Zurich, he changed my self-perception, the way I approach things, and how I lead my life the most. Should I ever be asked to picture my ideal brother – it would be Simon. As long as the sun sets orange, this friendship may last! Mentioning all my friends I found in Zurich and the stories that are behind them would probably take more pages than those at hand. However, there are always some that stick out. This is Christian, who convinced me to learn Swiss German, an idea he most likely regretted the following couple of weeks. He was the first to show me that Switzerland was home now. Dominik and Damian stand representative for all students who became my friends. Thanks to my presidents for all the fun we had! My time at ETH is over but I definitively hope that the friendships will continue. Last but not least I am deeply grateful to my parents who always believed in me and never doubted that my work will eventually bear fruit. They always gave me the confidence that I can succeed if I only wanted. Thanks!

ZUSAMMENFASSUNG

“From Game to Play” beschreibt wie innovative Anwender von Videospielen als Spie- ler starten und nachdem ihre Neugier beim Beobachten anderer Spieler geweckt wur- de diese Spiele als preiswerte Alternative zu 3D-Animationsprogrammen verwenden um damit selbstproduzierte Filme zu vermarkten. Das hierzu benötigte Wissen sam- meln die Spieler in Online Communities, in welchen sie auch ihre ersten Produktionen veröffentlichen. Während diesem Prozess des Spielens und Lernens entdecken die Mitglieder der Online Communities Möglichkeiten ihre produzierten Filme in der Filmindustrie zu kommerzialisieren, was sie zu „User Entrepreneuren“ werden lässst. User Entrepreneurship als relativ neues Forschungsgebiet hat seinen Ursprung in der Literatur über User Innovation, einem Feld das sich primär der Erforschung von nut- zergetriebenen Innovationen widmet. Nutzer von im Markt erhältlichen Produkten haben oftmals Änderungswünsche oder weitere Entwicklungsideen, welche von den Herstellern nicht oder erst sehr spät aufgegriffen werden. Innovationen von solchen Nutzern dienen daher in erster Linie der Befriedigung eigener Wünsche statt der Ge- winnerzielung, wie dies bei Herstellern der Fall ist. User Entrepreneure unterscheiden sich analog hierzu von klassischen Entrepreneuren indem sie Produkte oder Prototy- pen entwickeln, welche vor allem ihnen selbst dienen. Während der Benutzung dieser Prototypen erkennen sie aufgrund von Rückmeldungen anderer Nutzer ähnlicher Produkte das gegebenenfalls vorhandene Marktpotential und werden anschliessend zu kommerziell agierenden Produzenten. Das in dieser Dissertation präsentierte Modell für User Entrepreneurship über Indu- striegrenzen hinweg betrachtet Nutzer von Videospielen, welche durch die Verwen- dung dieser Spiele zur Filmproduktion, Machinima als neues Genre in der Animati- onsindustrie geschaffen haben. Spezielle Betrachtung erfahren die für den Erfolg der User Entrepreneure wichtigen Online Communities, in welchen die Anwender lernen sich auszutauschen und ihre Produktionen zu präsentieren. Hierbei spielen Kompo- nenten Sozialer Netzwerke zur Massenkommunikation, der Einfluss von Rückmeldun- gen und Ideen zu den Innovationen und persönliche Neugier der Mitglieder eine zen- trale Rolle. Verschiedene Erkenntnisse in Bezug auf den User Entrepreneurship Prozess und des- sen Relevanz für Forschung und Wirtschaft werden aufgezeigt. Hinzu kommen zahl- reiche Aspekte, welche sich mit dem Thema Community Management befassen. Die abschliessenden Betrachtungen umfassen Vorschläge zu möglichen weiteren For- schungsthemen sowie die Grenzen dieser Dissertation.

ABSTRACT

“From Game to Play” illustrates the path innovative users took when they started off as avid gamers who got curious about others’ use of video games as cost effective tools for 3D animation production. They acquired the knowledge and skills necessary to realize first little projects, later staged their own plays in games which they presented to fellow online community members, and after having recognized an opportunity eventually commercialized their Machinima projects as animated shorts in the film industry – thus turning user entrepreneurs. User entrepreneurship as a rather new field of research has its origins in the user innovation literature where users innovate to satisfy their own needs, hence for their own use-value instead of commercial interest. In contrast to the classic approach of entrepreneurship where opportunity recognition precedes prototype development, user entrepreneurs first develop prototypes and while using these for their own purpose receive feedback and thus recognize an opportunity for commercialization. This dissertation presents a model of user entrepreneurship across industry boundaries by studying users in the field of Machinima, a new film genre characterized by shooting film in video games, who enter the animation industry. Being important for the success of user entrepreneurs, the impact of online user communities on the commercialization process is discussed. Therefore, several aspects of online communities such as social network site features for mass communication, the impact of feedback on innovation activities of the sponsoring firm, and individual curiosity as predictor of members’ contribution behavior are studied and described in more detail. Several implications for research and management regarding the user entrepreneurship process with its impact on incumbent firms as well as online community management conclude the thesis complemented by a future research agenda.

TABLE OF CONTENTS

1. From Individual Curiosity to Initial Commercial Success 2

2. Innovating with Peers – How User Entrepreneurs Interact with and Learn from Communities 5 2.1. The Enabling Framework of User Entrepreneurship 5 2.2. Online Communities 7 2.3. Curiosity 8 2.4. User Innovation 8 2.5. User Entrepreneurship 9

3. Setting the Stage – Research Framework and Methodology 11 3.1. The Gaps – Extending Current Research 11 3.2. Change of Scene – Following User Entrepreneurs Beyond Industry Boundaries 12 3.3. Commercial, yet Social – SNS Features in Firm-sponsored Online Communities 12 3.4. Listening to the Prompter – Evaluating the Potential of Community Feedback for Producer Innovation 12 3.5. Turning Extras into Actors – Lurking Behavior in the Light of Individual Curiosity 13

4. Studies of the Core Components – Curiosity, Community and Commercialization 14 4.1. Under the radar: Industry entry by user entrepreneurs 14 4.2. Modding as Rating Behavior in Virtual Communities: The Case of Rooster Teeth Productions 15 4.3. A Directing Audience: How Specialized Feedback in a Virtual Community of Consumption Stimulates New Media Production 15 4.4. Curious Lurkers. Linking I/D-type Curiosity to Contribution Behavior in Online Communities 16

5. Conclusions, Implications and the Way Forward 18 5.1. Implications for Research and Management 18 5.2. Limitations and Future Research 19

References 21

Appendix 1 – Under the Radar: Industry entry by user entrepreneurs 27

Appendix 2 – Modding as Rating Behavior in Virtual Communities: The Case of Rooster Teeth Productions 55

Appendix 3 – A Directing Audience: How Specialized Feedback in a Virtual Community of Consumption Stimulates New Media Production 67

Appendix 4 – Curious Lurkers. Linking I/D-tpye Curiostiy to Contribution Behavior in Online Communities 93

Appendix 5 – Curriculum Vitae 115

FROM GAME TO PLAY

CURIOSITY, INNOVATION & ENTREPRENEURSHIP IN ONLINE COMMUNITIES

Peter Michael Jäger

Chair of Strategic Management and Innovation Department of Management, Technology, and Economics (MTEC) ETH Zürich, Switzerland

November 2010

“From Game to Play” illustrates the path innovative users took when they started off as avid gamers who got curious about others’ use of video games as cost effective tools for 3D animation production. They acquired the knowledge and skills necessary to realize first little projects, later staged their own plays in games which they presented to fellow online community members, and after having recognized an opportunity eventually commercialized their Machinima projects as animated shorts in the film industry – thus turning user entrepreneurs. User entrepreneurship as rather new field of research has its origins in the user innovation literature where users innovate to satisfy their own needs, hence for their own use-value instead of commercial interest. In contrast to the classic approach of entrepreneurship where opportunity recognition precedes prototype development, user entrepreneurs first develop prototypes and while using these for their own purpose receive feedback and thus recognize an opportunity for commercialization. This dissertation presents a model of user entrepreneurship across industry boundaries by studying users in the field of Machinima, a new film genre characterized by shooting film in video games, who enter the animation industry. Being important for the success of user entrepreneurs, the impact of online user communities on the commercialization process is discussed. Therefore, several aspects of online communities such as social network site features for mass communication, the impact of feedback on innovation activities of the sponsoring firm, and individual curiosity as predictor of members’ contribution behavior are studied and described in more detail. Several implications for research and management regarding the user entrepreneurship process with its impact on incumbent firms as well as online community management conclude the thesis complemented by a future research agenda.

FROM GAME TO PLAY PETER JÄGER • 2010

1. FROM INDIVIDUAL CURIOSITY TO INITIAL COMMERCIAL SUCCESS

“… well, I think a lot of it has to do with the fact that the community site that we have made … or at least at the time we made it … had features that weren’t that present in other places, we were a little ahead of the curve at that time, and so there were a lot of cool features that people were interested in. This is like before MySpace really had taken off … So we’ve always tried to give it a little functionality, things they do in a community website they’re interested in making … you know, interested in being a part of it. We tried to make the website almost like a game.”

Interview with Geoff Ramsey of Rooster Teeth Productions on April 3, 2007

Ventures in the Web 2.0 digital age can draw itself. The Journal of Business and Psychology upon resources (Barney, 1991; Wernerfelt, 1984) recently ran a special issue on “Millennials and and capabilities (Grant, 1996; Kogut and Zander, the World of Work” discussing perspectives for 1992) yet inconceivable (Hustad and Teigland, organizations and management brought about by 2008; Pascu et al., 2007). First Google started its a new generation unprecedented in their use of success in 1998 followed by MySpace, Doodle media and digital technologies (Hershatter and and SecondLife which developed in 2003 slightly Epstein, 2010) who have just started to challenge before Mark Zuckerberg launched his Facebook the legacies of the so acquainted corporate world. in the year 2004. Finally YouTube came online in Today, online communities serve as virtual play- 2005. Whether it be information seeking, enter- grounds for members with various interests tainment value or social bonding, any of these (Jeppesen and Frederiksen, 2006; Moore and applications is fueled by a never depleting re- Serva, 2007; Rheingold, 1993; Tafjel, 1978) and source – the curiosity of its users. thus constitute an incubator of new talent, foster Curiosity, the innate basic emotion of all human user-innovation (Franke and Shah, 2003; von beings, drives our behavior to explore, investigate Hippel and von Krogh, 2003; Shah, 2006; von and learn, thus acquire new knowledge (Berlyne, Hippel, 2007), and stimulate users to engage in 1950; Hall and Smith, 1903). Feelings of “insa- entrepreneurial activities (Alvarez and Barney, tiable curiosity” inspire scientific research and 2005; 2007; Sarasvathy, 2001; 2004). The com- eventually yield new innovations that satisfy the munities at the same time serve as testing ground needs and desires of society (Anderson et al., for user entrepreneurs to experiment with and 1996; Nowotny, 2009; Sakkab, 2007). Advances initially explore new ideas, products or services in information technology (IT) eventually enabled (Baldwin et al., 2006; Shah and Tripsas, 2007; the Internet as the new “final frontier” and in Haefliger et al., 2010) since fellow community doing so pushed the limits of information-seeking members provide instant feedback on product possibilities to entirely new extents. Curiosity, features or potential improvement opportunities among other motives, drives individuals to gather (Jäger et al., 2010). Considering these incubating online and engage in social bonding (Ahuja and characteristics of the Web 2.0, starting a new Galvin, 2003; Kozinets, 2002; Litman and Pezzo, business has never been easier. 2007; Wellman and Gulia, 1997; Wellman, 2001), However, entrepreneurship carried out by users share and explore entertainment content (Cha et as studied by Haefliger and colleagues (2010) in al., 2007; Kim et al, 2009; Roth et al., 2009), or the context of Machinima1 appears to be a rather engage in innovation activities (Fueller, 2006; fragile concept, which depends on several ena- Fueller et al, 2007; Mollick, 2005). In connecting bling factors that have not all been explored and people and making knowledge instantly available, understood to their entirety yet. Since user entre- everywhere and anytime, the success of Apple’s preneurship, in contrast to the conventional en- iPhone can be attributed to its variety of online features, changing the way in which individuals 1 Machinima, the production of 3D animation, is a new interact – with others and with the technology film genre characterized by shooting film in video games.

2 FROM INDIVIDUAL CURIOSITY TO INITIAL COMMERCIAL SUCCESS

trepreneurship approach, depends on the exposition of new product ideas to potential customers prior to commercialization (Shah and Tripsas, 2007), communities are of integral importance (Haefliger et al., 2010; Jäger et al., 2010). The opening statement to this section demonstrates how Rooster Teeth Productions – one of the focal Machinima companies of the research in this dissertation – dealt with community topics and community-firm interaction early on to establish a flourishing community of consumers who are at the same time co-creators (Jäger et al., 2010). Notably, Geoff Ramsey mentioned the impor- Figure 1. Machinima, the user-driven phenomenon tance of keeping members interested and making them “a part of it”. Keeping online communities alive strongly de- The doctoral thesis at hand is composed of this pends on the enduring participation of members opening chapter accompanied by four independ- (Ren et al., under review) who ideally never expe- ent research papers authored or co-authored by rience feelings of boredom or solitude (Jäger, the writer. All papers cover different aspects of 2010). In the end, whether online ventures are user entrepreneurship and online communities successful or not comes down to the supportive thereby drawing upon evidence from the Ma- attitude of its users who are very often members chinima phenomenon as illustrated in figure 1 of dedicated online communities as has been (Hancock and Ingram, 2007; Lowood, 2007; described by marketing scholars (Algesheimer et Marino, 2004; Morris et al., 2005). This intro- al., 2005; Cova and Pace, 2006; Kozinets, 2007; ductory chapter covers the literature review on McAlexander et al., 2002; Schau and Gilly, the respective topics, the research framework of 2003). However, with up to 90%, the number of the dissertation, abstracts of the four pivotal pa- passive members in online communities, so-called pers, and concludes with generalized findings to lurkers (Kozinets, 1999; Nonnecke and Preece, arrive at new perspectives for research and man- 2000; Rafaeli et al., 2004), is alarmingly high and agement. All studies are positioned in the man- research missed out on addressing the causes agement and psychology literature at the intersec- sufficiently (Katz, 1998; Mason, 1999; Nonnecke tion of nascent research on epistemic curiosity2, and Preece, 2000; 2001; 2003; Nonnecke et al., community-firm interaction, and user entrepre- 2004; Preece et al., 2004). Lurkers do not consti- neurship. The first paper “Under the radar: In- tute a threat but were highly valuable if re- dustry entry by user entrepreneurs” (Haefliger et activated due to their familiarity with the online al., 2010) draws upon the emerging research on community compared to complete outsiders. user entrepreneurship and presents a new model Establishing new ventures online – as easy as it how user entrepreneurs combine existing with appears in the Web 2.0 context – hinges on sev- newly acquired knowledge, interact with their eral critical factors. Understanding the possible peer and customer communities, and finally en- success factors of online ventures in conjunction gage in commercialization of Machinima in the with their mutual embeddedness in online animation industry. The second paper “Modding communities is therefore crucial. This thesis takes as rating behavior in virtual communities: The an early concept of user entrepreneurship (Shah case of Rooster Teeth Productions” (Haefliger et and Tripsas, 2007) and links it to three major al., 2009), examines a firm-sponsored virtual online community topics, namely curiosity to community that makes use of social network site explain lurking and information-seeking behavior, features which enable members to “mod”, hence member specialization and their innovation sup- to evaluate other members’ contributions. The port, and member-to-member communication to third paper “A Directing Audience: How Special- gain an advanced understanding of how entre- ized Feedback in a Virtual Community of Con- preneurial activities by users hatch in the online community context. 2 Special interest is paid to interest- and deprivation-type epistemic curiosity as personality trait, which constitutes a very recent research agenda (see Litman, 2005; 2008; 2010).

3 FROM GAME TO PLAY PETER JÄGER • 2010

sumption Stimulates New Media Production” Lurkers. Linking I/D-type Curiosity to Contribu- (Jäger et al., 2010) elaborates on online commu- tion Behavior in Online Communities” (Jäger, nities and their potential for value co-creation. 2010) advances recent research on interest- and Members’ combination of technical expertise informational deprivation-type epistemic curios- and knowledge about the product leads them to ity as inherent personality traits of individuals by generate specialized feedback for process and theorizing about community members’ wanting product innovation for the sponsoring firm. The and liking of new information to explain ongoing empirical study also yielded, that only 10% of all commitment and lurking behavior in online members contributed more than once, while the communities. rest became lurkers. The fourth paper “Curious

2. INNOVATING WITH PEERS – HOW USER ENTREPRENEURS INTERACT WITH AND LEARN FROM COMMUNITIES

“We’re talking about the mid 1990s, with Quake, you have a game that is based on client server architecture, so you could play now over the Internet. It’s the same technology that’s driving the game that’s making the communities possible. I think that’s kind of important to keep in mind. People are discovering about virtual communities at the same time as they’re learning about these competitive multiplayer games, and at the same time as they are learning how to make movies from them. It’s all part of the same sort of engagement with this still relatively new technology in the mid 1990s, which was used for a variety of purposes. So I think the relationship of the player to usually another player, but someone who is a spectator, was very important for the early development of the Machinima community.”

Interview with Henry Lowood of Stanford University on March 15, 2007

When first introducing the idea of user innova- edge for product development, and the role as tion in 1988, von Hippel deviated from the com- testing ground and initial customer base. mon understanding of innovation, which links an Next, some basic parameters of user innovation invention with subsequent market success and entrepreneurship are discussed followed by a (Schumpeter, 1934) in that he emphasized the description of various community types with use-value of a new product to the user innovator. their respective impact on value (co-) creation. Research soon embraced the concept of users as Then the literature on curiosity, user innovation, a legitimate source of innovation yielding numer- and user entrepreneurship is discussed in respect ous studies on user innovation in various industry of online communities. domains (see Bogers et al., 2010 for a thorough review). However, considering the potential commercial success of a user innovation (Baldwin 2.1. The Enabling Framework of User Entre- et al., 2006) and thus recognizing the user as en- preneurship trepreneur was not established till Shah and Trip- sas (2007) studied entrepreneurial activities of Advances in IT, especially the Internet, estab- users in the juvenile products industry.3 lished a magnitude of new opportunities for individuals as well as businesses to engage in transac- While the user innovation literature describes the tions and activities that appeared inconceivable importance of user communities extensively until recently. Social software provides new means (Franke and Shah, 2003; von Hippel and von to bond and stay connected irrespective of the Krogh, 2003; Shah, 2006; von Hippel, 2007), others’ present location, cultural background, or studies on user entrepreneurship just started to social situation (Ahuja and Galvin, 2003; Kozi- explore the impact of communities on the suc- nets, 2002; Litman and Pezzo, 2007; Wellman cessful outcome of users’ entrepreneurial activi- and Gulia, 1997; Wellman, 2001). Its speed and ties (Haefliger et al., 2010; Shah and Tripsas, cost effectiveness as distribution channel made 2007). Relevant aspects of community-firm inter- the Internet an invaluable resource (Barney, action have been described such as the impact on 1991; Wernerfelt, 1984) to firms for knowledge opportunity recognition, acquisition of knowl- sharing (Hustad and Teigland, 2008), (virtual) product distribution, e-commerce, and marketing activities in general (Armstrong and Hagel, 1996; Ghosh, 1998; Ranchhod and Gurau, 1999; 3 Earlier work by Fanke and Shah (2003), Luthje and colleagues (2005), and Baldwin and colleagues (2006) Varadarajan et al., 2008; Webb, 2002). As a pointed towards that direction but did not establish user downside, the Internet exposed firms to increased entrepreneurship as dedicated new concept. competition (Dewan et al., 2000) and in the case

5 FROM GAME TO PLAY PETER JÄGER • 2010 of digital products such as music or media at- tive action carried out by the community of de- tracted and supported piracy (Smith and Telang, velopers who at the same time benefit from the 2009). use of the software. This model is thus based on User-generated content such as wikis, blogs, fan mutual interest and benefit in the absence of sites, or media content is the new paradigm of exclusive appropriation. Web 2.0 (Hustad and Teigland, 2008). Taking User Entrepreneurship (Baldwin et al., 2006; this one step further, users carry out innovation Haefliger et al., 2010; Shah and Tripsas, 2007) activities online (Franke and Shah, 2003; Franke follows a slightly different path in that user entre- and von Hippel, 2003; Lee and Cole, 2003; Piller preneurs either innovate new products (Shah and and Walcher, 2006; Sawhney et al., 2005). They Tripsas, 2007) or draw upon existing user innova- thereby share their ideas with fellow community tion (Haefliger et al., 2010), which they commer- members or engage in value co-creation with cialize and hence exclusively appropriate. The firms (Fueller, 2006; Fueller et al., 2007; Jensen et community of users hereby inspires the user en- al., 2007; Jeppesen and Frederiksen, 2006). trepreneur, delivers relevant knowledge to inno- Firms increasingly draw upon ideas from their vate, assists in testing the products, provides feed- users to advance existing products or embrace back for improvements during use, and in doing new product ideas inspired by users. Boundaries so builds the basis for the user entrepreneur to between the firm and its environment are thus recognize an opportunity to commercialize the permeable allowing an in- and outflow of ideas. newly invented product (Haefliger et al., 2010; The resulting innovation practices based on an Shah and Tripsas, 2007). Innovation practice in interaction between firms and users or user this case constitutes an interaction between the communities lay the foundation for hybrid orga- nascent firm and the user community where nizations that apply open innovation principles community members benefit from reputational (cf. Chesbrough, 2003). One example constitutes gains grounded in supporting each other and the the collaborative development of open source prospect to obtain the new product that satisfies software where individuals who organize in their needs in a superior way compared to exist- communities contribute equally alongside firms ing, commercially available products. The sup- (Spaeth et al., 2008; Stuermer et al., 2009). von porting mechanisms, different innovation prac- Hippel and von Krogh (2003) termed this ap- tices, interaction patterns, and resulting organiza- proach the private-collective model of innovation tional forms are depicted as enabling framework given the private investment and later appropria- of user entrepreneurship in Figure 2. tion for financial returns by firms and the collec-

Figure 2. The enabling framework of user entrepreneurship

6 INNOVATING WITH PEERS – HOW USER ENTREPRENEURS INTERACT WITH AND LEARN FROM COMMUNITIES

Online communities thus play a vital role in at- novation scholars on the other hand are mainly tracting new users to engage in innovation activi- concerned with communities that produce soft- ties – which we later refer to curiosity, carry out ware or generate ideas for new products or serv- the innovation task, link consumers with produc- ices (Franke and von Hippel, 2003; Lee and Cole, ers, and assist in the commercialization process of 2003; Jeppesen and Frederiksen, 2006). With new products. This is illustrated by the quote by their members supporting a product or service, Henry Loowood, Curator for History of Science promoting a brand and spreading loyalty to a & Technology Collections and Film & Media product or firm, or acting as a resource for ideas Collections in the Stanford University Libraries, (Carlson et al., 2007; Kozinets, 1999; McAlexan- who describes how users recognized the potential der et al., 2002; McWilliam, 2000; Muniz and of Machinima as new animation production O'Guinn, 2001; Nambisan, 2002; Rowley et al., technology while simultaneously learning about 2007; von Hippel, 1988), online communities online communities. The relationship between positively impact on firm performance (Alge- producer and consumer was thus inevitable in the sheimer and Dholakia, 2006; Armstrong and development process of Machinima. Hagel, 1996, Porter, 2006). However, characteristics of online communities vary along several dimensions (Jäger et al., 2010). 2.2. Online Communities Considering value co-creation, firm involvement Being an important driver from the very begin- takes different levels from very active participa- ning, online social interaction is one of the major tion where firms delegate employees to develop activities on the Internet besides entertainment software (Spaeth et al., 2010; Stuermer et al., and information search (Rheingold, 1993; Tafjel, 2009), solve problems online, contribute to dis- 1978; Usmani et al., 2010). With the inception of cussions (Rowley et al., 2007; Luedicke, 2006), Web 2.0, online communities were taken to an and organize brand feasts for community mem- entirely new extent (Hustad and Teigland, 2008; bers (Bagozzi and Dholakia, 2006) to passive Massari, 2010; O’Reilly, 2007). Today, adoles- involvement where firms just offer the online cents learn how to leverage online content from space for interaction to their consumers (Cova social network sites to complement their relation- and Pace, 2006) to no involvement which hap- ships in real life (Subrahmanyam and Greenfield, pens for instance in brand or fan communities 2008). Although social network sites make only a organized by consumers themselves (Fueller et al., fraction of the various types of online communi- 2007; Janzik et al., 2006; Kozines, 1997). The ties, social network site features like user profiles scope towards innovation largely depends on the or commenting and rating systems render invalu- community (Jäger et al., 2010; Sawhney and able for online community sites in general (Hae- Prandelli, 2000). Communities of consumption fliger et al., 2009). deliver information in their discussion of prod- Online communities have been studied in several ucts or reports on usage (Algesheimer et al., 2005; contexts and across different disciplines including Bagozzi and Dholakia, 2006; Kozinets, 2007; information systems research, marketing, innova- McAlexander et al., 2002). Communities of creation and management, and psychology (Porter, tion, which operate at the intersection of con- 2006). While the information systems literature sumption and production (Sawhney and Pran- approaches the question “How” users interact, delli, 2000) generate knowledge by linking per- i.e. which technology they use, marketing scholars sonal experience with product tests or answering are most interested in “Who” is acting and for marketing surveys. While, communities of pro- “Whom”. Innovation and management research, duction innovate new virtual products or services showing an overlap with marketing, takes interest or – for physical products – contribute detailed in “What” is created while studies in the field of building instructions online (Franke and Shah, psychology mainly deal with the “Why” question 2003; Hienerth and Lettl, 2009; von Hippel, referring to the users’ motives. Marketing scholars 2005). The latter are thus most relevant for ob- thus tend to focus on online communities of con- serving user innovation activities. sumption (Algesheimer et al., 2005; Cova and Motivation, as mentioned earlier, is located at the Pace, 2006; Kozinets, 2007; McAlexander et al., intersection of online community research and 2002; Schau and Gilly, 2003), which unite mem- psychology. The most common framework used bers in their consumption behavior of products, to explain individual motivation for community services, or as fans in the form of fandom (Kozi- participation and contribution to value creation is nets 2001; Pichler and Hemetsberger, 2007). In- the self-determination theory (SDT) by Deci and

7 FROM GAME TO PLAY PETER JÄGER • 2010

Ryan (1985). According to the SDT, two basic berger et al., 1976). The model thus relates a sets of motives drive action, which Deci and relatively high wanting (appetite, need) of new Ryan refer to as intrinsic and extrinsic motiva- information combined with a relatively high lik- tion. While intrinsically motivated action is per- ing (pleasure) of this information with D-type formed for the mere joy and interest inherent to curiosity while a relatively low wanting paired the activity, extrinsically motivated action is with relatively high liking of new information is rather goal oriented and stimulated by an exter- associated with I-type curiosity. An illustrative nal entity. To name a few, fun and enjoyment, example for this is food, which can be enjoyed community identification, or altruism are exam- (liking, pleasure) stimulated by an appealing smell ples of intrinsic motivation while career or pay without previously experiencing hunger (wanting, exemplify extrinsic motives (for an extensive re- appetite) (Berridge, 2003; 2004). This dissertation view of the motivation literature see von Krogh applies curiosity to explain engagement in online et al., 2010). While studying Open Source Soft- communities as well as lurking behavior (Kozi- ware development, Roberts and colleagues (2006) nets, 1999; Nonnecke and Preece, 2000; Rafaeli advanced the classic SDT perspective in describ- et al., 2004), the passive state of online commu- ing reputation, reciprocity, learning, and own-use nity members (cf. Jäger, 2010). value as internalized extrinsic motivation (cp. Lakhani and von Hippel, 2003; Lerner and Ti- role, 2002). Applying the SDT framework, how- 2.4. User Innovation ever challenged by von Krogh and colleagues Manufacturers were for a long time taken to be (2010), to the context of online community par- the predominant source of innovation (Schum- ticipation and contribution to value co-creation, peter, 1934). In 1976, von Hippel changed our yields that intrinsic and internalized extrinsic understanding and added users as an important motives dominate in this domain (Janzik et al., source of innovation when observing user innova- 2009). Users are thus more motivated by joy and tion in the field of scientific instruments. Refer- interest than by external stimulation such as ring to the appropriation of the innovation, von monetary rewards (Lerner and Tirole, 2002). On Hippel termed innovation where innovators the contrary, imposing external rewards on an benefit from monetary rewards as manufacturer intrinsically motivated community might lead to innovation while users benefit from the use-value a “crowding out” effect that results in decreased of an innovation without commercial intention activity levels (Alexy and Leitner, 2008; Frey and (von Hippel, 1988). Approaching the question, Oberholzer-Gee, 1997; Frey and Jegen, 2001) why users innovate, Bogers and colleagues (2010) identified two basic reasons: the users’ low cost of innovation and the innovators’ benefit from use. 2.3. Curiosity Innovation costs are mainly associated with Curiosity, the inherent desire of humans to strive knowledge acquisition relevant for the innovation for new knowledge (Berlyne, 1966; Hall and process. Again, users show two distinct character- Smith, 1903; Loewenstein, 1994) constitutes itself istics that provide them with an advantage over as feeling of informational deprivation (Lowen- manufacturers rendering users more likely to stein, 1994) or as feeling of interest (Spielberger innovate in certain fields since users better know and Starr, 1994). Interest-type (I-type) curiosity about their needs than manufacturers do. Since drives individuals to explore new knowledge for transferring knowledge is costly (cf. Cohen and the pure joy of it while informational depriva- Levinthal, 1990; Nonaka, 1991), users can draw tion-type (D-type) curiosity accrues from states of upon their “sticky” knowledge (von Hippel, uncertainty that motivate individuals to seek new 1994), which provides them with a unique advan- knowledge in order to remedy the unpleasant, tage in the innovation process. Thus, the locus of negative feelings associated with uncertainty innovation is more likely to remain with the user (Litman, 2008; Litman & Jimerson, 2004; Litman than with the manufacturer (Lüthje et al., 2005; & Spielberger, 2003; Loewenstein, 1994). Ogawa, 1998). In addition, users usually have In linking I- and D-type curiosity to the need for lower opportunity costs than manufactures. This (appetite) and the liking of (pleasure) new infor- combined with the fact that user often posses mation, Litman (2010) developed the wanting- expert knowledge from either their professional liking model of information-seeking based on domain or a hobbyist activity endows the user prior work by Litman (2005) and the state-trait innovator with the necessary resources to create theory by Spielberger (Spielberger, 1975; Spiel- low-cost innovative solutions (Franke and Shah,

8 INNOVATING WITH PEERS – HOW USER ENTREPRENEURS INTERACT WITH AND LEARN FROM COMMUNITIES

2003; Lakhani and von Hippel, 2003; Lüthje, recognition over prototype development and 2004). finally to commercialization (Venkataraman, Regarding the knowledge, hence the cost aspect 1997; Shane and Venkataraman, 2000). When of innovation, explains why users under certain first describing user entrepreneurship, Shah and circumstances are more likely to innovate than Tripsas (2007) inverted one major aspect of the manufacturers but provides no motives or inten- process: they argued that users first develop a tions why users engage in actual innovation prac- prototype and while using this prototype for own tice. For this, the users expected benefits are of purposes recognize potential opportunities for relevance (von Hippel, 2005). The discussion here commercialization (Haefliger et al., 2007; Shah somewhat compares to the previous discussion on and Tripsas, 2007). User entrepreneurs show curiosity where information seeking behavior is general traits of user innovators (von Hippel, motivated by individual needs. Users, according 1988; 2005) in that they base their newly devel- to von Hippel’s definition benefit directly from oped prototypes on commercially available prod- using the innovation – the innovation thus satis- ucts. This brings about new challenges for user fies a need, which could not be addressed by ex- entrepreneurs regarding intellectual property (IP) isting products or services (von Hippel, 1976; rights that adhere to the original product and 1988; 2005) leading the users to innovate for hence remain with the producer (Haefliger et al., themselves. New solutions brought about by users 2010). User innovators who innovate for their thus have the potential to go beyond functionali- own benefit usually do not face these IP issues or ties and are fundamentally different to what at least IP holders are more lenient and generally manufactures might create. Taking this one step desist from enforcing IP rights due to the effort further, research introduced the concept of lead- and excessive costs of enforcement (Liebeskind, users (von Hippel, 1986; 1988; 2005). Lead-users 1996). Possible strategies how to deal with alien are ahead of trends and thus experience con- IP incorporated in own products are discussed in straints more frequently in using commodity one of the research papers of this dissertation products than the majority of users. These con- (Haefliger et al., 2010). straints, which might develop into unique, yet Generally low opportunity costs of user entre- unsatisfied needs may stimulate the lead-users’ preneurs are another common trait they share curiosity to seek alternative solutions or adapta- with user innovators. User entrepreneurs thus tions to already excising products. Such individu- freely experiment with new ideas since they are als are prone to become user innovators. A sec- not restrained by tight R&D budgets. This ex- ond yet understudied benefit for users who inno- perimentation in conjunction with a usually high vate is their enjoyment, which they derive from exposition to online communities of peers yields the process of problem solving (Lüthje, 2004; von valuable feedback for performance improvement Hippel, 2005). This again relates to curiosity in from other users with whom the innovation is that users experience pleasure in finding out shared and thus simultaneously enhances chances about new products or in seeking and recombin- for opportunity recognition (Haefliger et al., ing existing knowledge for innovation. 2010; Shah and Tripsas, 2008). In the entrepre- Motives, which drive users to become user inno- neurship literature such approaches where entre- vators are hence expected to be closely linked to preneurs start with a set of given means and curiosity where personality traits and momentary eventually find a reasonable opportunity for states of curiosity explain individuals’ behavior commercialization are termed effectuation (Read (cf. Jäger, 2010). This aspect has been addressed et al., 2009; Sarasvathy, 2001). However, as soon in one of the dissertation papers. as prototypes are commercialized, the innovation community turns into a community of consumers Beyond benefiting from use, recent research ob- that might impose new challenges for user entre- served the user innovators’ benefits from selling preneurs in motivating these consumers to pro- thus commercializing their ideas, which advances vide feedback and continue participation and co- the user from an innovator to an entrepreneur creation activities (Haefliger et al., 2010; Jäger et (Haefliger et al., 2010; Shah and Tripsas, 2007). al., 2010). When Shah and Tripsas (2007) first studied user 2.5. User Entrepreneurship entrepreneurship, they observed users who commercialized their products in the same industry in Entrepreneurship in its classic conception de- which they found the original products. However, scribes a series of activities from opportunity user innovators might draw upon products, i.e.

9 FROM GAME TO PLAY PETER JÄGER • 2010 assets from one industry and use the altered assets industry boundaries. User entrepreneurs show a as such or as tools for the development of new distinct talent in leveraging complementary assets products in another industry (Faulkner and (Teece, 1986) when they base their products on Runde, 2009) – they thus diffuse an existing tech- readily available assets which they alter or other- nology across industry boundaries (cf. Rogers, wise apply in a new context, use cost effective 1962). Users who changed industry dynamics distribution channels like the Internet, and draw have been studied by Baldwin and colleagues upon their existing community to learn and re- (2006) while Haefliger and colleagues (2010) de- cruit new talent (Haefliger et al., 2010). veloped a model of user entrepreneurship across

10 SETTING THE STAGE – RESEARCH FRAMEWORK AND METHODOLOGY

3. SETTING THE STAGE – RESEARCH FRAMEWORK AND METHODOLOGY

“I think that’s fascinating. To me that’s something you would never have in any other medium. And some users they watch it and they think, oh it’s just some guy messing around, which it is. But there’s, I think there’s a really … there’s like real theory going on behind that.”

Interview with Chris Burke of Bong+Dern Productions on March 2, 2007

The theory discussion presented in the previous von Hippel, 2003; Lee and Cole, 2003; Jeppesen chapter gives rise to several unexplored research and Frederiksen, 2006) are well studied. The questions in the field of online communities and reasons why community members contribute are user entrepreneurship. Some are remotely ad- covered by the motivation literature (Butler et al., dressed in that pivotal papers of this dissertation 2007; Jeppesen and Frederiksen, 2006; Moore have been cited. However, the primary objective and Serva, 2007; Rheingold, 1993; Tafjel, 1978; of this dissertation is to link nascent research on Wasko and Faraj, 2000; for a thorough review see user entrepreneurship and curiosity to existing von Krogh et al., 2010). However, with ratios up research streams of user innovation, knowledge to 90% of community members taking a passive creation, and entrepreneurship in that two key role – so called lurkers (Kozinets, 1999; Nonnecke questions about user motivation and online and Preece, 2000; Rafaeli et al., 2004), notably community interaction are addressed: 1) How do few studies have been conducted on explaining online communities of (innovative) users affect lurking behavior and potential remedies. user entrepreneurship and 2) How does curiosity Attempting to fill the gaps in current literature, as personality trait affect individuals’ online this dissertation presents four independent studies community (and innovation) behavior. that empirically and theoretically explore online communities and individual personality characteristics to present implications how value co- 3.1. The Gaps – Extending Current Research creation in online communities can best be lever- Existing studies on user innovation long neglected aged by user entrepreneurs and established firms. the fact that innovative users would commercial- Considerable validity is derived from the phe- ize their ideas. While few studies point towards nomenological approach and the grounding of the fact that users actually profit from selling their the studies in the area of Machinima with which innovations (Foxall and Tierney, 1984; Lee, 1996) the authors could familiarize themselves for an it only incurred recently that users indeed act as ample amount of time. Findings are based on entrepreneurs (Shah and Tripsas, 2007). How- large quantitative as well as qualitative data from ever, little is known about the role the community archival online sources as well as expert inter- of fellow users plays in this process. Another, yet views with unique sets of data for each project. unstudied aspect in the area of user entrepre- While the first study on user entrepreneurship neurship is the potential shift of industries, thus allowed for a broad overview of several aspects of the diffusion of technology (cf. Rogers, 1962) Machinima, the subsequent studies deal with while creating new market niches (Baldwin et al., remarkable single aspects identified during re- 2006; Faulkner and Runde, 2009). search for the first project. Combining the four Online communities, their role in social bonding studies as presented next provides a holistic view (Ahuja and Galvin, 2003; Kozinets, 2002; Litman on online communities and user entrepreneur- and Pezzo, 2007; Wellman and Gulia, 1997; ship. Wellman, 2001), their positive impact on firm performance (Algesheimer and Dholakia, 2006; Armstrong and Hagel, 1996, Porter, 2006).), and their impact on value co-creation (Franke and

11 FROM GAME TO PLAY PETER JÄGER • 2010

3.2. Change of Scene – Following User Entre- the ease-of-use of the community site, and estab- preneurs Beyond Industry Boundaries lish moderated communication through mutual rating, yielding a higher social embeddedness of The first paper “Under the radar: Industry entry members (Algesheimer et al., 2005). Schoberth by user entrepreneurs” (Haefliger et al., 2010) and colleagues (2006) thus propose future re- explores the Machinima phenomenon in its en- search on communication activities in online tirety. Machinima, the production of animated communities. shorts by using video games as tools, challenges current research on user entrepreneurship (Shah Research for this paper comprises a large quanti- and Tripsas, 2007) in that user entrepreneurs shift tative study on the firm-sponsored online com- the innovation activity from the video game to munity of Rooster Teeth Productions, the most the motion picture industry. Besides obtaining effective Machinima production company to innovation relevant complementary assets in the date. Rooster Teeth Productions was founded in video game industry, this shift requires the user 2003 after the creators of the trailer to their most entrepreneur to combine knowledge from differ- popular Machinima Series “Red vs. Blue” re- ent domains, attract a first potential customer ceived overwhelming feedback from the back- base to enable opportunity recognition, and even- then Machinima community. Starting off as user tually deal with IP issues tied to the complemen- entrepreneurs, the Rooster Teeth Productions tary assets, here the video games. crew soon recognized the beneficial activities of a flourishing user4 community (cp. the opening Since several of these aspects have never been statement to Chapter 1 by with Geoff Ramsey). addressed in the context of user entrepreneur- Over the years almost 43,000 community mem- ship, the study employs an inductive logic to ar- bers posted more than 480,000 comments on 165 rive at a process model of commercialization by episodes. Analyzing this data with SPSS and Ex- user entrepreneurship across industry boundaries cel provided an improved understanding of how (Cohen, 1980; Eisenhardt, 1989; Eisenhardt and community members interact when SNS features Graebner, 2007; Glaser and Strauss, 1967; are present in their community architecture. Strauss and Corbin, 1990). Propositions were formulated based on a multiple, non-embedded case study (Yin, 2003) of seven firms within the 3.4. Listening to the Prompter – Evaluating Machinima community. Altogether, 25 inter- the Potential of Community Feedback for views, two field trips, and an extensive study of Producer Innovation archival Internet data yielded important insights into the commercialization process user entrepre- The third paper “A Directing Audience: How neurs undergo and the involvement of the users’ Specialized Feedback in a Virtual Community of community in this process. Detailed results and Consumption Stimulates New Media Produc- findings of this and the other three research pro- tion” (Jäger et al., 2010) links research on com- jects are presented in the next chapter. munities of consumption (Algesheimer et al., 2005; Cova and Pace, 2006; Kozients, 2007; McAlexander et al., 2002; Schau and Gilly, 2003) 3.3. Commercial, yet Social – SNS Features in with the innovation literature on feedback and Firm-sponsored Online Communities value co-creation (Jeppesen and Frediksen, 2006; Sawhney et al., 2005). While communities of The second paper “Modding as rating behavior production such as open source software com- in virtual communities: The case of Rooster munities are a respected source of innovation Teeth Productions” (Haefliger et al., 2009) ex- (von Krogh and von Hippel, 2003; 2006), little is tends work at the intersection of social network known about the impact of online communities sites (Boyd et al., 2007) and online communities of consumption on innovation practices of firms. (Porter, 2004). While both research streams as So far in the domain of marketing research, on- part of the broader field of computer-mediated line communities of consumption have been communication have been studied independently, studied for aspects like loyalty building to existing little work exists on online communities that integrate SNS features for mass communication among community members. However, integrat- 4 It is debatable whether user community, consumer ing SNS features might increase the performance community or customer community is the better term of member-to-member communication, enhance here. For reasons of consistency throughout the work the term user community is preferred.

12 SETTING THE STAGE – RESEARCH FRAMEWORK AND METHODOLOGY

products (Cova & Pace, 2006; Kozinets, 1999, Online Communities” (Jäger, 2010) theoretically 2002; Muniz & O'Guinn, 2001) or brands (Picher applies the “wanting-liking model of informa- & Hemetsberger, 2007). Given the vast amount tion-seeking”, developed by Litman (2010) to of content mainly in the form of comments, gen- show individual differences in knowledge acquisi- erated by these types of communities, the question behavior, to explain lurking behavior in on- tion how firms could evaluate and thus use this line communities (Kozinets, 1999; Nonnecke and information for product improvements or innova- Preece, 2000; Rafaeli et al., 2004). An extensive tion arises. Supporting this claim, Kozinets and literature review on online communities, the colleagues (2008) propose the idea of creative and emerging field of lurking behavior, the motivation innovative users in online communities of con- literature referring to joining and contributing to sumption who produce collective innovation “as online communities, curiosity in general, and the an aggregated byproduct of everyday informa- nascent contributions on curiosity as personality tion consumption”. However, no empirical evi- trait is presented. dence has been presented yet to substantiate the Contributing to upcoming research in manage- value of mass communication, generated by ment and personality psychology, the paper theo- members of communities of consumption, for retically develops a combined model of commu- firm innovation. nity members’ wanting (appetite) and liking The focal community for the grounded theory (pleasure) of new information with community approach (Glaser & Strauss, 1967; Strauss and joining and lurking tendencies of individuals Corbin, 1990) was again the Rooster Teeth Pro- based on interest- and informational deprivation- duction Red vs. Blue user community. Applying type curiosity as inherent personality traits. The netnography (Kozinets, 1998; 1999; 2010) led to research thus interprets Litman’s (2005; 2008; the collection of 12,000 unsolicited comments by 2010) work to explain the results of existing stud- users of which 7,000 were coded individually. An ies on member motivation and lurking behavior in-depth analysis of the content of these com- in the light of curiosity as personality trait. ments allowed assessing their potential contribution to innovation in different categories like While chapter 3 presents the research framework product or process innovation. of this dissertation in describing the methods and linking the four pivotal papers, chapter 4 summa- 3.5. Turning Extras into Actors – Lurking Be- rizes each study and presents the core results havior in the Light of Individual Curiosity before turning to the aggregated and re-evaluated conclusions and implications in chapter 5. The fourth paper “Curious Lurkers. Linking I/D-type Curiosity to Contribution Behavior in

13 FROM GAME TO PLAY PETER JÄGER • 2010

4. STUDIES OF THE CORE COMPONENTS – CURIOSITY, COMMUNITY AND COMMERCIALIZATION

“Well, the first time I saw Machinima as a possible business opportunity was when I was working at id Software. I was the director of business development at id and I saw what people were doing with Quake. They were making little mini-movies within the Quake universe, where people would get together from all over the country and sometimes different parts of the world […] So, when I saw that, I thought this is a quite innovative little thing that they are doing. They are taking something that was meant to just record your game, and they are actually making little entertaining skits out of it. So I thought this might be something to keep an eye an in the future because it takes so much time and so much money to produce actual animation, that if you were able to multitask and take advantage of the videogame assets that are already being made. So when I left id Software I thought this is a really great opportunity and I decided to pursue it. The problem with pursuing it in the year 2000 was that nobody really knew what it was. […] But there’s enough of a fan base in Machinima that a lot of people started doing it.”

Interview with Anna Kang of Fountainhead Entertainment on March 30, 2007

Anna Kang, one of the Machinima user entre- the film industry, which constitutes a shift of in- preneurs’ studied throughout this dissertation, dustries (cf. Baldwin et al., 2006) – phase one. highlights the main aspects of the Machinima Since users had no commercial interest at first phenomenon in her statement on how Fountain- and the phenomenon was of minor importance head Entertainment started off: users of video to the game companies, which thus neglected to games were interested and got curious about the observe the users’ activities in the new industry, new, innovative medium Machinima. They get this shift is considered to happen “under the ra- together, form groups, organize in online fan dar” of incumbent firms. Then active in the ani- communities, and eventually seizing the potential mation industry, users soon attracted a sizeable of the technology recognize the commercial op- fan base that led user entrepreneurs to recognize portunity given by the use of video games and the potential opportunity of the new medium assets as production tool for animated shorts. (Shah and Tripsas, 2007). In commercializing This chapter provides an overview of the key products related to Machinima, user entrepre- results and findings gained during the four studies neurs became visible in the motion picture indus- on the core components of the user entrepre- try – phase two. neurship process: curiosity, community and However, the two-phase approach is less of a commercialization. preconceived path intended by user entrepreneurs than the result of various contextual factors that impact on the commercialization process. 4.1. Under the radar: Industry Entry by User User entrepreneurs heavily rely on access to Entrepreneurs complementary assets, hence they use the video When user entrepreneurs enter industries, they games as tools. Obtaining the rights from video do this in a two-phased approach. In the case of game producers to legally apply games and re- Machinima, user innovators first developed new spective assets for animation production needs to skills how to use video games to record their gam- precede commercialization. In contrast to user ing performance (Faulkner and Runde, 2009; von innovators who frequently make unauthorized Hippel, 2005). Learning about this new technol- use of copyrighted material (Lee, 2008), user ogy, users with an interest in the animation indus- entrepreneurs first operate under conditions of try applied the video games as production tools in uncertainty but later address IP issues with high sensitivity.

14 STUDIES OF THE CORE COMPONENTS – CURIOSITY, COMMUNITY AND COMMERCIALIZATION

Users who intend to turn entrepreneurs need to members posted at least one comment that re- combine knowledge on how to apply the com- ceived a mod. Considering the comment level, plementary assets with new domain knowledge of only 15% of all comments were modded. A pos- the film industry. They therefore draw on existing sible explanation also used in the context of lurk- user communities to acquire application relevant ing is information overload (de Valck et al., 2007; knowledge from peers (Franke and Shah, 2003; Rafaeli et al., 2004) assuming that members can- Jeppesen and Molin, 2003; von Krogh et al., not process the high volume of new posts. Ana- 2003). This requires profound knowledge of the lyzing the mod value revealed that the commu- initial user innovation process though user entre- nity is rather friendly with 60% of the mods be- preneurs are not necessarily user innovators in ing positive, 4% neutral, and 36% showing a the context of the Machinima production process negative attitude towards others’ comments. development. Complementing the technology Correlating a comment’s mod value with its time- application knowledge, new domain knowledge of-posting yielded a strong and significant corre- of the film industry necessary for commercializa- lation for the first 50 comments per thread5 while tion (Michael et al., 2002) is mainly integrated by after this threshold the data points became scat- hiring members from the filmmaker talent pool. tered. This finding is especially relevant for com- This result challenges the perception of the user munity members who intend to improve their innovation process (von Hippel, 1988; 1994; karma level – a member variable based on several 2005) where users hold both, “sticky” knowledge factors whereof one is the accumulated mods about their needs and the innovation relevant received – since early posting has a relatively knowledge for new product development. In the stronger influence on modding than the actual case of Machinima, user entrepreneurs seek and content of the post. From an information systems acquire knowledge from different domains that perspective, this finding implies that filter mecha- they need to obtain and combine prior to com- nisms based solely on member-to-member ratings mercialization. will not yield sufficient results for community User entrepreneurship as identified in this study sponsors such as firms who are skimming com- rests on the user entrepreneurs interest in the new munity activities in order to improve existing technology, continuous learning about the tech- products or gather ideas for innovation. nology, their ability to leverage user communities, and finally the talent to seize an opportunity for commercialization that relates to effectuation 4.3. A Directing Audience: How Specialized (Sarasvathy, 2001) thus the analysis of community Feedback in a Virtual Community of Con- feedback. sumption Stimulates New Media Production While the positive impact of online communities of consumption on firm performance is un- 4.2. Modding as Rating Behavior in Virtual doubted (Algesheimer et al., 2005; Cova and Communities: The Case of Rooster Teeth Pro- Pace, 2006; Kozients, 2007; McAlexander et al., ductions 2002; Schau and Gilly, 2003), empirical evidence Firms who operate online user communities face on how community activity impacts on innova- several challenges. Maybe the most important tion performance of the sponsoring firm is ab- challenge is the retention of community members sent. Since filtering member contributions by over time (Ren et al., under review). Since mem- ratings does not proof feasible (Haefliger et al., bers follow different motives when participating 2009), content-based filtering might warrant new in social network sites compared to when being insights. Coding almost 7,000 comments of active in online communities (cf. Litman, 2005; Rooster Teeth Productions’ community of con- Litman and Pezzo, 2007), integrating SNS fea- sumption showed that members generate special- tures in online communities might yield better ized, unsolicited feedback. 10% of the commu- social embeddedness and thus a sustained com- nity members contribute more then once, which mitment to the community (Algesheimer et al., is basically consistent with prior research on lurk- 2005). The study on the online community of ing behavior (Butler, 1999; Joyce and Kraut, Rooster Teeth Productions showed that modding 2006). – the rating of other members’ comments with a numerical and qualitative value – is a widely ap- 5 In the Rooster Teeth Productions Red vs. Blue community plied practice. Almost half of the community site, every of the 165 episodes has its own thread.

15 FROM GAME TO PLAY PETER JÄGER • 2010

Categorizing the comments revealed that 30% nets, 1999; Nonnecke and Preece, 2000; Rafaeli fall into specialized categories, 15% are consid- et al., 2004). Since lurkers usually exhibit a posi- ered general feedback while the remaining 55% tive attitude towards the community, show inter- constitute noise. Specialized feedback is related to est, and have familiarized themselves with the characters (12%), storyline (13%), technical issues community content, context and norms, stimulat- (4%), and flaws and inconsistencies (0.5%). ing lurkers to actively participate can be of great Studying the content of the comments warrants value compared to acquiring outsiders as new that the sponsoring firm integrated meaningful members. Lurkers and their motives have been feedback into their innovation process in adapt- studied (Ahuja and Galvin, 2003; Butler, 2001; ing production techniques, altering existing prod- Katz, 1998; Nonnecke and Preece, 2003; Preece ucts, or changing the expected course of new et al., 2004; Wasko and Faraj, 2000) but research product development. is mainly descriptive while models are missing Members of communities of consumption – and mechanisms are not yet fully understood. without involvement of the sponsoring firm – Drawing on concepts from personality psychol- thus offer specialized feedback that is of value for ogy, this study links epistemic curiosity as person- the firm since incorporated into products or ality trait of lurkers to their contribution behavior processes. Knowing the products’ history well in online communities. Curiosity, the stable emo- enables community members to comment on the tional tendency driving individuals to seek and possible continuity of product development that acquire new information to various levels of results in product innovation. Members’ technical wanting and liking new information, is applied to expertise combined with knowledge about the establish a process model of lurkers’ online be- product allows users to provide feedback relevant havior. Lurkers primarily lurk while getting famil- for process innovation while a strong interest in iar with the community or when they are in a and current monitoring of product details sup- transitional phase of fade-out (Nonnecke and ports the firm’s quality assurance. These findings Preece, 2003). Once the different stages of lurk- challenge existing theory, which so far distin- ing throughout membership tenure (Joyce and guished between communities of production and Kraut, 2006; Kim, 2000; Wenger, 1988) are un- communities of consumption while the latter derstood, explaining an individual’s propensity to were not expected to contribute to firm innova- de-lurking might become feasible. tion. Curiosity rests on two pillars – curiosity as inher- However, intelligent filter mechanisms for content ent personality trait that explains why individuals selection are still missing. Thus, members have to generally show different levels of curiosity (if) and cope with the overwhelming load of information curiosity as emotional-motivational state that presented by new comments, which might force explains to what extend individuals experience members to lurk (de Valck et al., 2007; Rafaeli et curiosity (how), which is usually stimulated exter- al., 2004). Provided the relevance of feedback as nally. Regarding their information-seeking behav- identified in this study and the high level of lurk- ior, individuals experience curiosity in two ways, ing of up to 90% calls for more research on lurk- as feeling of informational deprivation (D-type ing behavior, especially why members cease to epistemic curiosity) and as feeling of interest (I- contribute after their initial post (cf. Butler, 1999; type epistemic curiosity) (Litman, 2005). Either Joyce and Kraut, 2006), which usually coincides one motivates action since I-type curiosity is di- with their sign-up date. rectly associated with positive feelings when acquiring new information while D-type curiosity is associated with unpleasant feelings of uncertainty 4.4. Curious Lurkers. Linking I/D-type Curios- that are expected to be relieved upon information ity to Contribution Behavior in Online Com- acquisition. Litman (2010) combined both views munities in his model of information-seeking relating I- Studies on member motivation constitute a sig- and D-type curiosity to different levels of wanting nificant part of the online community literature (appetite) and liking (pleasure) new information. (Jeppesen and Fredrikesn, 2006; Moore and I-type curiosity is thereby characterized by rela- Serva, 2007; Rheingold, 1993; Tafjel, 1978; von tively low wanting combined with relatively high Krogh et al., 2010). However, less is known about liking of new information while D-type curiosity lurkers, the passive members of online communi- is characterized by relatively high wanting com- ties, which dominate at levels up to 90% (Kozi-

16 STUDIES OF THE CORE COMPONENTS – CURIOSITY, COMMUNITY AND COMMERCIALIZATION

bined with relatively high liking of new informa- gaged in browsing online community content yet tion. does not suffice to engage them in active partici- Applying Litman’s (2010) wanting-liking model pation. Considering differences in individual of information-seeking to the different stages of traits as well as in momentary emotional- lurking yields interesting insights into lurkers’ motivational states of members may explain why potential behavior. The new, integrated model of members initially join and lurk, join and contrib- lurker dynamics thus theorizes that curiosity as a ute once, or join and contribute continuously. feeling of informational deprivation (D-type epis- Incentivizing members to continuously contribute temic curiosity: need for information) accounts thus depends on their individual levels of trait for members’ active contribution behavior in curiosity since stimulating a sufficient level of online communities while curiosity as a feeling of state curiosity, i.e. stimulating a need for knowl- interest (I-type epistemic curiosity: pleasure of edge, appears to be difficult and costly. acquiring new knowledge) keeps members en-

17 FROM GAME TO PLAY PETER JÄGER • 2010

5. CONCLUSIONS, IMPLICATIONS AND THE WAY FORWARD

“… we found a group of people who really like innovative cool stuff […] and it was really great, it was surprising. I was really amazed and I continue to be amazed at how great they are …”

Interview with Burnie Burns of Rooster Teeth Productions on April 3, 2007

Concluding the work on curiosity, innovation and entrepreneurship as in the case of Machinima entrepreneurship in online user communities, this might render entry barriers to an industry porous last chapter of the dissertation presents the ag- (Porter, 1980), challenge existing business models, gregated and contextually compiled implications or innovate radically new technologies (Christen- for research and management. Limitations as sen et al., 1998) and thereby change industry apply are presented and finally a future research structures. Users who innovate or commercialize agenda is proposed. new ideas thus play an important role in industry “From Game to Play” illustrates the path innova- dynamics (Baldwin et al., 2006; Bogers et al., tive users took when they started off as gamers 2010; Faulkner and Runde, 2009; Haefliger et al., who got curious about others’ use of the games as 2010; Shah and Tripsas, 2007). tools for animation production, acquired the However, user entrepreneurs differ from tradi- knowledge necessary to realize first little projects, tional user innovators in that they approach intel- later staged their own plays in games which they lectual property rights with increased sensibility. presented to fellow community members, and Instead of relying on informal copyright practices after recognizing an opportunity eventually (Lee, 2008), user entrepreneurs seek agreement commercialized their Machinima projects – thus with producer firms early before commercializa- turning into user entrepreneurs. tion. Managing this barrier to commercialization User communities play a vital role for the success well distinguished between successful user entre- of established and new ventures (Algesheimer preneurs and those who failed to establish a sus- and Dholakia, 2006; Armstrong and Hagel, tainable business model (Haefliger et al., 2010). 1996; Williams and Cothrel, 2000). This research The question, how actively firms should monitor contributes to current literature by contemplating their industry as well as adjacent industries to and describing important aspects of user com- identify emerging user innovation and entrepre- munities’ impact on the user entrepreneurships neurial activities remains a challenging one. Costs process (Haefliger et al., 2010), the ways mem- associated with monitoring and enforcing IP bers interact in online communities when social rights are astronomical (Liebeskind, 1996) and network site features are present (Haefliger et al., the initial impact of user entrepreneurial activi- 2009), the value of feedback from a community ties might not be detrimental for incumbent of consumption for firm innovation (Jäger et al., firms. However, once gaining momentum, the 2010), and the relation between curiosity as per- new technology could erode margins of incum- sonality trait and lurking behavior, hence passive bents or shift the power within an entire industry community participation (Jäger, 2010). to new players. Especially the Internet with Web 2.0 technologies featuring user-generated content and an increasing integration of users into online 5.1. Implications for Research and Manage- communities changes the distribution channels of ment various industries, not only media (Hustad and Teigland, 2008; Pascu et al., 2007). Online user communities are the breeding ground for various kinds of activities. While ef- Online communities can shift bargaining power fects like feedback for innovation, incubation of (Porter, 2010) from producers to consumers new talent, and loyalty spreading are highly sup- (Kozinets, 1999). Research to date missed de- portive for firms, others like the inception of user scribing the potential of SNS features in online communities other than social network sites. In-

18 CONCLUSIONS, IMPLICATIONS AND THE WAY FORWARD

cluding such features like personal profiles, rating bers with relatively high levels of D-type trait other members’ comments, or totally new ways curiosity are expected to contribute continuously of mass communication might further increase to the online community. However, this model the power of consumers in that (critical) com- neglects the fact that individual trait curiosity ments, which are posted by credible members are might reside at different levels depending on the better exposed due to high ratings. On the other general interests of the individual. With changing hand, ratings could be used as part of a filtering hobbies, the interest might change and members mechanism for sponsoring firms to assess quality thus move on to other communities. After all, feedback by users. However, the study by Hae- curiosity as an emotional-motivational state con- fliger and colleagues (2009) showed that ratings stitutes the affect that eventually drives an indi- are not sufficient as sole filter criterion yet. vidual’s behavior (cf. Bateman et al., 2006 for Skimming for quality feedback generated by users affective commitment to online communities). of online communities of consumption yields Since cognition impacts on individual behavior valuable insights for sponsoring firms. Jäger and and community design might impact the cogni- colleagues (2010) provided an altered view of tive process (Beenen et al., 2004; McLaughlin, communities of consumption in studying the 1999) community hosts as well as fellow commu- content of users’ comments that proved innova- nity members6 could have a certain influence on tion relevant since delivering direct feedback for members’ contribution behavior. product and process innovation as well as quality assurance. Thus investing in an online community of consumption for pure loyalty-building 5.2. Limitations and Future Research reasons without harvesting the additional “fruits” Most of the research in this thesis relates to the constitutes an opportunity loss. However, as dis- context of Machinima as a new phenomenon. As cussed before, meaningful filter mechanisms have described by Henry Lowood in the interview not been developed yet and coping with the statement above, Machinima is characterized by overwhelming load of comments resembles cost the simultaneity of different enabling aspects that for members and firms alike. all relate to the Internet as new medium. This Motives why users contribute to online commu- might render the generalization of the findings nity activities have been studied in various fields difficult hence calling for future research in simi- and domains (von Krogh et al., 2010). However, lar settings but different to pure Internet ventures. SDT used as predominant framework is appar- For the user entrepreneurial context, studies ently not sufficient to explain all kinds of online should apply the model to different industries (social) behavior (cf. von Krogh et al., 2010). Cu- than media. It would be interesting to evaluate if riosity from the field of personality psychology certain industries have a higher propensity to give might yield new insights why a large number of rise to user entrepreneurship than others and community members take the passive role of what the discrimination characteristics are. This lurkers (Jäger, 2010). Understanding lurking as a might relate to the firms’ preference towards process model in which members lurk at different sharing of complementary assets thus their leni- points in time and thus at different stages ence in letting others experiment with and ap- throughout their community tenure combined propriate their IP or to the availability of user with the wanting-liking model of information communities that affect learning and opportunity seeking (Litman, 2010) allows to not only catego- recognition. Expanding the topic on IP, the study rize lurking behavior but also to predict the possi- revealed that user entrepreneurs are rather open ble chances of re-activating lurkers. Individuals to sharing production relevant knowledge of the with low levels of either I- or D-type trait curios- Machinima process but are highly reluctant to ity are highly unlikely to return to the community sharing product related knowledge, e.g. the conti- after their initial need for knowledge is satisfied. nuity of the story line, etc. with their user com- These members sign up, post a question once and munity. Hence, user entrepreneurs first rely on vanish thereafter. Members with high I-type trait informal copyright practices in phase one of the curiosity but low D-type trait curiosity show a commercialization process but later become pro- certain propensity to return to the community if a sufficient level of D-type state curiosity can be stimulated – these members return from time to 6 Should a strong influence of members’ behavior on cognition exist, member behavior might additionally be time yet do not continuously contribute. Mem- influenced community norms.

19 FROM GAME TO PLAY PETER JÄGER • 2010

tective about their own IP. A finding that is un- Are members who rate but do not post still con- known from the OSS context where usually the sidered lurkers? Does this depend on how valu- same license terms apply for the derivative prod- able ratings are for firms or fellow community ucts. members? Future research on online communities needs to Lurking behavior has been explained by drawing develop viable filter algorithms that enable the on curiosity as concepts from personality psy- sponsoring firm to identify lead users (von Hip- chology. However, the theoretical model has not pel, 1986; 1988; 2005) and categorize comments been tested empirically yet. Validating the model into different classes from noise to specialized would stimulate new approaches to incentivizing innovation relevant information. A possible solu- community members with low levels of D-type tion lies in the further integration of SNS tools trait curiosity to participate while members with into online communities of consumption that low levels of I-type trait curiosity might be per- allow attributing comments to various categories suaded to stay. Regardless of the provided incen- while posting. However, rating behavior by mem- tives – these incentives can only stimulate state bers should be attended to as well since rating curiosity – a significant proportion of members others’ comments might constitute cost for the will contribute relatively less new content and individual. Less is known how rating influences some will always lurk. Thus, from an economical future contribution behavior or whether rating perspective, the costs of incentivizing members to depends on member demographics. New mem- participate increase heavily after a certain thresh- bers might be more inclined to rate others’ com- old has been reached. Assessing this threshold is ments than post comments themselves. This thus of importance. might in turn challenge the definition of lurkers:

20 REFERENCES

REFERENCES

Ahuja, M.K. & Galvin, J.E. 2003. Socialization in Virtual Boyd, D.M. & Ellison, N.B. 2007. Social network sites: Defi- Groups. Journal of Management. 29(2): 161-185. nition, history, and scholarship. Journal of Computer-Mediated Communication. 13(1): 210-230. Algesheimer, R.; Dholakia, U.M. & Herrmann, A. 2005. The social influence of brand community: Evidence from Butler, B.; Sproull, L.; Kiesler, S. & Kraut, R. 2007. Com- European car clubs. Journal of Marketing. 69(3): 19-34. munity effort in online groups: Who does the work and why? In S. Weisband. (Ed.), Leadership at a distance. (pp. 171- Alexy, O. & Leitner, M. 2008. Norms, Rewards, and their 194). Hillsdale, NJ.: Lawrence Erlbaum Associates. Effect on the Motivation of Open Source Software Develop- ers. Technical report, Technical University of Munich. Cha, M.; Kwak, H.; Rodriguez, P.; Ahn, Y.-Y. & Moon, S. 2007. I tube, you tube, everybody tubes: Analyzing the Alvarez, S.A. & Barney, J.B. 2005. How do entrepreneurs world’s largest user generated content video system. In ACM organize firms under conditions of uncertainty? Journal of Internet measurement conference (IMC’07) (pp. 1–14). Management. 31(5): 776–793. Chesbrough, H.W. 2003. Open Innovation: The new im- Alvarez, S.A. & Barney, J.B. 2007. Discovery and creation: perative for creating and profiting from technology. Boston: alternative theories of entrepreneurial action. Strategic Entre- Harvard Business School Press. preneurship Journal. 1(1-2) 11-26. Christensen, C.; Suárez, F.F. & Utterback, J.M. 1998. Strate- Anderson, J.; Williams, N.; Seemungal, D.; Narin, F. & Oli- gies for Survival in Fast-Changing Industries. Management vastro, D. 1996. Humen Genetic Technology: Exploring the Science. 44(12): 207-220. Links between Science and Innovation. Technolog y Analysis & Strategic Management. 8(2): 135-156. Cohen, B.P. 1980. Developing Sociological Knowledge: Theory and Method. Englewood Cliffs, NJ: Prentice Hall. Armstrong, A. & Hagel, J III. 1996. The Real Value of On- Line Communities. Harvard Business Review. May-June: 134- Cohen, W.M. & Levinthal, D.A. 1990. Absorptive capacity: 141. A new perspective on learning and innovation. Administrative Science Quarterly. 35(): 128-152. Bagozzi, R.P. & Dholakia, U.M. 2006. Open source software user communities: a study of participation in Linux user Cova, B. & Pace, S. 2006. Brand community of convenience groups. Management Science. 52(7): 1099-1115. products: new forms of customer empowerment - the case "my Nutella The Community". European Journal of Marketing. Baldwin, C.; Hienerth, C. & von Hippel, E. 2006. How user 40(9-10): 1087-1105. innovations become commercial products: A theoretical investigation and case study. Research Policy. 35 (9): 1291– Cummings, J.N.; Butler, B. & Kraut, R. 2002. The Quality 1313. of Online Social Relationships. Communications of the ACM. 45(7): 103-108. Barney, J.B. 1991. Firm Resources and Sustained Competi- tive Advantage. Journal of Management. 17(1): 99-120. de Valck, K.; Langerak, F.; Verhoef, P.C. & Verlegh, P.W.J. 2007. Satisfaction with virtual communities of interest: Bateman, P.; Gray, P. & Butler, B. 200. Community Com- Effect on members' visit frequency. British Journal of Manage- mitment: How Affect, Obligation, and Necessity Drive ment. 18(3): 241-256. Online Behaviors. International Conference on Information Systems (ICIS) ICIS 2006 Proceedings. Deci, E.L. & Ryan, R.M. 1985. Intrinsic Motivation and Self-determination in human Behavior. New York: Plenum Beenen, G.; Ling, K.; Wang, X.; Chang, K.; Frankowski, D.; Press. Resnick, P. & Kraut, R.E. 2004. Using social psychology to motivate contributions to online communities. Proceedings Dewan, R.; Freimer, M. & Seidmann, A. 2000. Organizing of the 2004 ACM conference on Computer supported co- Distribution Cannels for Information Goods on the Internet. operative work. 212-221. Management Science. 46(4): 483-495. Berlyne, D.E. 1950. Novelty and curiosity as determinants of Eisenhardt, K.M. 1989. Building Theories from Case Study exploratory behavior. British Journal of Psychology. 41(): 50–68. Research. Academy of Management Review. 14(4): 532–550. Berlyne, D.E. 1966. Curiosity and exploration. Science. 153(): Eisenhardt, K.M. & Graebner, M.E. 2007. Theory Building 25–33. from Cases: Opportunities and Challenges. Academy of Man- agement Journal. 50(1): 25–32. Berridge, K.C. 2003. Pleasures of the brain. Brain and Cogni- tion. 52(): 106–128. Faulkner, P. & Runde, J. 2009. On the identity of technological objects and user innovations in function. Academy of Berridge, K.C. 2004. Pleasure, unfelt affect and irrational Management Review. 34(3): 442–462. desire. In A.S.R. Manstead, A.H. Fischer & N. Frijda (Eds.), Feelings and Emotions: The Amsterdam Symposium (pp. Foxall, G.R. & Tierney, J.D. 1984. From CAP1 to CAP2: 43-62). Cambridge, UK: Cambridge University Press. User-initiated innovation from the user’s point of view. Management Decision. 22(5): 3-15. Bogers, M.; Afuah, A. & Bastian, B. 2010. Users as Innova- tors. A Review, Critique, and Future Research Directions. Journal of Management. 36(4): 857-875.

21 FROM GAME TO PLAY PETER JÄGER • 2010

Franke, N. & Shah, S. 2003. How communities support Computer-Controlled Music Instruments. Organization Sci- innovative activities: An exploration of assistance and shar- ence. 17(1): 45-63. ing among end-users. Research Policy. 32(1): 157-178. Jeppesen, L.B. & Molin, M.J. 2003. Consumers as co- Franke, N. & von Hippel, E. 2003. Satisfying heterogeneous developers: learning and innovation outside the firm. Technol- user needs via innovation toolkits: The case of apache secu- ogy Analysis & Strategic Management. 15(3): 363–383. rity software. Research Policy. 32(7): 1199-1215. Katz, J. 1998. Luring the Lurkers. Available at Frey, B.S. & Jegen, R. 2001. Motivation Crowding Theory. http://slashdot.org/features/98/12/28/1745253.shtml Journal of Economic Surveys. 15(5): 589-611. Kim, K.; Kim, G-M. & Kil E.S. 2009. Measuring the Com- Frey, B.S. & Oberholzer-Gee, F. 1997. The Cost of Price patibility Factors in mobile Entertainment Service Adoption. Incentives: An Empirical Analysis of Motivation Crowding- Journal of Computer Information Systems. (Fall): 141-148. Out. American Economic Review. 87(4): 746-755. Kogut, B.; Zander, U. 1992. Knowledge of the Firm, Com- Fueller, J. 2006. Why Consumers Engage in Virtual New binative Capabilities, and the Replication of Technology. Product Development initiated by Producers. Advances in Organization Science. 3(3): 383-397. Consumer Research. 33(): 639-646. Kozinets, R.V. 1997. ‘I want to believe’: a netnography of Fueller, J.; Jawecki, G. & Mühlbacher, H. 2007. Innovation the X-philes’ subculture of consumption. In Advances in Creation by Online Basketball Communities. Journal of Consumer Research, M. Brucks & D.J. MacInnis. 24(): 470- Business Research. 60(1): 60-71. 475. Association for Consumer Research, Provo, UT. Ghosh, S. 1998. Making Business Sense of the Internet. Kozinets, R.V. 1999. E-tribalized marketing? The strategic Harvard Business Review. March-April(): 126-135. implications of virtual communities of consumption. Euro- pean Management Journal. 17(3): 252-264. Glaser, B. & Strauss, A.L. 1967. The Discovery of Grounded Theory. Chicago, IL: Aldine. Kozinets, R.V. 2002. The Field behind the Screen: Using Netnography for Marketing Research in Online Communi- Grant, R.M. 1996. Toward a Knowledge-Based Theory of ties. Journal of Marketing Research. 39(Feb): 61-72. the Firm. Strategic Management Journal. 17(Winter Special Issue: Knowledge and the Firm): 109-122. Kozinets, R.V. 2007. Inno-tribes: Star Trek as Wikimedia. in Consumer tribes, B. Cova, R.V. Kozinets & A. Shankar, 194- Hall, G.S. & Smith, T.L. 1903. Curiosity and Interest. Peda- 211. Oxford and Burlington, MA: Butterworth-Heinemann. gogical Seminary. 10(3): 315-358. Kozinets, R.V.; Hemetsberger, A. & Schau, H.J. 2008. The Haefliger, S.; Reichen, P.; Jaeger, P. & von Krogh, G. 2009. Wisdom of Consumer Crowds. Collective Innovation in the Modding as rating behavior in virtual communities: The Age of Networked Marketing. Journal of Macromarketing. case of Rooster Teeth Productions. Online Communities 28(4): 339-354. and Social Computing, Lecture Notes in Computer Science. 5621(): 197-206. Kozinets, R.V. 2010. Doing Ethnographic Research Online. Sage Publications Ltd. Haefliger, S.; Jäger, P. & von Krogh, G. 2010. Under the radar: Industry entry by user entrepreneurs. Research Policy. Lakhani, K.R. & von Hippel, E. 2003. How open source 39(9): 1198–1213. software works: “Free” user-to-user assistance. Research Policy. 32(): 923-943. Hancock, H. & Ingram, J. 2007. Machinima for Dummies. For Dummies. Lee, K.R. 1996. The role of user firms in the innovation of machine tools: The Japanese case. Research Policy. 25(): 491- Hustad, E. & Teigland, R. 2008. Implementing social net- 507. working media and Web 2.0 in multinationals: Implications for knowledge management. Proceedings of the 9th Euro- Lee, G.K. & Cole, R.E. 2003. From a Firm-Based to a pean Conference on Knowledge Management. 323-331. Community-Based Model of Knowledge Creation: The Case fo the Linux Kernel Development. Organization Science. Jäger, P.M. 2010. Curious Lurkers. Linking I/D-type Curios- 14(6): 633-649. ity to Contribution Behavior in Online Communities. ETH Zurich Working Paper 2010. Lee, E. 2008. Warming up to user-generated content. Univer- sity of Illinois Law Review. 5(): 1458–1548. Jäger, P.M.; von Krogh, G. & Haefliger, S. 2010. A Directing Audience: How Specialized Feedback in a Virtual Commu- Liebeskind, J.P. 1996. Knowledge, strategy, and the theory of nity of Consumption Stimulates New Media Production. the firm. Strategic Management Journal. 17(Winter Special ETH Zurich Working Paper 2010. Issue): 93-107. Janzik, L. & Raasch, C & Herstatt, C. 2009. Motivation in Litman, J.A. & Spielberger, C.D. 2003. Measuring epistemic innovative online communities: Why join, why innovate, shy curiosity and its diversive and specific components. Journal of share? International Journal of Innovation Management. forthcom- Personality Assessment. 80(): 75-86. ing Litman, J.A. & Jimerson, T.L. 2004. The measurement of Jensen, M.B.; Johnson, B.; Lorenz, E. & Lundvall, B.-Å. curiosity as a feeling of deprivation. Journal of Personality 2007. Forms of knowledge and modes of innovation. Re- Assessment. 82(): 147-157. search Policy. 36(5): 680-693. Litman, J.A. 2005. Curiosity and the pleasures of learning: Jeppesen, L.B. & Fredriksen, L. 2006. Why Do Users Con- Wanting and liking new information. Cognition and Emotion. tribute to Firm-Hosted User Communities? The case of 19(6): 793-814.

22 REFERENCES

Litman, J.A. & Pezzo, M.V. 2007. Dimensionality of interpersonal Nambisan, S. 2002. Designing Virtual Customer Environ- curiosity. 43(): 1448-1459. ments for New Product Development: Toward a Theory. Academy of Management Review. 27(3): 392-413 Litman, J.A. 2008. Interest and deprivation factors of epistemic curiosity. Personality and Individual Differences. 44(): 1585- Nonaka, I. 1991. The knowledge-creating company. Harvard 1595. Business Review. 69(6): 96-104. Litman, J.A. 2010. Relationships between measures of I- Nonnecke, B. & Preece, J. 2000. Lurker Demographics: and D-type curiosity, ambiguity tolerance, and need for Counting the silent. Proceedings of the CHI 2000. The closure: An initial test of the wanting-liking model of infor- Hague. ACM. mation-seeking. Personality and Individual Differences. 48(): 397- 402. Nonnecke, B. & Preece, J. 2001. Why Lurkers Lurk. Ameri- cas Conference on Information Systems. 2001. Loewenstein, G. 1994. The psychology of curiosity: A review and reinterpretation. Psychological Bulletin. 116(): 75-98. Nonnecke, B. & Preece, J. 2003. Silent participants: Getting to know lurkers better. In: Lueg, C. & Fisher, D. From Use- Lowood, 2007. High-performance play: The making of net to Cowebs: Interacting with social information spaces. Machinima. In: Clarke, A. & Mitchell, G. (Eds.) Videogames Computer Supported Cooperative Work Series. 2003: 110- and art. Bristol: Intellect Books. 132. Luedicke , M. 2006. Brand Community Under Fire: The Nonnecke, B.; Preece, J.; Andrews, D. & Voutour, R. 2004. Role of Social Environments for the HUMMER Brand Online lurkers tell why. Paper presented at AMCIS 2004. Community. Advances in Consumer Research. 33(): 486-493. New York, NY. Lüthje, C. 2004. Characteristics of innovating users in a Nowotny, H. 2008. Insatiable Curiosity: Innovation in a Fragile consumer goods field: An empirical study of sport-related Future. Translated by M. Cohen. The MIT Press, Cam- product consumers. Technovation. 24(): 683-695. bridge, MA. Lüthje, C.; Herstatt, C. & von Hippel, E. 2005. User- Ogawa, S. 1998. Does sticky information affect the locus of innovators and “local” information: The case of mountain innovation? Evidence from the Japanese convenience store biking. Research Policy. 34(): 951-965. industry. Research Policy. 26(): 777-790. Marino, P. 2004. 3D game-based filmmaking: The art of O’Reilly, T. 2007.What is web 2.0: Design patterns and Machinima. Paraglyph Press business models for the next generation of software. Commu- nications & Strategies. 65(): 17–37. Mason, B. 1999. Issues in virtual ethnography. In Proc. http://ssrn.com/paper=1008839. Ethnographic Studies in Real and Virtual Environments: Inhabited Information Spaces and Connected Communities. Pascu, C.; Osimo, D.; Ulbrich, M.; Turlea, G. & Burgelman, Edinburgh. 1999: 61-69. J.C. 2007. The potential disruptive impact of Internet 2 based technologies. First Monday. 12(3). Massari, L. 2010. Analysis of MySpace user profiles. Informa- tion Systems Frontiers. 12(4): 361-367. Piller, F.T. & Walcher, D. 2006. Toolkits for idea competitions: A novel method to integrate users in new product McAlexander, J.H.; Schouten, J.W. & Koenig, H.F. 2002. development. R&D Management. 36: 307-318. Building Brand Communitiy. Journal of Marketing. 66(Jan): 38-54. Porter, C.E. 2006. A Typology of Virtual Communities: A Multi-Disciplinary Foundation for Future Research. Journal McLoughlin, C. 1999. Culturally responsive technology use: of Computer-Mediated Communication. 10(1): 00-00. developing an on-line community of learners. British Journal of Educational Technology. 30(3): 231–243. Porter, M.E. 1980. Competitive strategy. Techniques for analyzing industries and competitors. New York: Free Press. Michael, S.; Storey, D. & Thomas, H. 2002. Discovery and coordination in strategic management and entrepreneurship. Preece, J.; Nonnecke, B. & Andrews, D. 2004. The top 5 In: Hitt, M., Ireland, R.D., Camp, S.M., Sexton, D.L. (Eds.), reasons for lurking: Improving community experiences for Strategic Entrepreneurship. Blackwell, Oxford, pp. 45–65. everyone. Special Issue of Computers in Human Behavior: An Interdisciplinary Perspective. 20(2). Mollick, E. 2005. Tapping Into the Underground. MIT Sloan Management Review. 46(4): 21-24. Rafaeli, S.; Ravid, G. & Soroka, V. 2004. De-Lurking in Virtual Communities: A Social Communication Network Moore, T.D. & Serva, M.A. 2007. Understanding Member Approach to Measuring the Effects of Social and Cultural Motivation for Contributing to Different Types of Virtual Capital. HICSS vol. 7, pp.70203, Proceedings of the 37th Online Communities: A Proposed Framework. SIGMIS Annual Hawaii International Conference on System Sci- CPR 2007: Proceedings fo the 2007 ACM SIGMIS CPR ences (HICSS'04) - Track 7, 2004 Conference Global Information Technology Workforce. 153- 158. Ranchhod, A. & Gurau, C. 1999. Internet-enabled distribution strategies. Journal of Information Technology. 14(): 333-346. Morris, D.; Kelland, M. & Lloyd, D. 2005. Machinima. Course Technology PTR Read, S.; Song, M. & Smit, W. 2009. A meta-analytic review of effectuation and venture performance. Journal of Business Muniz, A. M. & O'Guinn, T. C. 2001. Brand community. Venturing. 24(6): 573–587. Journal of Consumer Research. 27(4): 412-432. Ren, Y.; Harper, F.M.; Drenner, S.; Terveen, L.; Kiesler, S.; Riedl, J. & Kraut, R.E. XXXX. Increasing Commitment to

23 FROM GAME TO PLAY PETER JÄGER • 2010

Online Communities: Designing Theory. Under Review at Spaeth, S.; Stuermer, M. & von Krogh, G. 2010. Enabling MIS Quarterly. knowledge creation through outsiders: towards a push model of open innovation. International Journal of Technology Manage- Rheingold, H. 1993. The Virtual Community (1st. ed.). ment. (forthcoming) Addison-Wesley Pub. Co. Spielberger, C.D. 1975. Anxiety: state-trait-process. In C.D. Roberts, J.A.; Hann, I-H. & Slaughter, S.A. 2006. Under- Spielberger & I.G. Sarason (Eds.), Stress and anxi- standing the Motivations, Participation, and Performance of ety.Washington: Hemisphere/Wiley. Open Source Software Developers: A Longitudinal Study of the Apache Projects. Management Science. 52(7): 984-999. Spielberger, C.D.; Peters, R.A. & Frain, F. 1976. The State- Trait Curiosity Inventory. Unpublished manual, University of Rogers, E.M. 1962. Diffusion of Innovations. Free Press, Glen- South Florida, Tampa. coe. Strauss, A.L. & Corbin, J. 1990. Grounded Theory Re- Roth, C.; Vorderer, P & Klimmt, C. 2009. The Motivational search: Procedures, Canons and Evaluative Criteria. Appeal of Interactive Storytelling: Towards a Dimensional Zeitschrift für Soziologie. 19(6): 418 ff. model of the User Experience. In: I.A. Iurgel, N.Zagalo, and P.Petta (Eds.): ICIDS 2009, Lecture Notes in Computer Stuermer, M.; Spaeth, S. & von Krogh, G. 2009. Extending Science 5915: 38-43. private-collective innovation: a case study. R&D Management. 39():170-191. Rowley, J.; Kupiec-Teahan, B. & Leeming, E. 2007. Cus- tomer Community and Co-creation: A Case Study. Marketing Subrahmanyam, K. & Greenfield, P. 2008. Online Commu- Intelligence & Planning. 25(2): 136-146. nication and Adolescent Relationships. The Future of Children. 18(1): 119-146. Sakkab, N.Y. 2007. Growing through Innovation. Research- Technology Management. 50(6): 59-64. Tafjel, H. 1978. Differentiation between social groups. Lon- don: Academic. Sarasvathy, S. D. 2001. Causation and effectuation: Toward a theoretical shift from economic inevitability to entrepre- Teece, D.J. 1986. Profiting from technological innovation: neurial contingency. The Academy of Management Review. 26(2): implications for integration, collaboration, licensing and 243–263. public policy. Research Policy. 15(): 285–305. Sarasvathy, S. 2004. Making it happen: Beyond theories of Usmani, Z.; Alghamdi, F.Q.; Tariq, A. & Puri, T.N. 2010. I the firm to theories of firm design. Entrepreneurship Theory and Know What You Did This Summer – Users’ Behavior on Practice. 28(6) 519-531. Internet. In: T.Sobh (Ed.) Innovations and Advances in Computer Sciences and Engineering. Springer Sci- Sawhney, M.; Verona, G. & Prandelli, E. 2005. Collaborat- ence+Business Media. ing to create: The Internet as a platform for customer engagement in product innovation. Journal of Interactive Market- Varadarajan, R.; Yadav, M.S. & Shankar, V. 2008. First- ing. 19(4): 4-17. mover advantage in an Internet-enabled market environment: conceptual framework and propositions. Journal of the Schau, H. J. & Gilly, M. C. 2003. We are what we post? Self- Academy of Marketing Science. 36(): 293-308. presentation in personal Web space. Journal of Consumer Research. 30(3): 385-404. Venkataraman, S. 1997. The distinctive domain of entrepreneurship research. In Advances in Entrepreneurship, Schoberth, T.; Heinzl, A.; & Preece, J. 2006. Exploring Firm Emergence, and Growth, Vol. 3, Katz, J. (ed). JAI communication activities in online communities: A longitu- Press: Greenwich, CT; 119-138. dinal analysis in the financial services industry. Journal of Organizational Computing and Electronic Commerce. 16(3-4): 247- von Hippel, E. 1976. The dominant role of users in the 265. scientific instrument innovation process. Research Policy. 5(): 212-239. Schumpeter, J.A. 1934. The Theory of Economic Development. Harvard University Press, Cambridge. von Hippel, E. 1986. Lead users: A source of novel product concepts. Management Science. 32(): 791-805. Shah, S.K. 2006. Motivation, Governance, and the Viability of Hybrid Forms in Open Source Software Development. von Hippel, E. 1988. The Sources of Innovation. Oxford Uni- Management Science. 52(7): 1000-1014. versity Press. Shah, S. & Tripsas, M. 2007. The accidental entrepreneur: von Hippel, E. 1994. “Sticky information” and the locus of The emergent and collective process of user entrepreneur- problem solving: Implications for innovation. Management ship. Strategic Entrepreneurship Journal. 1: 123–140. Science. 40(): 429-439. Shane, S. & Venkataraman, 2000. The promise of entrepre- von Hippel, E. & von Krogh, G. 2003. Open source software neurship as a field of research. Academy of Management Review. and the “Private-Collective” innovation model: Issues for 25(1): 217-226. organization science. Organization Science. 14(2): 209–223. Smith, M.D. & Telang, R. 2009. Competing with free: The von Hippel, E. 2005. Democratizing innovation. Cambridge. Impact of Movie Broadcasts on DVD Sales and Internet MA: MIT Press. Piracy. MIS Quarterly. 33(2): 321-338. von Hippel, E. 2007. Horizontal innovation networks – By Spaeth, S.; Haefliger, S.; von Krogh, G. & Renzl, B. 2008. and for users. Industrial and Corporate Change. 16(2): 293–315 Communal resources in open source software development. Information Research. 13(1): . von Krogh, G. & von Hippel, E. 2003. Special issue on open source software development. Research Policy. 32(): 1149-1157.

24 REFERENCES

von Krogh, G. & von Hippel, E. 2006. The promise of Wellman, B. & Gulia, M. 1999. Net surfers don’t ride alone: research on open source software. Management Science. 52(): Virtual communities as communities. In: P. Kollock & M. 975-983. Smith (Eds.): Communities and Cyberspace. Routledge, New York. von Krogh, G.; Haefliger, S.; Spaeth, S. & Wallin, M. 2010. Open Source Software: A Review of Motivations to Con- Wellman, B. 2001. Computer Networks As Social Networks. tribute. ETH Zurich Working Paper. Science. 293(September): 2031-2034. Wasko, M.M. & Faraj, S. 2000. “It is what one does”: why Wernerfelt, B. 1984. A Resource-Based View of the Firm. people participate and help others in electronic communities Strategic Management Journal. 5 (2): 171-180. of practice. Journal of Strategic Information Systems. 9(): 155-173. Yin, R.K. 2003. Case study research: design and methods. Thou- Webb, K.L. 2002. Managing channels of distribution in the sand Oaks, CA: Sage. age of electronic commerce. Industrial Marketing Management. 31(): 95-102.

25 FROM GAME TO PLAY PETER JÄGER • 2010

UNDER THE RADAR

INDUSTRY ENTRY BY USER ENTREPRENEURS

Stefan Haefliger Peter Jäger Georg von Krogh

Chair of Strategic Management and Innovation Department of Management, Technology, and Economics (MTEC) ETH Zürich, Switzerland

Published in Research Policy vol. 39 issue 9 2010 page 1198-1213

We inductively develop a model of the commercialization process for new products or services user entrepreneurs undertake when entering an industry while drawing on proprietary technology developed in another industry. Extending the growing field of user entrepreneurship, we identify a two-phase approach to industry entry by user entrepreneurs who start “under the radar” of incumbent firms, gain experience, attract a first potential customer base, and then, in a second phase, engage in commercialization. During this process, a community of fellow users is of major importance for the entrepreneur, serving as a knowledge pool for skills development and experimentation with different commercialization paths. We study a nascent group of firms founded by users of video games who became entrepreneurs on entering the animation industry by producing Machinima, a new film genre characterized by shooting film in video games. We explain how user entrepreneurs gain access to complementary assets (video games) for their new use (shooting film), how they deal with intellectual property issues when using other firms’ assets, and how user entrepreneurs combine domain knowledge about film production with their experience in video games and the art of Machinima. Our propositions hold implications for management and policy.

27 FROM GAME TO PLAY PETER JÄGER • 2010

1. INTRODUCTION

Entrepreneurship, conventionally understood, is a commercialization occurred in the industry in process where opportunity recognition precedes which the original user activity took place. In the prototype development (Venkataraman, 1997; case of juvenile products, dissatisfied parents Shane and Venkataraman, 2000). In the case of innovated more sophisticated designs, like a user entrepreneurship, however, this process is stroller that can be used while jogging, which they reversed: users first develop prototypes and, while eventually commercialized after other parents using and gaining experience with the new de- became aware of its superior performance. In the sign, recognize a potential for commercialization case of rodeo kayaking, users redesigned kayaks of their product7 or service (Shah and Tripsas, to better perform under extremely challenging 2007). Thus, user entrepreneurship can be seen conditions. In either case, user entrepreneurs as taking a position within a standing debate in developed new concepts on the back of existing entrepreneurship theory about the discovery or products. They demonstrated and promoted their creation of opportunities (Alvarez and Barney, designs in competitions. In scanning for products 2005, 2007; Sarasvathy, 2001, 2004) by describ- that satisfied their needs prior to developing their ing entrepreneurial opportunities as created in own, user entrepreneurs gained familiarity with use. Generally showing traits of user innovators the general market structure and the competitive (von Hippel, 1988), user entrepreneurs derive environment of their industry. their designs from existing products or technolo- Products and services are fungible in their appli- gies. If commercial value is created in a different cation. They can be modified, developed, or re- industry from that in which the original product interpreted to be used as a tool or basis for new was located, we should expect new challenges products in another context (Faulkner and connected to technology diffusion and knowledge Runde, 2009; von Hippel, 1988). User entrepre- recombination (Geroski, 2000). The aim of this neurs may shift the use activity from one industry paper is to fill a gap in the fast growing literature to another or, to put it differently, diffuse technol- on user entrepreneurship, by theorizing about the ogy across industries through different use (cf. process of user entrepreneurship as users move Rogers, 1962). Research has so far neglected this from one industry to another in order to com- important aspect: user entrepreneurs who de- mercialize. User entrepreneurs face low opportu- velop products based on assets from one industry, nity costs and exhibit a high willingness to ex- which they apply as complementary assets in the periment and high potential to explore commer- industry where they commercialize. Based on a cial opportunities by entering existing markets or study of the Machinima phenomenon, a new creating new ones, especially when the target animation genre, and a sample of firms started markets are turbulent and demand is uncertain by users, we develop a new process model of user (Shah and Tripsas, 2007). Usually embedded in a entrepreneurship across industries. The firms in community of users with similar needs, user en- our sample apply video games as production trepreneurs operate under favorable conditions: technology for animation and commercialize the community plays a vital role in diffusing new their films via online distribution or DVD sales in designs while user entrepreneurs are granted the animation production industry, a subdivision early access to feedback and information relevant of the motion picture industry. In the 1990s, user to commercialization prior to firm foundation innovators started to record their game play and (Shah and Tripsas, 2007). introduced recording technology to video games, Examples of user entrepreneurship have been transforming their use (Faulkner and Runde, studied in the fields of sporting equipment 2009). With the publication of the first Ma- (Baldwin et al., 2006; Luthje et al., 2005; Franke chinima films, shot in video games, the technol- and Shah, 2003) and juvenile products (Shah and ogy became amenable to story telling that ex- Tripsas, 2007). Both fields present cases where tended beyond the story elements contained in the game. User entrepreneurs publish animated

7 To reduce complexity throughout the paper, we use the term “product” to refer to products, technologies, or processes.

28 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

shorts8 using video games as production technol- ucts or assets remain with the original producer ogy and create opportunities to commercialize and create legal uncertainty for the user entre- from their experience with Machinima. Figure 1 preneur (Depoorter, 2009). Companies holding shows the process of user entrepreneurship in the IP of products affected by such user activity two phases as we observe and develop it in this may show tolerance toward the application of paper, where an initial user innovation in industry their assets (Harhoff et al., 2003); otherwise, A enables user entrepreneurs to enter industry B, permission may be granted through a variety of in a first phase, under the radar of incumbents means, such as research exemption in the field of and, in a second phase, to commercialize ani- science (O’Rourke, 2000; Strandburg, 2008), fair mated shorts based on the new use of a technol- use under US copyright law, or the growing prac- ogy from industry A. For example Rooster Teeth tice of (informal) unauthorized use of copyright Productions, a Machinima firm in our sample, material (Lee, 2008). made use of the game Halo, produced by Micro- While firms in the industry where the original soft/Bungie, to create a Machinima franchise9 user activity was located (industry A) may apply called Red vs. Blue under which they publish and enforce patents or copyright to secure profits episodes online. Based on their experience and in their home market, they might adopt a more audience gained, they subsequently sold DVDs, lenient position toward others’ exploitation of sponsorship access, and merchandising. these assets in industries where they do not compete (industry B), especially given the high cost and effort of monitoring and enforcing intellectual property rights (Lie- beskind, 1996). Hence, firms may be willing to share selectively and tolerate the application of specific assets that are core to their “home” market but complementary in other markets, and so provide user entrepreneurs with a foundation to commercialize. The knowledge needed to commercialize in an- Figure 1. Process of user entrepreneurship in two phases other industry’s markets may extend beyond the knowledge acquired in use and, possibly, beyond Products, goods, or services are usually sold or the experience available through the community licensed under conditions that restrict their use or of users (Baldwin et al., 2006). Domain knowl- application, such as enduser license agreements edge important to commercialization includes (EULAs). User entrepreneurs drawing on existing market-relevant education and experience, pro- products experience use restrictions as obstacles duction techniques, work flows and processes, to commercialization. Relying on third party insights about genres and market demand, and assets complicates commercialization; intellectual industry-specific marketing knowledge. We study property rights attached to either modified prod- commercialization by user entrepreneurs outside the industry where the products they use origi- 8 Animated shorts are the predominant products while nated. This offers new insights into tolerance (feature) films appeared rarely, especially early on. toward application of IP, opportunity creation by Throughout this paper, we use the terms film, animation, user entrepreneurs in new industries with frag- and shorts interchangeably. mented markets, and community support across 9 Walt Disney pioneered this marketing concept evoking industries. The particular mix of competition, various sources of revenue based on their characters (Wasko et al., 1993; Yoon and Malecki, 2009). versatility of commercialization-relevant assets

29 FROM GAME TO PLAY PETER JÄGER • 2010

across industries, intellectual property rights, and both industries show similar characteristics and the role of knowledge acquisition raises an im- boundaries are blurring (Calantone et al., 2010; portant question that we address in this paper: Eliashberg et al., 2006; Yoon and Malecki, 2009). Under what conditions do users, who apply (se- The animation industry is probably the closest lectively shared proprietary) assets from one in- link to the video game industry, in that both share dustry, commercialize in markets of another in- substantial talents (Aoyama and Izushi, 2003; dustry (thus becoming user entrepreneurs)? Izushi and Aoyama, 2006). With the increasing We examine this question in the motion picture digitalization of the value chain, entry barriers to industry, which represents an ideal context to the animation industry are lowering, enabling study user entrepreneurship, considering Shah everyone with a personal computer to participate and Tripsas’ (2007) proposition that markets with (Eliashberg et al., 2006). Internet and animation high turbulence and demand uncertainty favor production technologies overlap (Britton et al., user entrepreneurship: besides a huge main- 2009). Producing animated shorts is considered a stream market for theatrical film production, point of entry for small studios that might later several niche markets within the industry are in a attract a growing audience. Being able to shift state of revolt due to technological advances. among markets means that these animation stu- Today, major studios show little interest in provid- dios can eventually move into feature films, as ing content for the Internet even though there is AKOM demonstrated with The Simpsons (Yoon excess demand for it and customers are willing to and Malecki, 2009). pay. Industry analysts have commented that the Our study extends work on user entrepreneurship low profitability of the Internet market, com- by defining core constructs and explaining the pared to their mainstream business, has meant commercialization patterns of users who create that major studios have missed the opportunity to or enter new markets in different industries. Ef- develop sustainable business models to serve on- fectual strategies that assume that opportunities line customers (Papies and Clemet, 2008). This emerge when created by an entrepreneur have neglect may prove critical as the Internet is gain- been positively associated with venture performing in importance as an outlet for media content ance (Read et al., 2009). We pay special attention (Scott, 2004; Yoon and Malecki, 2009). Further- to the strategies users follow to remedy legal un- more, it is still unknown whether theatrical and certainty when applying borrowed assets, the new non-theatrical outlets substitute or complement knowledge they need to acquire, and the support each other (Eliashberg et al., 2006). “The devel- they receive from their community of peers. opment of new delivery systems will in principle Based on case studies, we inductively generate a open up the market to more effective contestation model describing key elements of the strategies by smaller independent film production and dis- user entrepreneurs follow when commercializing tribution companies (cf. Leyshon, 2001). Thus, products or services in new and economically the eventual attainment of film distribution by relevant markets. means of the Internet will no doubt give rise to a After introducing our research design, we de- great increase in the amount of cinematic mate- scribe the relevant cases and present the results of rial available to consumers, thereby widening the our study in the form of descriptive propositions. market and almost certainly making inroads on We conclude with a discussion of our findings, blockbuster audiences” (Scott, 2004: 58). the implications for research, management, and The motion picture industry can expect to see policy, and outline a future research agenda for new entrants from the video game industry since user entrepreneurship and strategy.

30 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

2. SAMPLE AND RESEARCH DESIGN

Machinima offered an ideal context to explore sufficient complexity without creating too much the research question of this paper because (1) data. Such a sample size allows researchers to user entrepreneurs could be observed entering describe and analyze the cases in a systematic and the animation industry10 over the past ten years methodical manner, leading to thorough contex- while the genre was in the process of emerging; tual interpretation. (2) the animation industry was traditionally characterized by high entry barriers, leading to “creative” entry strategies; (3) Machinima 2.1. Sample production involves proprietary software as well The production of animated films used to be as artwork, allowing a nuanced observation of restricted to media professionals who could afford how users manage IP conflicts; (4) users the expensive software packages needed. These frequently possess advanced gaming skills but restrictions led users in the animation industry to need to acquire film production knowledge to produce films with games, inspired by innovative sustain a business in the target industry; and (5) gamers who developed methods to record their users display high variance in their entrepreneu- game play. Such games are relatively cheap com- rial approaches over time, ranging from product pared to traditional production tools. In addition, sales to diversification into consulting, software most of the in-game assets, like characters and development, and online services.11 landscapes that resemble actors and scenes, are Our research comprised three phases: case sam- already available, thus reducing overall produc- pling, data gathering, and data analysis. We con- tion cost and time.12 ducted a multiple, non-embedded case study Defined as “shooting film in a real-time 3D envi- (Yin, 2003), and gathered data from seven firms ronment” (AMAS, see footnote 6), Machinima is within the Machinima community representing (1) a production technology, and the entire population of Machinima-based busi- (2) the name for the genre. It is deeply rooted in nesses at the time of this study. We follow an in- the gaming culture where gamers, early on, expe- ductive logic to theorize about industry entry and rienced the need to record, edit, and distribute market fragmentation by user entrepreneurs and proof of their gaming skills on film to demon- generate propositions derived from the cases strate their proficiency as gamers. Adding story (Cohen, 1980; Eisenhardt, 1989; Eisenhardt and elements to their films (FK—see Appendix for Graebner, 2007; Glaser and Strauss, 1967; full names and affiliation of interview partners), Strauss and Corbin, 1990). Theory development Machinima users in the animation industry later on a small sample size has been discussed by introduced a new genre that can be clearly distin- March et al. (1991), Eisenhardt (1989), and Sig- guished from traditional animation (Mezias and gelkow (2007), while exemplary works with the Mezias, 2000; Peretti and Negro, 2007) when methodology include Vaughn (1990), Lawrence they produced low-cost films for themselves or et al. (2002), and Pervez et al. (2008). Eisenhardt close friends (Morris et al., 2005). The unique (1989: 545) proposes “a number between four characteristics of the Machinima production and ten cases to usually work well” and allows for process enabled users to become entrepreneurs, applying gaming technology in the animation 10 According to the International Standard Industrial industry as well as related industries, like film Classification of All Economic Activities, Rev.4 (ISIC), distribution and production support. The cases in software game development (6201) belongs to another our sample cover all aspects of the product de- industry than motion picture production and distribution velopment and commercialization process. (5911, 5912, 5913, and 6020, respectively). The first two digits are sometimes used to denote a specific industry. (See Farjoun, 1994, for a discussion of the relatedness of industries in terms of knowledge, an aspect we use to 12 Estimates of costs alone show that Machinima production define complementary assets and show the market entry.) amounts to a fraction of animation production. See also 11 A number of publications (Marino, 2004; Morris et al., an article by the BBC 2005; Hancock & Ingram, 2007) cover the topic of http://news.bbc.co.uk/1/hi/technology/7045018.stm Machinima for the general reader. (October 26, 2007).

31 FROM GAME TO PLAY PETER JÄGER • 2010

Table 1. Sample of firms for the cases

Every important step in the history of Ma- firm, Machinima.com, started producing anima- chinima was initiated and conducted by users tion later on to supplement their portfolio. Thus, who played games and experimented with the all firms once entered the animation industry. Of Machinima production technology. Some of the three firms that altered their sources or reve- these users went on to form firms, including nue, one chose to turn away from Machinima. Rooster Teeth Productions, the ILL Clan, We conducted literal and theoretical replication Strange Company, Machinima.com, Bong + (Yin, 2003) by considering firms with operations Dern, and Fountainhead Entertainment. Table 1 in the animation industry as well as firms that provides an overview of the samples in this study. developed alternative sources of revenue different Interviewees brought these seven firms, which we than, but still based on, the Machinima produc- later identified as the population, to our attention tion experience. The fact that user entrepreneurs as predominant examples of Machinima-based show a tendency to display their capabilities at commercialization. Thus, we defined “successful” championships (Baldwin et al., 2006), like the Machinima firms according to informants annual AMAS Film Festival, provides further (Brown and Eisenhardt, 1997). All seven firms confidence that our sample covers the entire had produced Machinima films and won at least population at the time of data gathering. one award at an AMAS Film Festival.13 Three Within the community, the visibility of Strange started by producing animation and continued to Company, the ILL Clan, and Machinima.com do so; another three altered their sources of reve- was extraordinarily high due to their involvement nue after having produced Machinima; and one in the formation of Machinima. Rooster Teeth Productions was the prime example of a firm 13 A ﬁlm festival held by the Academy of Machinima Arts that generates revenues solely from the produc- and Sciences (AMAS) tion of films and related products common to this

32 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

industry. Fountainhead Entertainment used to be munity defined itself, who participated, the mo- a Machinima pioneer in producing animation as tives of the different users, and (most importantly) well as supplying a production tool.14 It originally active firms as well as those that had ceased op- focused on a Machinima-centered business and eration. Understanding the phenomenon helped then refocused its activities on mobile phone us establish more effective relations with key in- games. Short Fuze first released the film Desert formants. Hosting Machinima films, we retrieved Combat: James Bond—No License in 2004; later, plenty of information from community websites15 Matt Kelland and Dave Lloyd wrote the first including the names of the producer, the director, book describing Machinima for the general and the year of publication. Moreover, some of reader (Morris et al., 2005). As a result of techni- the films and credit files provided information cal problems experienced during the production about the production process, individuals, and of their film and their observation of the com- firms involved. Altogether, we read 32 articles, munity’s needs, Short Fuze began to develop studied a 340-page report on the video game Moviestorm, an easy-to-use Machinima produc- industry, watched more than 100 short films, and tion tool, for which they received £450,000 in browsed roughly 50 web pages to gain a thorough seed funding in 2005, followed by a first round of understanding of the phenomenon (for the use of funding of £950,000 in 2007. This Spartan Life, rich information sources, see Vaughan, 1990). produced by Bong + Dern productions, is an Second, one of the authors participated in a four- award-winning virtual talk show that received day Machinima workshop to conduct field obser- substantial publicity because of its high-profile vations and build relations with the Machinima guests from game companies. community. The workshop covered the entire Although the Electric Sheep Company (ESC) Machinima production process and the author hired the workforce of the ILL Clan in 2007, we created a film to gain first-hand experience of the regard the ILL Clan as a separate entity in our process. sample for three reasons: (1) the ILL Clan had Third, in November 2006, we identified inter- been producing Machinima since 1997, that is, viewees from a variety of backgrounds and began for most of our observation period; semi-structured interviews. This approach al- (2) the founders retained their brand and contin- lowed us to react to replies and adapt the ques- ued to make their existing and new animation tions to a candidate’s profile. We usually started products available under this label; and (3) the the interview with questions about an individual’s group continued to produce Machinima for ESC, background and education to decide later on which operated in virtual worlds such as Second whether the interviewee was an objective ob- Life. server or a key informant on a topic. The initial The resulting sample combines all incorporated set of questions, based mainly on desk research, Machinima businesses that were at least a year was tested during the workshop. The question- old. Their common denominator is the entry of naire was subsequently refined and tailored to the its founders into the animation industry as Ma- specific background of each interviewee, based chinima users at one point in time, ultimately on information taken from online résumés or leading to commercialization. The firms differ in previous interviews. terms of their commercialization activities, fi- Fourth, before entering the second round of in- nancing, size, age, ambitions, and goals. We also terviews we analyzed our preliminary results. included one firm that discontinued its Ma- Having a general understanding of Machinima chinima-related revenue source. and the community that supported it, we focused on user entrepreneurs, some of their legal advi- sors, user innovators who played an important 2.2. Data gathering role during the development of Machinima— Data gathering took place in five phases, includ- people who provided valuable background ining both real-time observations and retrospective formation, and finally games companies to com- data (Brown and Eisenhardt, 1997). Desk re- plement the picture. From February to March search delivered insights on how the user com- 2007 we completed seven interviews, one face-to- face, and another six conducted by telephone (see

14 Machinimation is a real-time 3D ﬁlmmaking software add-on representing a dedicated solely to Machinima. 15 Such as www.Machinima.com, www.mprem.com modiﬁcation (mod) of id Software’s Quake III Arena. (Machinima Premiere) or www.gamevideos.com.

33 FROM GAME TO PLAY PETER JÄGER • 2010

Table A2 in the Appendix for a full list and de- 2.3. Data analysis scription of interview partners). We compiled individual case studies, based on the Fifth, another author travelled to New York and data gathered from the five phases. First, bearing Texas and conducted a total of ten interviews in mind the process of traditional user entrepre- and participant observation during a week on site neurship from existing literature, the data were with Rooster Teeth Productions, the ILL Clan, prescreened to derive a common coding scheme Fountainhead, and an ESC representative. Daily (see Appendix). Our interviewees’ frequent and records of working routines, the Machinima pro- unexpected references to IP issues and legal un- duction process, and other office tasks were kept; certainty indicated that the entire process of notes were taken while attending meetings, lunch, commercialization was greatly influenced by the and evening events (Brown and Eisenhardt, legal aspects of game engine use. Hence, we con- 1997). Five formal interviews were conducted, sidered relevant literature in this field to support three of which were recorded on video for later the coding scheme. analysis and classroom use. In addition, more A rough sketch of what industry entry involved than 100 photographs of Machinima working supported us in describing the phenomenon to environments were taken during the field trip. interviewees. We then coded the transcribed in- In total we conducted 25 interviews, 21 in Eng- terviews using MAXQDA, a software tool for text lish and four in German, each lasting 55 min on analysis. While analyzing the interviews, we average. We transcribed 19 interviews (20–30 coded statements in the text, which allowed us to pages each) verbatim. In addition, we cross- sort and evaluate information. Two researchers checked relevant information with other inter- working in parallel conducted the coding. After view data or facts from desk research incorporat- the initial coding, results were merged and the ing external links and comments in the text. second coder recoded selected interviews, con-

Table 2. Overview of propositions and grounding in cases

34 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

tributing to inter-coder reliability. construct was obtained for each case. This led us Applying an iterative process with an overlap of to drop one early construct but to formulate an- data analysis and data collection (Eisenhardt, other, enabling further consolidation of the cod- 1989; Glaser and Strauss, 1967), we used the ing scheme. The final case study write-ups facili- distilled interview data, including the codes and tated the comparison of all ventures’ positions in the higher code categories (e.g.: 2 Domain knowl- terms of user entrepreneurship. Searching for edge, 2.1 Education, 2.2 Work experience, 2.3 cross-case patterns enabled us to extract the gen- Prior film production), as well as secondary in- eral sequence of actions users followed to com- formation, to refine our case studies by gathering mercialize their products or services as well as to additional data whenever gaps were identified identify the different stakeholders or stakeholder during coding. We evaluated the interview data groups involved in the different phases of the and compared coding of key informants (e.g., commercialization process (Eisenhardt, 1989). director) with objective observers’ (e.g., lawyer) Based on the flow of actions, we derived five views where applicable. We triangulated the find- propositions and drew up a table summarizing ings from the interviews with our observation the findings for each (see Table 2). To enhance records (Pettigrew, 1988), independent informa- construct validity, all propositions were compared tion gathered from the Internet, image, and video and discussed in the light of the existing litera- material, and third-party newspaper articles. We ture. The maintenance of individual case studies produced detailed case write-ups for each firm to during the analysis supported replication across cope with the magnitude of data. While structur- firms. A final working paper was sent to all ining and analyzing the within-case data, we formants for comments and feedback, which were checked whether information depicting every integrated with the text.

35 FROM GAME TO PLAY PETER JÄGER • 2010

3. TOWARD A PROCESS MODEL OF INDUSTRY ENTRY BY USER ENTREPRENEURS

In this section we present findings on how, in the “Initially we worked in Burnie’s spare room in Burnie’s case of Machinima, users become entrepreneurs house. And it was a small room [...] and it wasn’t so who commercialize their products or services bad when we were just Burnie and I. But you know, two based on assets from another industry in a two- people turning into five people in a room that’s maybe phase process. We inductively develop a model [...] 250 square feet became pretty cramped, and at some from the cases and formulate five propositions point Red vs. Blue [...] became successful enough, where (Table 2) to explain user entrepreneurs’ behavior I quit the day job and devoted full time to Red vs. Blue. and the consequences of their activities (Fig. 2, Burnie eventually did it as well.” (GR) explained throughout the text). We emphasize the interactions between user entrepreneurs and While one might argue that users are not located firms that selectively share complementary assets, in a specific industry without commercialization, between user entrepreneurs and their community they nonetheless produce and distribute anima- of peers who exchange knowledge about the tion. This compares to cases of open source soft- application of complementary assets, and be- ware, where no commercial intentions may exist tween user entrepreneurs and talent from the but where market share of incumbent software motion picture industry to acquire domain companies is affected. Geoff Ramsey’s anecdote knowledge. First, we describe the context of in- provides a typical illustration of the transition dustry entry, which occurred in two phases. from phase one to phase two. “Machinima was a very cheap way to produce animation or, you know, any kind of narrative, but as it becomes 3.1. Under the radar of incumbent firms: a more popular the cost increases. And, unfortunately, the two-phase process of industry entry more popular you are the more money you generate ...[and The first phase of industry entry consists of a the more] bandwidth you’re paying for people to be able to lateral move from one industry to another under watch your series. ...So we were always looking for ways the radar of incumbents: When first applying to help offset costs. So, I reckon June of 2003, this is video games as tools to produce animated shorts, around episode 12 of Red vs. Blue, I put a thread in the forums: I said, if I made a Red vs. Blue T-shirt, just a users emerged from the domain of the video white T-shirt with a logo on it, would you be interested in game industry to enter and become producers in buying it? And the response was overwhelming. And so I the animation industry where they used games as made a T-shirt and people bought it.” (GR) alien production tools—most of them continuing to be avid gamers. Neither companies in the In phase two, user entrepreneurs experimented video game nor in the animation industry paid with various ways to commercialize their ideas. attention to or followed the activities of these Rooster Teeth introduced merchandising to help “hobbyists” who operated on a very small scale offset costs, besides DVD sales, an increasingly of product development (cp. Depoorter, 2009), popular revenue window in the motion picture which we consider under the radar. At the same industry in times when box office sales do not time, video game companies were harsh with suffice to break even provided the production users who altered or modified video games re- costs (Eliashberg et al., 2006; Wasko et al., 1993; gardless of the scope of modifications or their Yoon and Malecki, 2009). Table 1 summarizes commercial intention (AC, CB). The same holds the activities undertaken by the firms in our sam- true for film production companies that saw their ple during under-the-radar entry to the anima- media assets re-used and or distributed over the tion industry (phase one) and first opportunity Internet. recognition and commercialization strategies The following quote from Geoff Ramsey (GR) (phase two). illustrates how the founders of Rooster Teeth It is important to note that the user entrepreneurs started out to reach an audience from a spare- were the producers as well as the distributors, time activity and moved on to a full-time activity, exhibitors, and the licensees of their own mer- following the success they achieved in terms of chandise—roles usually separated along the value audience and attention received. chain or at least split into various subdivisions of

36 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

they had either to seek alternative sources of revenue or obtain explicit permission from IP holders.

3.2. Asset holders’ preference for selective sharing of proprietary assets (industry A) Dependency on video games as complementary assets for animation production (see Teece, 1986) com- plicated industry entry by user entrepreneurs for a number of reasons. Game Figure 2. Process model of commercialization by user entrepreneurs across industries engines (movements in 3D space, sound, artificial intelligence, look-and-feel, corporate groups in the motion picture industry. etc.) and artwork (settings, scenery, characters, Thus, user entrepreneurs in our case entered skins, textures) represent significant investments several related industries subsequently based on by companies in the game industry. Copyright potential business opportunities they recognized law and EULAs protect these assets, prohibiting during display of their core product, the Ma- any (commercial) application of the acquired chinima films. video game or the artwork that comes with it. Wendy Selzer (WS), of Brooklyn Law School, These copyright agreements made it difficult if explained the critical legal situation in which user not impossible for user entrepreneurs to sell their entrepreneurs found themselves when entering Machinima films directly (see also Marino, 2004; phase two. Here, their foremost concern is the Hancock and Ingram, 2007). Fred von Lohmann uncertainty regarding a potential legal dispute (FvL), IP lawyer at the Electronic Frontier Foun- with the game companies: dation, clarified the difference in copyright for the “They typically ask that question when they’re trying to two main components of a game, the engine and do ... when they move from the non-commercial into the the assets. While copyright on the assets is closely commercial space because another of the factors in the defined and attributed to the creator, the output fair use inquiry is whether your use is commercial or non- generated by the game engine is far from easy to commercial as well as that somebody is far less likely to classify and thus vulnerable to lawsuits, giving sue you if you’re just doing non-commercial. Or at once user entrepreneurs little leeway to negotiate with you start making money that they see it as something that distributors: they might be getting a cut of. So at that point, you see “I’m sure a game company would say, ‘We own copy- more of them [user entrepreneurs] asking the question.” rights in all of the graphics that comprise the game, so (WS) the character designs, the textures, the landscape arrange- Operating within uncertain legal boundaries, ments,’ you know, all of the graphical elements. I’m sure resulting from the copyright restrictions that they would claim that there’s a copyright there, and they came with the video games, made access to dis- may also argue that they have a copyright in the engine tribution channels anything but easy. TV net- and the output of the engine is, you know, therefore a works exerted pressure on user entrepreneurs by derivative work. I think that is a more far-fetched argu- passing on the legal clarification regarding copy- ment, that argument would quickly lead to the conclusion right infringement (WS). User entrepreneurs in that Microsoft owns everything that is produced with our sample (AK, MK, CB, PM) reported having Microsoft Word, and I think that argument is far fetched experienced conservative reactions to Machinima but, again, it’s not inconceivable. So those are the two from industry incumbents such as TV networks elements that are most likely to arise in a copyright dis- and film studios. This negative attitude meant pute: the graphical objects themselves and then the actual

37 FROM GAME TO PLAY PETER JÄGER • 2010

...the algorithm—the engine that puts it all together.” tions to the standard licensing terms include the (FvL) promotion or advertisement of games through Machinima (AC, BB, FvL, PD). Phil DeBevoise Using games for animation production thus re- (PD) of Machinima.com reported: quires user entrepreneurs to gain legal access to the pertaining assets. Chris Burke of Bong + “The game publishers recognize this is being a terrific free Dern explained his approach to third-party IP as promotional and marketing vehicle for their games and reaching an agreement with the game company they are very much, you know, we have spoken to pretty to avoid any risk, despite incurring high transac- much all of them ...are very excited, and very much want tion costs originating from the negotiations: to engage the gaming community and have them be, you know, very much invested in their games. Like when they “Generally the intellectual property rights are always a are making Machinima, that’s a true act of love and big restriction, you know, even if the game company is passion. They are working many hours not only playing willing to work with you, it’s still going to take you six their game, but they’re also making this.” (PD) months, sometimes a year to work out something which will allow you to do, what you want to do.” (CB) Without the game companies’ consent, users face the dilemma of producing films without permis- Users solved the dilemma of uncertainty in two sion to distribute and run legal risks of infringing ways: they either negotiated a contract with the the copyright or the EULA. To date, no case game company who granted use and thus selec- involving Machinima had been tried in court nor tively shared game assets for clearly defined pur- have EULAs with special Machinima clauses poses, or they remodeled large parts of the game. spread widely. One game publisher that recently While the ILL Clan and Strange Company ap- engaged in encouraging its users to produce Ma- plied specialized tools to “re-skin” avatars and chinima under terms of fair use is Blizzard Enter- avoid using existing artwork, Rooster Teeth struck tainment, the producer of World of Warcraft.17 a deal, which allowed them to use the artwork In general, the right to produce Machinima un- from Halo for their Red vs. Blue franchise. We der fair use of video games (U.S. copyright law), discuss the advantages and disadvantages of ei- as users hope, remains uncertain and a sense of ther strategy later. Burnie Burns of Rooster Teeth frustration about this dilemma is widespread in described their relationship with Microsoft as the Machinima community (voiced by all inter- being grounded in excitement about the possibili- view partners). ties user entrepreneurs discovered, while at the same time showing generosity and interest in a Another approach was to create large amounts of long-term engagement: artwork from scratch. The ILL Clan pioneered this work with their film Hardly Workin’ for “[Microsoft] what we found was we found a group of which “we changed the entire look of the game” people who really like innovative cool stuff and they saw (PM), accomplished by two members during two something that I think they thought was unique and they years of part-time work. Shooting the film, which worked with us and it was really great, it was surprising. involved five people, and integrating the improvi- I was really amazed and I continue to be amazed at how sors’ dialogues, took a further six months. Creat- great they are to us and how much freedom they give us ing own assets by modifying the game gave confi- and how long we are able to work with such a big com- dence of not infringing any IP rights, as Paul pany.” (BB) Marino (PM) and Frank Dellario (FD) described: Selective sharing by game companies takes differ- “Quake allowed you to re-skin your characters, that was ent forms: they license assets under creative one of the things that id Software did ...that was a very commons or open source (OS) licenses (such as innovative approach to customizing the game and [...] we parts of the Unreal Tournament engine released used this feature to do that for our film.” (PM) by Epic Games in 199916 or the tool set for Neverwinter Nights by Bioware); or they close “Using Quake 2 we then made Hardly Workin’ which contracts with Machinima producers that allow Paul Marino directed and that we created all on our own user entrepreneurs to use the games (engines assets. The only thing we used was the engine. The map, and/or artworks) for commercial purposes. The everything we used, we created from scratch, cause we game companies’ incentives to grant these excep- said ‘Let’s get away from other people’s IP.”’ (FD)

16 Ports to Linux are available from the Open UT project: 17 http://www.wow- http://openut.sourceforge.net/info.php. europe.com/de/community/Machinima/letter.html.

38 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

Four firms in the sample (The ILL Clan, Strange, with a story and produce a film is illustrated by Fountainhead, Short Fuze) opted for the costly Burnie Burns (BB) of Rooster Teeth: approach of creating artwork themselves. Bong + “I think, probably since Red vs. Blue started, we’ve never Dern and Rooster Teeth were the exception, in had an idea and then found a game to make the idea that they primarily used game assets after negoti- work. That’s how Red vs. Blue started, we had the idea ating with the game company. Since the contract for Red vs. Blue and we found Halo as the way to do it. details between game companies and Machinima Everything else has been where we’ve had a game with producers remain undisclosed, this study cannot great tools and great Machinima capabilities and then we conclude whether this involved great cost or have found a story to match the theme of the game. We whether the companies placed their assets at the try not to make it about the game but thematically it disposal of selected user entrepreneurs. However, makes sense, if you are going to have elements in the game the evidence shows that support tends to flow already, we want to include them in the story. You can’t from game companies to users, in terms of per- take a game like F.E.A.R. and turn it into a romantic mission to publish, submit to film festivals, con- comedy, well you could but you know...It’s more of a tract work (commissioned Machinima, TV adver- suspension of disbelief in that case. It’s more about just tisements and game commercials), and feature position this. We try to do stories that are believable, support. The Sims 2 features integrated video within the context.” (BB) capturing and Blizzard recently added tutorials for Machinima in World of Warcraft. The latest All firms in the sample used and relied on com- sequel of Halo offers a save film feature, which plementary assets such as game engines, tool sets, allows players to view and record their game play and artwork stemming from video games, which in retrospect from camera angles they did not use were incorporated in the production process. In during actual game play. This also demonstrates some cases, the artwork even stimulated the inno- that the Machinima community, after gaining vation process (BB). enough leverage, had a reverse impact on the Following a different approach, Short Fuze relied game companies (compare von Hippel, 1988, on on complementary assets they produced them- producers who incorporate user innovation). selves: a software product called Moviestorm. Thus, we propose (1): If firms (in other industries) The firm’s business does not rely directly on the show a preference toward selective sharing of proprietary application of game assets but on the user experi- assets, this positively impacts the user entrepreneurs’ access ence associated with them. After seizing the op- to complementary assets for commercialization. portunity while active in the animation industry, Short Fuze created a game-like environment for User entrepreneurs often first use complementary users to shoot films, thus actually supplying com- assets regardless of potential IP infringements plementary assets to users. Matt Kelland (MK) of (phase one) and later recognize an opportunity to Short Fuze depicts legal uncertainty as the incen- commercialize (phase two)—a central finding in tive to create Moviestorm, which substitutes the the user entrepreneurship process. The failure of formerly used video games: asset holders to approve the commercial application of complementary assets may represent a “One of the biggest issues with Machinima is that there road-block for user entrepreneurs, since access to is a big debate about copyright issues, because when peo- complementary assets is critical for commerciali- ple are making videos, using games made by professional zation, even if access is granted ex post (the initial game companies, they say, well, you know, you’re using use). our sounds, you’re using our assets, you’re using our an- imations, our levels and so on. And everybody in the Ma- chinima community is waiting for the first big lawsuit to 3.3. User entrepreneurs’ access to comple- happen and one of the things we decided was just to move mentary assets for commercialization (indus- around this by saying, well, we intend to own the engine, try B) the assets, maps and everything so when somebody makes a movie with Moviestorm, they own it. In just the same User entrepreneurs need access to a range of way that they would own anything they created with complementary assets, including those of game Word or Photoshop.” (MK) companies, as we discussed in the previous section. We now relate these assets to commerciali- Fountainhead quickly ceased to commercialize zation. The process of finding and locating rele- animation despite its access to complementary vant assets in the video game industry to match assets that allowed it to create groundbreaking Machinima early on. Machinima.com stands

39 FROM GAME TO PLAY PETER JÄGER • 2010

alone, due to its special role as a web platform edge exchange as being part of the maturing supporting the community18 in various ways, Machinima community. He describes the inter- ranging from hosting to education. The remain- weaving of technology and knowledge about the ing four firms, however, lend support to our sec- motion picture industry: ond proposition by relying on complementary “Presenting [Machinima] to the community is less im- assets for their commercialization strategy. portant. [...] early on it was about [technology]. That Thus, we propose (2): User entrepreneurs’ access to has changed a little today because the community moved complementary assets positively impacts commercialization. to a field that is less concerned with the technology. A few Proposition two also reflects the two-phase proc- years ago, technology was more central ... to consider ess, since access to complementary assets is pro- things from a technical perspective. There were different vided by user innovators, whom we regard as games one could have possibly worked with. Then pros separate from user entrepreneurs, as well as the and cons were discussed, that was definitively important, game companies’ willingness to share these assets. because back then in the Machinima.com community one For completeness, it should be noted that while didn’t come from the game but from the wish to produce a complementary assets play a crucial role in creat- movie. And because of that, one could be talked into using Machinima, they do not ing another game or at least checking out another game, because it was used as a tool. [...] In this regard, Ma- replace the creation of original artwork like chinima matured a little or at least the community thinks voice-over and audio effects by user entrepre- it matured a bit. And this has more to do with how they neurs. The integration of assets, artwork, and deal with the newcomers. [...] Machinima producers potential post-production with the actual shoot- simply don’t know anything about the 180-degree rule ing requires considerable skill and experience in and cutting and directing in the broadest sense. There, making Machinima all in areas of domain they try to very much catch up and tutorials are written knowledge, which we turn to next. and hints are exchanged and that kind of stuff.” (FK, translated from German by the authors) 3.4. Acquisition of new domain knowledge Knowing each other’s work well, users exchanged experiences and opinions during film festivals Critical new domain knowledge for commerciali- (AMAS, Sundance, Tribeca, Bit-film, and others), zation represents the qualification necessary not game conferences, and via online discussion fo- only to produce but also to distribute animation rums (Machinima.com and others). Some even shot in video games that appeals to a greater wrote Machinima beginners’ books about their audience. It includes education and experience in specific production knowledge (Hancock and cinematography, the creation of a narrative, Ingram, 2007; Marino, 2004) or kept regularly screenwriting, post-production and editing skills, updated blogs (FK, PM, FD). Key contributions as well as knowledge of the film business such as by community members in the form of tools marketing, reaching an audience, creating se- (Uwe Girlich’s Little Movie Processing Center or quels, and sustaining the interest of an audience Friedrich Kirschner’s Movie Sandbox19) altered in film characters and stories—in other words, the way users worked and frequently facilitated the creative skills that are crucial to producing the art of Machinima for new talent. Certain any kind of film, including Machinima. The ap- Machinima pieces reverberated inside the com- plication of complementary assets refers to the munity and inspired new work because of their users’ technical skills at applying and exploiting demonstrable feasibility. The way tools were ap- given features in a video game, such as special- plied was critical for the development process ized tools to reshape the appearance of charac- since it would shape the final product and, ulti- ters, puppeteering, or the capturing of raw video mately, the extent of commercialization. A global material, basically all functions related to execu- community of users sustained discussions about tion. the most effective ways to use tools, both legally Friedrich Kirschner (FK), member of the AMAS, and artistically. Chris Burke (CB) of Bong + Dern explained early approaches to production knowl- exemplifies this stimulation of the community with extraordinary contributions:

18 Because www.Machinima.com did not start commissioning films until the end of August 2007, after we interviewed their CEO Philip DeBevoise, we have not considered Machinima.com for some of our propositions. 19 http://www.moviesandbox.com/

40 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

“Something that really fascinates me is non-narrative Despite the importance of domain knowledge, Machinima, which is, I guess, what grew out of what commercializing Machinima seemed to rely on used to be speed run and all that stuff [recording of game the user entrepreneurs’ ability to combine do- play]. Things like Warthog Jump. [...] I think that’s main knowledge and experience with comple- fascinating. To me that’s something you would never have mentary assets. Users of video games approached in any other medium. And some users they watch it and animation production in new ways (Lowood, they think, oh it’s just some guy messing around, which it 2007; Marino, 2004) regarding their use of tech- is. But there’s, I think there’s a really ... there’s like real nical tools and cinematography, as Matt Kelland theory going on behind that.” (CB) (MK) of Short Fuze commented. The crew at These observations demonstrate that another Short Fuze knew that there were important trade- indispensable aspect of the user entrepreneurship offs when making animation with games or with process in this case involved the community of user experience of gamers. Moviestorm, a game- users (Machinimators), which impacted positively like animation production environment, caters to not only on domain knowledge from the motion users who want to produce Machinima but have picture industry, but also on the combination of little or no film-related education or experience: domain knowledge with complementary assets. “As a game player, my preconception is ‘I don’t have to While the community of Machinimators helped do anything, the computer just works it all out for me.’ to promote the art of producing high-quality And it may not be exactly what I want, but it’s good animation and provided the resources to develop enough and it was easy. Whereas an animator would say, and deepen domain knowledge, it also promoted well you have to be able to decide where they walk and the integration and application of complemen- how they walk and get it all absolutely right, which we tary assets. say, it actually doesn’t matter.” (MK) Burnie Burns (BB) and Geoff Ramsey (GR) of Our cases demonstrate that a community of users Rooster Teeth produced the first series of Red vs. positively impacts on the accumulation of rele- Blue in their homes while voice actors called in vant domain knowledge from different industries. over the phone to contribute their part of the We can observe an important distinction between scripts. Later, they recruited friends, some of user innovators and user entrepreneurs. User whom had worked in Hollywood studios, to join innovators, within a community or by themselves, the team. Domain knowledge proved crucial combine complementary assets (video games) when it came to commercialization. Not only did with cinematography skill (e.g. Randall Glass with Matt Hullum bring valuable film experience to his film Warthog Jump, published in 2002, the team, but the development of the series’ achieving effects once thought impossible). User characters and epic story elements over 100 epi- entrepreneurs, however, take their animation sodes, commitment to the web as the preferred products one step further and commercialize channel of distribution, as well as contracts that them in the animation industry. Depending on secured the rights to sell DVDs and merchandise, the user entrepreneurs’ origin, the respective all represent thorough knowledge about the mo- other domain knowledge has to be acquired. tion picture industry that few video game users To commercialize Machinima, user entrepre- possess. neurs need to combine domain knowledge with The case of Fountainhead Entertainment dem- their skills at applying complementary assets. onstrates the contrary situation. Its founder, Anna Consider again the example of Rooster Teeth: Kang (AK), one of the co-founders of AMAS, the tale of their successful series starts with the created groundbreaking Machinima work. De- discovery of a bug in Microsoft’s Halo game that spite critical acclaim, their efforts did not trans- allowed them to make their avatars look straight late to sustained commercial activity in the ani- ahead while pointing their guns down. This non- mation industry, possibly because there was lim- feature enabled the dialogue scenes in Red vs. ited industry experience within Fountainhead— Blue. The deep experience with a video game, their critically acclaimed music video In the Wait- the cinematographic skill to exploit this bug artis- ing Line had been produced with an outside di- tically to create entertaining products, and the rector who was interested in trying new technolo- business knowledge of how to market the product gies. The other Machinima producing firms in need to come together for commercialization. All the sample had at least one core member or of our cases show that access to both kinds of founder with an education in film or extensive knowledge—domain knowledge as well as experi- industry experience in film or animation or both.

41 FROM GAME TO PLAY PETER JÄGER • 2010

ence and skills—is necessary for commercializa- ence as users of Machinima tools, the insights tion. gained from a community of peers, and the We propose (3a): The community of users positively knowledge to be gained in the domain of the impacts on user entrepreneurs’ skills at applying comple- animation industry, all contribute to the possibil- mentary assets. ity of entering phase two, commercialization. The ability of all firms to generate revenue could In summary, commercialization was impossible, be traced to talent from the motion picture indus- in our sample, without the combined knowledge try, whether by temporary arrangements, hiring, of the domain of filmmaking and of the applica- or via the founders. We thus propose (3b): Access to tion of games as complementary assets to film the talent pool of the target industry positively impacts on production. We propose (4): User entrepreneurs com- user entrepreneurs’ acquisition of domain knowledge. bine domain knowledge with the skills to apply complementary assets to commercialize their products. Note that this finding is compelling in the sense that all but one user entrepreneur followed this Table 2 gives a summary of the propositions and course and phase one (under the radar) appeared how they relate to the respective firms, with a as a prerequisite for phase two. Understanding short description of how each proposition is sup- how to produce Machinima film and the experi- ported throughout the cases.

42 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

4. DISCUSSION

Our findings relate to the emerging literature on Proposition 2. The innovation literature has so user entrepreneurship which shares with an effec- far generated only limited insights into the com- tuation view on entrepreneurship theory the per- mercial use of selectively shared assets. We show spective that opportunities are created given a set that selective sharing may apply to different types of means (Read et al., 2009), here, the user expe- of asset (game engine, artwork, toolkits, etc.), rience. In this section we discuss where and how which in turn represent different trade-offs for the our propositions resonate with existing theory or user entrepreneur in terms of gaining access to, depart from it. or substituting for, the asset. We observe that Our finding that industry entry occurs in two while user innovators frequently make unauthor- phases extends existing work on user entrepre- ized use of copyright work (as described by Lee, neurship (Baldwin et al., 2006; Shah and Tripsas, 2008), user entrepreneurs display great sensitivity 2007) in that it separates the user innovator from when working with others’ copyright protected the user entrepreneur while describing a coherent work. They either secure owners’ explicit permis- process of user entrepreneurship. A user innova- sion to create and distribute Machinima using tor may enable phase one by extending the use of video game artwork, or they completely re-create a product or technology (von Hippel, 2005; artwork to avoid conflict over intellectual prop- Faulkner and Runde, 2009), in our case enabling erty rights. Commercializing with others’ assets in a video game to be used as a tool for animation a new industry is a facet of the user entrepre- production. Phase two may evolve without major neurship process not yet described by the litera- user innovations if user entrepreneurs draw on an ture. existing innovation for their own use and for later Proposition 3. An important finding in the user commercialization. Nevertheless, a deep familiar- innovation literature is that users tend to organize ity with the use of the innovation was a prerequi- their innovation projects in communities (von site for commercialization for all entrepreneurs in Hippel and von Krogh, 2003; Shah, 2006; von our sample. Hippel, 2007). Members of these communities bring their individual domain knowledge to bear on technical problems, share solutions, promote 4.1. Propositions in the light of existing theory their work, and develop and improve on technol- Proposition 1. Some authors have argued that ogy (Franke and Shah, 2003; von Krogh et al., firms frequently solicit the use of their knowledge 2003; Jeppesen and Molin, 2003). As users who assets for a licensing fee (Chesbrough, 2003; become entrepreneurs and enter a new industry, Arora et al., 2001). Uncompensated tolerance of the firms in our sample provide a unique oppor- asset exploitation has been documented for cases tunity to observe where knowledge is sourced and of informal know-how trading (von Hippel, to distinguish between knowledge of production 1987), and in areas where participating firms do methods, application or development of com- not compete directly and find it beneficial to sup- plementary assets, and domain knowledge. Little port each other (Henkel, 2006; Dahlander, 2007). is known about this distinction in the user innova- User entrepreneurs who deploy proprietary assets tion literature, which is vague about when knowl- depend on the owner’s tolerance of the applica- edge emerges from a user community and when tion of these assets for commercialization. Our it needs to be acquired through hiring, or as part cases extend previous research on selective shar- of the founding team. ing and tolerance toward application by demon- Within user communities, members share and strating the potential relevance of assets for user exchange experience with peers from other entrepreneurs’ entry into commercial market- knowledge domains (Lee and Cole, 2003; Spaeth places. They also document a nascent group of et al., 2008). Kogut and Zander (1992) suggested firms that contribute to developing a new genre studying how firms combine knowledge from in the industry they enter, which underscores the internal as well as external sources for innovation novelty of their products and the associated risks (see also Schumpeter, 1934). Successful innova- of IP infringement. tors need the ability to identify external knowledge as an important input to innovation and

43 FROM GAME TO PLAY PETER JÄGER • 2010

commercialization, and must have the capability industry) to include other industries and the crea- of combining new and existing knowledge (Alva- tion of new market fragments. rez and Barney, 2005; Brockman and Morgan, 2003; Chirico and Salvato, 2008). Users are considered a source of domain knowledge from idea- 4.2. A summary of stakeholders and activities tion to complete product development (von Hip- In exploring the case of Machinima, we studied a pel, 1988, 2007; Baldwin et al., 2006; Ogawa and new genre and a growing community of film and Piller, 2006; Füller et al., 2007); however, produc- gaming enthusiasts. Table 3 provides an overview tion methods are frequently considered proprie- and summary of the stakeholders involved and tary (Henkel, 2006: 962). We identify the types their activities before and during the phases of and sources of knowledge combined by user en- commercialization described by our model. User trepreneurs and propose ways in which a user innovation began before the users became entre- community provides insights into the production preneurs and the stories of their businesses un- technology for Machinima. By developing and folded. Our data cover the beginnings of user evolving the use of video games for animation innovation in this field and the birth of the genre, production, the community participates in entry since many of the individuals who went on to into a new industry. become entrepreneurs are today considered Ma- The firms in our sample exchange insights about chinima pioneers (Paul Marino, Hugh Hancock, video games in user communities (Cohendet and Burnie Burns, Anna Kang, and others). They Simon, 2007) and hire market insiders to acquire reported their first encounters with the technolo- domain knowledge. At the same time, they de- gies of recording and sharing game play, the velop proprietary production knowledge, such as creativity these spurred, and the first teams they experience in narrative development and plotting, formed to manage early productions. or post-production techniques. Revealing of In our account of the propositions, some knowledge by user entrepreneurs depends on the stakeholders’ positions were omitted for the sake industry context, in that production tools—the of brevity. However, more detail about the roles application of game engines for Machinima pro- played by the video game companies and the duction—are more liberally shared within the audience may facilitate the testing of our model user community than production skills, such as in other contexts. We convey a comprehensive cinematography and story development, key ele- account of the Machinima case, reaching back as ments for revenue generation in the animation far as possible in its history, to open the possibility industry. for others to discover and identify nascent mar- Proposition 4. When building new ventures in kets and industries by recognizing user activity an industry, entrepreneurs need domain knowl- and behavior in terms of knowledge acquisition edge relevant to commercialization (Michael et and re-interpretation of assets. al., 2002). When entering new industries, domain The columns in Table 3 show critical activities knowledge about target markets plays a key role within the two relevant industries, video games in commercializing innovations. The case of Ma- and motion picture. The rows follow the chinima shows that users’ experience can trans- stakeholders over time, going from top to bottom, late into an ability to apply tools that serve as from an enabling phase to user innovation and complementary assets in a new industry, and that then to the two commercialization phases covered a combination with domain knowledge enables by our propositions. The relevant stakeholders, them to commercialize their products. This find- apart from user entrepreneurs, include gamers ing departs from the literature on user innovation who are forming a community of Machinimators and user entrepreneurship, which assumes users’ (the lead user community), incumbent firms in stronghold to be their domain of experience, the video games industry, and consumers of vid- their access to knowledge generated by their eos. Shaded cells contain user entrepreneurs’ key practice as users, as well as their network (Luthje as described above, and may serve as reference et al., 2005; von Hippel, 2007). Our sample sug- points to compare with other stakeholders’ activi- gests that users’ realm of entrepreneurial activity ties. extends beyond the market where initial user innovation could be observed (the video game

44 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

Table 3. Stakeholder analysis considering the sequence of activities and industries

45 FROM GAME TO PLAY PETER JÄGER • 2010

5. IMPLICATIONS FOR RESEARCH AND MANAGEMENT

Our findings contribute to the literatures on user well as more Machinima. The point here is not entrepreneurship and (remotely) strategy, in that that user entrepreneurs’ creativity is unlimited we motivate future research on entrants keeping a but that their accumulated experience in the pro- low profile to overcome entry barriers. Four con- duction of Machinima, and the gaming culture tributions stand out. that the team shares with its audience, are sources First, a central finding of our study relates to user of new ideas that can give rise to new opportuni- entrepreneurship. Industry entry by user entre- ties and ultimately to commercialization. User preneurs could be observed to occur in two entrepreneurs search for new applications for phases (see Fig. 1 and Table 1). This demon- assets and skills (their givens) rather than focusing strated that entrants could outmaneuver entry on a predefined target by discarding alternatives, barriers by operating under the radar of incum- a process that Sarasvathy (2001) terms “effectua- bent firms before commercializing. Although they tion.” This process could also be observed in were already distributing film, users did not ap- other firms in our sample. pear as commercial players. Once they had This finding confirms the notion that user inno- gained a foothold through their audience, users vation rests on the advantage provided by sticky began to commercialize. Their two-phase entry knowledge (von Hippel, 1994) and users’ “local” into the animation industry avoided apparent perception of their own and their community’s conflict over IP rights by first relying on informal needs. However, it also shows that user entrepre- copyright practices (see Lee, 2008), and also neurs effectively leverage community knowledge avoided large investments in established distribu- for entry into new industries. Studying the user tion channels by using the Internet. entrepreneur’s role in the context of fellow users Second, our findings contribute to research in and consumers echoes the recent call by Rindova entrepreneurship and effectuation more specifi- et al. (2009) to study entrepreneurial activity cally. Rooster Teeth produced Machinima in more deeply to “understand the relationships Halo 1–3, Sims 2, F.E.A.R., and Shadowrun. between change intent and the nature of relation- The team’ skill in attracting a community of over ships with other social actors” (2009: 480). We half a million subscribing viewers and fans re- contribute to the body of research on effectuation sulted in several sources of revenue through their by delineating the characteristics of a process franchises: DVD sales, sponsorship, merchandis- where entrepreneurs enter a new industry in two ing with various items, as well as the production phases. Future research may help to predict tar- of a comic series that features the Rooster Teeth get industries and entrepreneurial activity based crew as main characters. The comic was available on user innovation and the principal of givens in for free through their website and sold in print. effectual thought (compare Read et al., 2009). The ventures grew around Machinima but devel- Third, our theory contributes to a differentiated oped Rooster Teeth into a web-based entertain- understanding of the user’s role in technology ment group with artistic products, such as the diffusion. Users, the Machinimators in our case, comic and the communication contents. Rooster have radically changed the use of video games by Teeth’s team learned by experimentally setting up using them as stage and input for animation pro- a shop around its brand and around specific duction rather than game play. Faulkner and characters from its Machinima series. Rooster Runde (2009) describe the diffusion of these user- Teeth’s strategy was driven by its creative use of triggered changes in products. Video games con- complementary assets. Using another video game tain both technology (the physics engine that led to an all-new series. According to our inter- creates a basis for the virtual world) and art (the viewees, using the team’s gaming experience to artwork that makes the virtual world visible and relate to other gamers in the audience creates allows for the experience of sensations inside the customer loyalty and sparks creativity, in terms of virtual world). The two can be separated, as we customer-generated content on the community demonstrated, and their diffusion follows slightly site as well as within the Rooster Teeth team (BB, different paths, given their creators’ choices to GR). A traditional sitcom around video gaming protect them from exploitation in different con- could fit with the Rooster Teeth’s strategy just as texts. The new context is animation; since the

46 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

advent of Machinima in the late 1990s, video into the motion picture industry themselves, or games have been used for animation production the audience could turn away from Machinima. in various ways. Hence, users diffuse technologies However, given the Machinima firms’ diverse across industry boundaries depending on their sources of revenue, their dependence on game assessment of the components of the technology, companies is limited. Future research needs to on their own costs of re-creating it, and on the follow up on the phenomenon. outcome of the negotiations with the owners of the IP rights to the technology, which may in turn be a question of timing. 5.1. To be continued... Fourth, our study raises a few important questions Our analysis covered a number of topics beyond about atomistic versus collective user entrepre- and in addition to the process model, which we neurship and the logic of selective revealing cut in favor of length and accessibility (see also (Henkel, 2006). While the user community of Table 3). We thus propose an agenda for future Machinimators was instrumental in diffusing research. Successful industry entry implies a gain knowledge about the application of video games in market share by the new entrant at the expense to animation production, process and business of incumbent firms. To date, Machinima is insig- knowledge (sequel and storyline development, nificant compared to the revenues generated by distribution channels) were revealed to a far lesser big players in the motion picture industry. Meas- extent. Gamer communities that became Ma- uring the market share to justify successful entry chinima communities preceded commercial ef- thus proves impractical. Machinima represents an forts, a finding that corresponds to other areas of emerging genre in film today and user entrepre- user entrepreneurship (Baldwin et al., 2006, Shah neurs, entering under the radar, created a new and Tripsas, 2007). What is new, in our case, is niche in the animation industry, fragmenting that the choice of industry lies in the nature of existing markets. The new niche is characterized the users’ activity. We were not observing a few by low barriers to entry, which in turn enable renegade firms that decided to venture into an further talent to enter the Machinima market. unknown industry; ours were entrepreneurial With the continuously advancing graphical capa- users who took innovation one step further and bilities of video games, the full potential of this commercialized in the industry most akin to the market might yet to be unleashed. How this in- work they had done: animation. The Machinima dustry will evolve over time, and whether it will community does not discuss game play or exten- capture market share or disentangle from the sions of games for gaming’s sake but develops conventional media markets, remains to be seen. ever more sophisticated knowledge that combines Machinima represents a fertile ground for study- the use of games and cinematography, as the ing the strategic impact of industry entry and quote by Friedrich Kirschner (above) illustrates. market fragmentation on incumbents as well as Some limitations apply to this study. First, while new entrants. the Machinima phenomenon is growing rapidly, The cases demonstrate that learning can occur in generalization of the propositions to areas other phase one, prior to full entry. Hence, important than games must be tested in future research. industry entry barriers do not deter user entry Second, two effects were almost impossible to and may need reconsideration in terms of their disentangle and could jeopardize the future ap- effectiveness for this type of activity (Porter, 1980; plicability of the theory to contexts other than Lieberman, 1987). If de novo entrants can inno- virtual environments: the diffusion of broadband vate without entering an industry commercially access and advances in distribution technology (users learn by experimenting with products and for media content over the Internet. The popu- processes and from peers), learning curve effects larity of Machinima, and thus the commercial could in fact moderate other entry barriers. Op- viability of the firms studied, might be connected portunity costs of users are rather low, so could to the general trend of viewing film over the In- offset entry costs, while they rise with commer- ternet. Finally, Machinima represents a very re- cialization when the user entrepreneur engages in cent phenomenon that provides scarce data on firm foundation. This eventually leads to the the survival of our sample firms (none of which question of which entry barriers remain intact can demonstrate a track record of more than ten during phase one, industry entry, and which dur- years). Game companies could eventually cease ing phase two, entry to the commercial market- selective sharing of engines and tools and venture place—a question we defer to future research.

47 FROM GAME TO PLAY PETER JÄGER • 2010

Interaction with their customers helped user en- a similar license in the same year.21 A recent trepreneurs to improve their products and correct analysis by Lee (2008) documents the rise of in- flaws. Posting Machinima on sites with social formal copyright practices for user-generated software features (such as commenting and rating) content. This “warming to unauthorized use” allowed user entrepreneurs to read viewers’ (Lee, 2008) includes commercial applications and comments within minutes.20 Rooster Teeth en- may, given appropriate and balanced policy (Ro- couraged short feedback cycles by establishing a quilly, 2010), influence the behavior of both users dual release structure granting sponsors early and entrepreneurs. However, this is an emerging access to new episodes. Sponsors’ feedback thus legal practice and the firms in our sample enabled the production team to correct flaws avoided building their ventures on the unauthor- prior to the public rollout. Consumer motivation ized use of copyright work. to engage in such activities, as well as the impact Innovation policy needs to take into account the of social software features to support these proc- way protected assets can be used across indus- esses, need close examination in order to imple- tries. Crucially, the transaction costs to create ment value-adding IT solutions. The nature of legal agreements should be lowered so that rights this interaction contributes to the continuous holders have incentives to enter negotiations with blurring of boundaries between the traditionally prospective entrepreneurs. Today, these incentives separated media of film, Internet, and video frequently point the other way and lead to whole- gaming. sale decline. Recent Internet technology, such as Firms in one industry increasingly witness how video compression and distribution, allows new freely or restrictively shared tools and assets can talent to create entertainment products on a be used in another industry. This can be both broad scale (Eliashberg et al., 2006). Encouraging beneficial and detrimental to the core business innovation in this domain calls for more flexible and thus needs close monitoring. In general, our and informal copyright agreements, since the sample suggests that positive effects dominate for current uncertainty may deter user innovators game industry incumbents. Ill-considered han- and user entrepreneurs. dling of IP, whether in the form of a too lenient or too strict position on exploitation of complementary assets in users’ target industries, can negatively affect the core business. Game companies might want to take a clear stance and en- courage or discourage users from certain practices from the point of product launch. While game companies did this rigorously for their home industry, they largely ignored setting up directives for user entrepreneurs’ target industries. These policies can define both the kind of exploitation and possible target industries. Laws differ about the fair use of IP across coun- tries and create uncertainties for users, as our study confirmed. Users can be frustrated with EULAs and select games based on their availability and flexibility of use. EULAs do not necessarily need to be rewritten, nor will individual contracts and agreements resolve the problem in the long run. Amendments can give certain rights back to users, as happened on Blizzard’s World of Warcraft site where the company made a 180- degree policy change in 2007 and started to share assets for certain uses selectively. Microsoft issued 21 www.worldofwarcraft.com/community/Machinima/letter 20 Season five of Red vs. Blue averaged 1,163 comments per .html. An article in Wired discussed the license change by episode, the first 50 posts usually arriving within 30 min Microsoft: www.wired.com/culture/art/ of the episode’s release. news/2007/09/Machinimalicenses.

48 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

ACKNOWLEDGEMENTS

The authors would like to thank our interview support from the Swiss National Science Founda- partners and the community of Machinimators tion (grant 100014 125513/1). Thanks also to for their time and valuable input, which made Bob Young whose book (Young and Rohm, 1999) this research possible. Corinne and Pascal Gyger, inspired our title. Lee Fleming and two anony- Mario Bischof and Alex Beninca’ provided excel- mous reviewers offered most helpful guidance lent research assistance. This project received throughout the publication process.

49 FROM GAME TO PLAY PETER JÄGER • 2010

APPENDIX

Table A1. Coding scheme

Table A2. Interview partners

50 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

REFERENCES

Alvarez, S.A., Barney, J.B., 2005. How do entrepreneurs Eisenhardt, K.M., Graebner, M.E., 2007. Theory building organize firms under conditions of uncertainty? Journal of from cases: opportunities and challenges. Academy of Man- Management 31 (5), 776–793. agement Journal 50 (1), 25–32. Alvarez, S.A., Barney, J.B., 2007. Discovery and creation: Eliashberg, J., Elberse, A., Leenders, M.A.A.M., 2006. The alternative theories of entrepreneurial action. Strategic motion picture industry: critical issues in practice, current Entrepreneurship Journal 1 (1–2), 11–26. research, and new research directions. Marketing Science 25 (6), 638–661. Aoyama, Y., Izushi, H., 2003. Hardware gimmick or cultural innovation? Technological, cultural, and social foundations Farjoun, M., 1994. Beyond industry boundaries: human of the Japanese video game industry. Research Policy 32, expertise, diversification and resource-related industry 423–444. groups. Organization Science 5 (2), 185–199. Arora, A., Fosfuri, A., Gambardella, A., 2001. Markets for Faulkner, P., Runde, J., 2009. On the identity of technologi- Technology: Economics of innovation and corporate strat- cal objects and user innovations in function. Academy of egy. MIT Press, Cambridge. Management Review 34 (3), 442–462. Baldwin, C., Hienerth, C., von Hippel, E., 2006. How user Franke, N., Shah, S., 2003. How communities support inno- innovations become commercial products: a theoretical vative activities: an exploration of assistance and sharing investigation and case study. Research Policy 35 (9), 1291– among end-users. Research Policy 32 (1), 157–178. 1313. Füller, J., Jawecki, G., Mühlbacher, H., 2007. Innovation Britton, J.B.H., Tremblay, D.-G., Smith, R., 2009. Contrasts creation by online basketball communities. Journal of Busi- in clustering: the example of Canadian New Media. Euro- ness Research 60 (1), 60–71. pean Planning Studies 17 (2), 211–234. Geroski, P.A., 2000. Models of technology diffusion. Re- Brockman, B.K., Morgan, R.M., 2003. The role of existing search Policy 29 (4–5), 603–625. knowledge in new product innovativeness and performance. Decision Sciences 34 (2), 385–419. Glaser, B., Strauss, A.L., 1967. The Discovery of Grounded Theory. Aldine, Chicago, IL. Brown, S.L., Eisenhardt, K.M., 1997. The art of continuous change: linking complexity theory and time-paced evolution Hancock, H., Ingram, J., 2007. Machinima for Dummies. in relentlessly shifting organizations. Administrative Science Wiley, NY. Quarterly 42, 1–34. Harhoff, D., Henkel, J., von Hippel, E., 2003. Profiting from Calantone, R.J., Yeniyurt, S., Townsend, J.D., Schmidt, J.B., voluntary information spillovers: how users benefit by freely 2010. The effects of competition in short product life-cycle revealing their innovations. Research Policy 32 (10), 1753– markets: the case of motion pictures. Journal of Product 1769. Innovation Management 27, 349–361. Henkel, J., 2006. Selective revealing in open innovation Chirico, F., Salvato, C., 2008. Knowledge integration and processes: the case of embedded Linux. Research Policy 35 dynamic organizational adaptation in family firms. Family (7), 953–969. Business Review 21 (2), 169–181. Izushi, H., Aoyama, Y., 2006. Industry evolution and cross- Chesbrough, H.W., 2003. The era of open innovation. sectoral skill transfers: a comparative analysis of the video Sloan Management Review 44 (3), 35–41. game industry in Japan, the United States, and the United Kingdom. Environment and Planning A 38, 1843–1861. Cohen, B.P., 1980. Developing Sociological Knowledge: Theory and Method. Prentice Hall, Englewood Cliffs, NJ. Jeppesen, L.B., Molin, M.J., 2003. Consumers as co- developers: learning and innovation outside the firm. Tech- Cohendet, P., Simon, L., 2007. Playing across the play- nology Analysis & Strategic Management 15 (3), 363–383. ground: paradoxes of knowledge creation in the video-game firm. Journal of Organizational Behaviour 28 (5), 587–605. Kogut, B., Zander, U., 1992. Knowledge of the firm, com- binative capabilities, and the replication of technology. Dahlander, L., 2007. Penguin in a new suit: a tale of how de Organization Science 3 (3), 383–397. novo entrants emerged to harness free and open source software communities. Industrial and Corporate Change 16 Lawrence, T.B., Hardy, C., Phillips, N., 2002. Institutional (5), 913–943. effects of interorganizational collaboration: the emergence of proto-institutions. Academy of Management Journal 45 Depoorter, B., 2009. Technology and uncertainty: the shap- (1), 281–290. ing effect on copyright law. University of Pennsylvania Law Review 157, 1831–1868. Lee, G.K., Cole, R.E., 2003. From a firm-based to a community-based model of knowledge creation: the case of the Eisenhardt, K.M., 1989. Building theories from case study Linux kernel development. Organization Science 14 (6), research. Academy of Management Review 14 (4), 532–550. 633–649. Lee, E., 2008. Warming up to user-generated content. Uni- versity of Illinois Law Review 5, 1458–1548.

51 FROM GAME TO PLAY PETER JÄGER • 2010

Leyshon, A., 2001. Time-space (and digital) compression: Rindova, V., Barry, D., Ketchen, D.J., 2009. Introduction to software formats, musical networks, and the reorganization Special Topic Forum: entrepreneuring as emancipation. of the music industry. Environment and Planning A 33, 49– Academy of Management Review 34 (3), 477–491. 77. Rogers, E.M., 1962. Diffusion of Innovations. Free Press, Lieberman, M.B., 1987. The learning curve, diffusion, and Glencoe. competitive strategy. Strategic Management Journal 8 (5), 441–452. Roquilly, C., 2010. The control over virtual worlds by game companies: issues and recommendations. MIS Quarterly Liebeskind, J.P., 1996. Knowledge, strategy, and the theory (Forthcoming). of the firm. Strategic Management Journal 17 (Winter Special Issue), 93–107. Sarasvathy, S.D., 2001. Causation and effectuation: toward a theoretical shift from economic inevitability to entrepreneu- Lowood, 2007. High-performance play: the making of rial contingency. The Academy of Management Review 26 Machinima. In: Clarke, A., Mitchell, G. (Eds.), Videogames (2), 243–263. and Art. Intellect Books, Bristol. Sarasvathy, S., 2004. Making it happen: beyond theories of Luthje, C., Herstatt, C., von Hippel, E., 2005. User- the firm to theories of firm design. Entrepreneurship The- innovators and “local” information: the case of mountain ory and Practice 28 (6), 519–531. biking. Research Policy 34, 951–965. Schumpeter, J.A., 1934. The Theory of Economic Devel- March, J.G., Sproull, L.S., Tamuz, M., 1991. Learning from opment. Harvard University Press, Cambridge. samples of one of fewer. Organization Science 2 (1), 1–13. Scott, A.J., 2004. Hollywood and the world: the geography Marino, P., 2004. 3D Game-Based Filmmaking: The Art of of motion-picture distribution and marketing. Review of Machinima. Paraglyph Press, Scottsdale, AZ. International Political Economy 11 (1), 33–61. Mezias, J.M., Mezias, S.J., 2000. Resource Partitioning, the Shah, S., 2006. Motivation, governance, and the viability of Foundation of Specialist Firms and Innovation: the Ameri- hybrid forms in open source software development. Man- can Feature Film Industry, 1912–1929. Organization Sci- agement Science 52, 1000–1014. ence 11 (3), 306–322. Shah, S., Tripsas, M., 2007. The accidental entrepreneur: Michael, S., Storey, D., Thomas, H., 2002. Discovery and the emergent and collective process of user entrepreneur- coordination in strategic management and entrepreneurship. ship. Strategic Entrepreneurship Journal 1, 123–140. In: Hitt, M., Ireland, R.D., Camp, S.M., Sexton, D.L. (Eds.), Strategic Entrepreneurship. Blackwell, Oxford, pp. 45–65. Shane, S., Venkataraman, 2000. The promise of entrepreneurship as a field of research. Academy of Management Morris, D., Kelland, M., Lloyd, D., 2005. Machinima. Review 25 (1), 217–226. Course Technology PTR. Siggelkow, N., 2007. Persuasion with case studies. Academy Ogawa, S., Piller, F.T., 2006. Reducing the risks of new of Management Journal 50 (1), 20–24. product development. MIT Sloan Management Review 47 (2), 65–71. Spaeth, S., Haefliger, S., von Krogh, G., Renzl, B., 2008. Communal resources in open source software development. O’Rourke, M.A., 2000. Toward a doctrine of fair use in Information Research 13 (1). patent law. Columbia Law Review, 100 (5): 1177–1250. Strandburg, K.J., 2008. Users as innovators: implications for Papies, D., Clemet, M., 2008. Adoption of new movie distri- patent doctrine. University of Colorado Law Review, 79 bution services on the Internet. Journal of Media Economics (46): 467–544. 21, 131–157. Strauss, A.L., Corbin, J., 1990. Grounded theory research: Peretti, N., Negro, G., 2007. Mixing genres and matching procedures, canons and evaluative criteria. Zeitschrift für people: a study in innovation and team composition in Hol- Soziologie 19 (6), 418 ff. lywood. Journal of Organizational Behavior 28 (5), 563– 586. Teece, D.J., 1986. Profiting from technological innovation: implications for integration, collaboration, licensing and Pervez, N.G., Tarnovskaya, V., Elg, U., 2008. Market driving public policy. Research Policy 15, 285–305. multinationals and their global sourcing network. Interna- tional Marketing Review 25 (5), 504–519. Vaughan, D., 1990. Autonomy, interdependence, and social control: NASA and the space shuttle challenger. Administra- Pettigrew, A. 1988. Longitudinal field research on change. tive Science Quarterly 35 (2), 225–257. Theory and practice. Paper presented at the National Sci- ence Foundation Conference on Longitudinal Research Venkataraman, S., 1997. The distinctive domain of entre- Methods in Organizations, Austin, Texas. preneurship research. In: Katz, J. (Ed.), Advances in Entre- preneurship, Firm Emergence, and Growth, vol. Porter, M.E., 1980. Competitive Strategy. Techniques for Analyzing Industries and Competitors. Free Press, New 3. JAI Press, Greenwich, CT, pp. 119–138. von Hippel, E., York. 1987. Cooperation between rivals: informal know-how trading. Read, S., Song, M., Smit, W., 2009. A meta-analytic review of effectuation and venture performance. Journal of Busi- Research Policy 16, 291–302. ness Venturing 24 (6), 573–587. von Hippel, E., 1988. The Sources of Innovation. Oxford University Press, Oxford.

52 UNDER THE RADAR • INDUSTRY ENTRY BY USER ENTREPRENEURS

von Hippel, E., 1994. Sticky information and the locus of Wasko, J., Phillips, M., Purdie, C., 1993. Hollywood meets problem solving: implications for innovation. Management Madison Avenue: the commercialization of US films. Media, Science 40 (4), 429–439. Culture & Society 15, 271–293. von Hippel, E., von Krogh, G., 2003. Open source software Yin, R.K., 2003. Case study research: design and methods. and the “Private-Collective” innovation model: issues for Sage, Thousand Oaks, CA. organization science. Organization Science 14 (2), 209–223. Yoon, H., Malecki, E.J., 2009. Cartoon planet: worlds of von Hippel, E., 2005. Democratizing Innovation. MIT production and global production networks in the animation Press, Cambridge, MA. industry. Industrial and Corporate Change 19 (1), 239–271. von Hippel, E., 2007. Horizontal innovation networks—by Young, R., Rohm, W.G., 1999. Under the Radar: How Red and for users. Industrial and Corporate Change 16 (2), 293– Hat Changed the Software Business–And Took Microsoft by 315. Surprise: How the Open Source Sneak Attack Is Transform- ing the Technology War First Edition. Coriolis Group, von Krogh, G., Spaeth, S., Lakhani, K.R., 2003. Commu- nity, joining, and specialization in open source software U.S. innovation: a case study. Research Policy 32 (7), 1217–1241.

53 FROM GAME TO PLAY PETER JÄGER • 2010

MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES

THE CASE OF ROOSTER TEETH PRODUCTIONS

Stefan Haefliger Philip Reichen Peter M. Jäger Georg von Krogh

Chair of Strategic Management and Innovation Department of Management, Technology, and Economics (MTEC) ETH Zürich, Switzerland

Published in Lecture Notes in Computer Science vol. 5621 2009 page 197-206

Virtual communities that make use of social network site features blend known applications of virtual communities. These communities can be simultaneously social and commercial, organization sponsored and heavily relying on member interaction. We explore modding behavior that allows members to evaluate other members’ contributions both with numerical value and qualitative rating. We show that approximately half of all members received mods on their comments, that the majority of mods given were positive, and that the amount of mods received for a comment was related to the position of the comment in the community website’s thread. Contributing to the emerging literature of social network sites and virtual communities, we discuss implications for theory, future research and management.

55 FROM GAME TO PLAY PETER JÄGER • 2010

1. INTRODUCTION

In a recent MIT Sloan Management Review by Tim O'Reilly23, social network sites (SNS) article, Bernoff and Li (2008) suggested “People emphasize member profiles and direct interaction are connecting with one another in increasing and links between members, provide content numbers, thanks to blogs, social networking sites ratings, and enable rating behavior (de Valck et like MySpace and countless communities across al., 2007; Kim et al., 2008; Pascu et al., 2007). the Web. Some companies are learning to turn “Modding” (derived from “moderation”) refers to this growing groundswell to their advantage.” a type of trust rating that “allows members […] With close to one billion22 people connected to to evaluate other users’ reviews with numerical the Internet, firms not only face unprecedented ratings” (Kim et al., 2008: 532). Modding is a opportunities but also considerable threats in direct feedback mechanism between community such a digital economy. Numerous firms have set members. up “virtual communities,” a term coined by Both streams of research on virtual communities Rheingold (1993). These communities are, mostly and social network sites belong to the field of but not exclusively, online spaces in which cus- computer-mediated communication. The combi- tomers and non-customers can interact with the nation of features within one online environment firm and each other. Porter (2004) defines virtual triggered new forms of behavior that warrant communities as an “aggregation of individuals or analysis. If virtual communities make use of SNS business partners who interact around a shared features the combination results in a new type of interest, where the interaction is at least partially virtual community that cannot easily be under- supported and/or mediated by technology and stood by the frameworks used to classify virtual guided by some protocols or norms.” (see also communities (Porter, 2004). Porter and Donthu, 2008). A virtual community that makes use of social Virtual communities can have a positive impact software features may be organization sponsored, on firm performance. According to one study, yet dominated by direct interaction among com- revenues have increased more than 50% for some munity members, hence, social and at the same firms (Algesheimer and Dholakia, 2006) that have time commercial. SNS features offer communica- managed these communities well. In addition tion structures that make member-to-member members of virtual communities remain twice as communication easier and more frequent. Mod- loyal to and buy almost twice as often from the erated communication makes members become sponsoring firm. Armstrong and Hagel (1996) more socially embedded in the virtual community found that “companies that create strong online (de Valck et al., 2007; Algesheimer et al., 2005). communities will command customer loyalty to a The availability of new communication struc- degree hitherto undreamed of and, consequently, tures that allow direct feedback on contributions will generate strong economic returns”. In addi- calls for research exploring rating behavior. Spe- tion, virtual communities can shift bargaining cifically, modding of member comments by other power from suppliers to customers (Kozinets, members extends the communication options 1999); spread positive word-of-mouth (Dholakia, usually associated with virtual communities and Bagozzi and Pearo, 2004); help firms learn from call for more research on mass communication in customers (Kardaras, et al., 2003); increase web- virtual communities (Schoberth et al., 2006). site traffic (Hagel and Armstrong, 1997); raise Schoberth and colleagues (2006) found, among entry barriers for competitors (Hagel and Arm- other things, heterogeneity in community partici- strong, 1997); facilitate product development pants' activities. Scholars have also called for efforts (Nambisan, 2002); and increase customer more quantitative research using behavioral data satisfaction and loyalty (Shankar, et al., 2003). from virtual communities (de Valck et al., 2007; Recently, social network sites caught the attention Casalo et al., 2008). Thus, we ask: how do mem- of users, firms, and researchers (Boyd and Elli- son, 2007). Sometimes labeled “Web 2.0” coined

23 For more details: 22 Source: http://www.oreillynet.com/pub/a/oreilly/tim/news/200 http://www.comscore.com/press/release.asp?press=2698 5/09/30/what-is-web-20.html

56 MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES • THE CASE OF ROOSTER TEETH PRODUCTIONS

bers of a virtual community make use of modding?

57 FROM GAME TO PLAY PETER JÄGER • 2010

2. METHODOLOGY

We conducted a large quantitative study on the turing two teams of soldiers in the game Halo virtual community of Rooster Teeth Productions, who are stationed in an isolated canyon where a Machinima Production company creating and their sole purpose is to fight each other. The publishing animated videos made in computer popularity of the show that first aired April 1st, games. We present the sample case as well as data 2003 profited from the humorous dialogues be- gathering and analysis in this section. tween the different characters. While the comedy was first aimed at other gamers, a broad audience swiftly appreciated RvB. To date, Rooster Teeth 2.1 Sample Case has released five seasons of RvB ‘The Blood Producing animated videos was previously re- Gulch Chronicles,’ and one season of RvB ‘Re- stricted to media professionals because of the construction’ comprising 20 to 25 episodes each high cost of software packages. These restrictions as well as several spin-off mini-series. Over the led innovative users to produce animated shorts years, shooting the movies has advanced from the with computer games by using the underlying 3D game Halo 1 on the xBox to the latest release render technology thus creating Machinima. Halo 3 running on xBox 360 with overwhelming Game engines were relatively cheap compared to new possibilities in graphics and artistic composi- traditional production tools. In addition, most of tion. In addition, most of their merchandising the in-game assets like characters and landscapes articles were related to RvB, which remained the were already at hand, which reduced the overall flagship show. Apart from RvB, Rooster Teeth production time for an animated movie signifi- produced several other shows including ‘The cantly. Strangerhood’, ‘P.A.N.I.C.S.’, or ‘1-800-Magic’, using different game engines to shoot the films. Rooster Teeth Productions is one of the most successful Machinima companies (von Krogh et Each series had its own website on which the al., 2009). They sell sponsorship subscriptions, videos were shown, important announcements merchandising, and DVDs and reach a large user from Rooster Teeth staff members published, community. The latter, in fact, was triggered early and where fans discussed topics around the show. on when Rooster Teeth introduced an elaborate The discussion took place where the videos were community platform offering SNS features: viewed – especially while viewers waited for the download to finish or directly after watching vid- “… well, I think a lot of it has to do with the fact that eos online. Users did not have to actively go to a the community site that we have made … or at least at website to express their thoughts about the prod- the time we made it … had features that weren’t that uct as is the case with most websites of communi- present in other places, we were a little ahead of the curve ties of consumption. at that time, and so there were a lot of cool features that Due to a steadily growing fan base, over the last people were interested in. This is like before MySpace four years the segment of the RvB community really had taken off … So we’ve always tried to give it a actively contributing to discussions grew to little functionality, things they do in a community website 42,000 members who posted more than 400,000 they’re interested in making … you know, interested in comments on 165 episodes. Members could being a part of it. We tried to make the website almost choose their level of engagement. They could be like a game.” Geoff Ramsey, Rooster Teeth Productions mere “consumers” who just watched the videos Rooster Teeth Productions was founded in 2003 and/or bought merchandising products without by Burnie Burns, Matt Hullum, Geoff Ramsey, interacting, or they could interact with other Jason Saldaña, and Gus Sorola in Austin, Texas. community members. The Rooster Teeth com- Their first and most widely known Machinima munity cannot be neatly classified as either VC or production was Red vs. Blue (RvB), a show fea- SNS since different users engaged differently.

58 MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES • THE CASE OF ROOSTER TEETH PRODUCTIONS

ture and therefore the amount of comments was low in section 1. Section 2 followed the launch of the new website before Season 3 leading to a steep increase in comments eventually coming to a slow- down after the end of the season. Section 3 was the most commented section ever covering Season 4. Half of the total comments-population was found in this section. This Figure 1. Age Distribution of Members Commenting on Red vs. Blue. finding does not imply that all comments were made during Season 4 since it With 16% of all members, the 18 year olds repre- was possible to comment 4 and 5 covering Season sented the largest group (see figure 1). The aver- 5 and the start of RvB ‘Reconstruction’ respec- age age of members was 21 years with a standard tively contain again relatively little commented deviation of eight years. The age distribution was products. biased and positively skewed by the fact that The basic units of analysis were the RvB-related members who didn’t enter any age were listed as comments made by members and the mod-points zero, and that few members who apparently en- associated with the comments. These were dis- tered the maximum age of 88 years. 93% of the played chronologically below the corresponding members were under 30 and the bulk was either video similar to YouTube with the difference that in high school or college-age. the comments in our case were ordered by as- Tracing the amount of members over four years, cending post date (i.e. the oldest post was dis- we found that the community had been growing played first). All comments and the associated at different speeds, but steadily in volume in a mod-points were publicly accessible. In order to nearly linear fashion (r square= 0.95). There were leave a comment one had to be signed in as four visible gaps in signups, which were located in community member. Member accounts were free the first two years of its existence with the longest of charge and did not have to be activated by a gap lasting for two weeks (see figure 2). moderator or an administrator. Hence, members The amount of comments per episode varied were able to sign up at any time and start posting. from a minimum of 58 to a maximum of more than 28,000 with an average of 2,400 comments per video (Figure 3). Five different sec- tions could be identified with a strong cyclicity given that the amount of comments increased notably during seasons: Section 1 represents the comments to Season 1 and 2 that were aired on a former version of the Rooster Teeth community website.

Those comments were not migrated to the new and more Figure 2. Accumulated Daily Sign-ups of Members over the last four Years. elaborate software infrastruc-

59 FROM GAME TO PLAY PETER JÄGER • 2010

Figure 3. Comments per Video. Episodes ordered chronologically.

Comment modding is the act of rating another who has one account can mod each comment by member’s comment(s). Synonyms are ‘rating’, another user only once. The mod is then publicly ‘giving mod-points’ or simply ‘modding’. In the displayed next to the comment. Rooster Teeth community each modding of a comment consists of a combination of two values: a numerical value and a qualitative rating. 2.2 Data Collection and Analysis The numerical value is either ‘+1’ or ‘-1’. Each For the purpose of the quantitative data analysis, numerical value has to be combined with one of we built up a database dedicated to the case un- four qualitative ratings from which users can der study. All available data from the Rooster choose in a drop down menu. The four qualita- Teeth RvB website concerning the episodes, the tive ratings corresponding to ‘+1’ are ‘Cool, members, and the comments was automatically 24 Ditto, Funny and Zing! ’. The four qualitative fetched during a three-day period from Septem- 25 ratings corresponding to ‘-1’ are ‘WTF , Lame, ber 20th to September 22nd, 2008 and transferred 26 27 Flamebait and Noob ’. Mod points could only to a local MySQL database for further analysis. be given once per user and per comment. A user To be granted full access to all the data, we obtained a sponsorship account. After screening and 24 Three possible definitions for our purpose: 1) New term evaluating the data, we discovered some missing for “owned”, said after saying something witty to someone in an insulting manner. 2) If someone makes an absolutely data sets that had been left out due to server awful joke, or says something completely random or maintenance by Rooster Teeth. For this, we ob- pointless. One member of the group may "zing" them. 3) tained the missing data sets on October 2nd. All A noise made when a person, place, or thing is data entries in the local database indicate their discriminated against in a humorous manner. Source: fetch time stamp to check for possible inconsis- www.urbandictionary.com tency. We rebuilt the relational database structure 25 Internet slang acronym for “What the fuck?” of the original website using a separate table for 26 A message posted to a public Internet discussion group, episodes, members, and comments which were such as a forum, newsgroup or mailing list, with the intent of provoking an angry response (a "flame") or argument linked by their dataset identification number ‘id’ over a topic the troll often has no real interest in. Source: that remained the same as the online PHP web www.wikipedia.org queries. 27 Short for “Newbie”. A slang term for a newcomer to We fetched a total of 42,771 member accounts online gaming or an Internet activity. Source: www.wikipedia.org and 483,272 comments with their corresponding

60 MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES • THE CASE OF ROOSTER TEETH PRODUCTIONS

information. Out of all 737,000 registered Rooster Teeth community members28, only those who at least commented once on a video of RvB were considered. Cleaning the fetched data sets from invalid information (either comments which link to a NULL member id or comments which link to empty member profiles) left us with 406,173 comments and 41,016 user profiles (see Table 1). SPSS, Excel and the phpMyAdmin interface of the local server were used for the quantitative data analysis.

28 http://rvb.roosterteeth.com/members/stats/

61 FROM GAME TO PLAY PETER JÄGER • 2010

3. RESULTS

Almost half of all members posted at least one same position (post number) [1 to 32,780]1 and comment, which has been modded, but only 15% divided this value by the amount of modded of all comments were modded (see Table 1). One comments [1 to 165]29 per position. We thus cal- possible explanation could be information over- culated the average mod points per comment load (de Valck et al., 2007). Members cannot position. In effect, we used the absolute mod browse the overwhelming amounts of comments value for a better representation of the attention that are posted. Observing modding behavior in a comment received, than the net value. more detail, we find that 60% of the modded We then ordered this quotient by descending comments obtained positive values (cumulated value, i.e. starting with the highest value in order mod rating > 0), 36% obtained negative values to receive a non-scaled ranking. For example, the (cumulated mod rating < 0), and for 4% of the data point at y=1 is calculated as follows: We

Table 1. General Statistics for Modding Behavior in the Rooster Teeth Online Community. comments the mods evened out (cumulated mod considered all comments with post number 1. rating = 0). The fact that 60% of all comments Since all 165 videos were commented at least carried a positive rating, with “Cool” being the once, 165 comments resulted. Some of these predominant rating class, showed that members comments appeared to have invalid data base generally tended to give friendly mods. entries on the website. After pruning those, we Next, casual observation suggested that modding were left with 147 valid comments. Out of these behavior centered around early comments. We 147 comments we only examined those that were analyzed the attention different comments re- modded. In post number 1, all 147 were modded. ceived based on their position in the comments We then summarized 147 absolute mod values, thread (see Figure 4). This position corresponded and divided the sum by 147, resulting in to the time the comment had been posted in as- 36,921/147= 251. Repeating this procedure for cending order i.e. the comment that was posted all comments that were posted second (position 2) first is at position 1, second at position 2, and so we get 107 respectively. Next, we ordered the forth. Post numbers are displayed on the x- quotients by descending value and displayed coordinate. For the y-coordinate we defined and them as ranking. Thus, the y-axis represents the calculated a ranking variable. We summed [0 to ‘attention’ or the valuation (negative or positive) 36,921] the absolute values [0 to 2,920] of all mod points given to comments which share the 29 The maximum amount of 165 videos that could be commented on limited this number.

62 MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES • THE CASE OF ROOSTER TEETH PRODUCTIONS

members accorded a comment where 1 is the top average was positive and statistically significant rank with the highest attention. (beta(14665) = 0.578 ; p ≤ 0.01). The results show that the first comments on each video received on average more (absolute) mod points than the respective subsequent comments. This holds true to a certain comment position from which on the data points become scattered. A threshold seems to appear around post number 50. The relationship between the post number of a comment, that is its position, and the mod value it received on

Figure 4. Relationship between the order in which comments were posted (ascending) and the rank based on the average (absolute) mod value the respective comments received (ascending; average to rank inversely propor- tional): both scales were logged.

63 FROM GAME TO PLAY PETER JÄGER • 2010

4. DISCUSSION

This exploratory study of modding behavior in a longitudinal studies of posting and modding be- virtual community revealed three findings that havior in order to identify changes in community open up for future research. First, just under half behavior over time. of the community members received mods on Third, comments that appeared early and high their contributions, while the other half did not up in the (chronological) list of comments re- received mods. Looking at the entire volume of ceived disproportionately more mods than subse- comments, only 15% were modded. Second, the quent, less visible comments. A first interpreta- community studied leaned towards ‘positive tion suggests that community members may suf- modding’, with 60% of all mods being positive. fer from information overload and pay far less Third, the time and location of a comment mat- attention to later comments than to early com- tered strongly for the likelihood that it would be ments. This finding raise doubts regarding the modded. Comments that appeared early after the high expectations by some authors attached to release of a new product and appeared on the modding and rating systems as quality signaling first two pages of comments, received dispropor- or filtering tools (Deng et al., 2008; Yang et al., tionately high amounts of mod-points. After ap- 2009). Should the timing and position of a com- proximately 50 contributions, the direct link be- ment matter more than its quality in predicting tween the position and the rank of mod-points the number of mods received, the modding sys- weakened. tem may be of little use to managers, marketing These findings warrant further research on vir- experts, and users of virtual communities. How- tual communities with SNS features in three ar- ever, this issue needs much more investigation in eas: individual behavior, collective behavior, and future studies. We observed that after a certain community structures. First, roughly half of the threshold the post number did not predict the community members never receive mods on their number of aggregate mods received. This calls comments. The behavior does not seem to catch for a refined analysis across multiple contexts and on throughout the member base. The extension communication structures. Is it important that to mass communication in virtual communities the first page contains 30 comments? Does the provided by modding seems to be used unevenly. chronological order matter or could it be reversed Hence our results extend the findings of and produce the same pattern? Schoberth and colleagues (2006) on heterogene- Managers of virtual communities and social net- ous communication behavior in online communi- work sites may take away three insights from our ties. Future research should analyze the factors study. First, virtual communities gain significantly that explain this behavior. Is modding considered new characteristics by adopting features associ- to be costly, either in giving or in receiving? Is ated with social network sites. Managers may modding contested? Do member demographics think of more effective ways of distinguishing explain modding behavior? Further, how does communities, possibly based on posting or mod- modding impact on contributions? Do members ding behavior by members. The case of Rooster who received negative mods learn or change their Teeth Production provides evidence as to the behavior? Do positive mods (or mods at all) in- successful combination of product feedback and duce participation? social network site features. Community members Second, we observed a friendly community who comment on a firm’s products when they are distributed more positive than negative mods. released. They evaluate each other’s comments This result may impact on community growth, and make use of the social infrastructure pro- the willingness of members to contribute, and vided. Second, the modding behavior confirmed ultimately, consumer behavior. The result could the impression of a friendly community. While support the idea that mods express trust rather this is only a first, preliminary finding it shows than distrust or disapproval (Kim et al., 2008). that the option of modding other community What explains this bias? How does this finding members’ contributions was being used in a ‘pro- compare to other communities’ behavior? Can ductive and supportive manner’. In general social communities that lean towards offensive behavior network site features could be meaningful exten- be sustained? Also, researchers should conduct sions to existing virtual communities. Third, fil-

64 MODDING AS RATING BEHAVIOR IN VIRTUAL COMMUNITIES • THE CASE OF ROOSTER TEETH PRODUCTIONS

tering valuable comments with the use of mem- logically. This may mean that after the first rush ber-based modding tools may not be a simple by people to make their comments visible, per- matter. Our results show that only after about 50 haps later mods may signal high-quality com- comments the mods received started to deviate ments. from the comment number as received chrono-

65 FROM GAME TO PLAY PETER JÄGER • 2010

REFERENCES

Algesheimer, R., Dholakia, U. M., & Herrmann, A. 2005. Kim, Y. A., Le, M. T., Lauw, H. W., Lim, E. P., Liu, H. F., The social influence of brand community: Evidence from Srivastava, J., & Ieee. 2008. Building a web of trust without European car clubs. Journal of Marketing, 69(3): 19-34. explicit trust ratings. Paper presented at the 24th IEEE International Conference on Data Engineering, Cancun, Algesheimer, R., & Dholakia, P. M. 2006. Do customer MEXICO. communities pay off ? Harvard Business Review, 84(11): 26- +. Kozinets, R. V. 1999. E-tribalized marketing?: the strategic implications of virtual communities of consumption. Euro- Armstrong, A., & Hagel, J. 1996. The real value of on-line pean Management Journal, 17(3): 252-264. communities. Harvard Business Review, 74(3): 134-&. Nambisan, S. 2002. Designing virtual customer environ- Bernoff, J., & Li, C. 2008. Marketing - Harnessing the power ments for new product development: Toward a theory. of the oh-so-social web. Mit Sloan Management Review, Academy of Management Review, 27(3): 392-413. 49(3): 36-+. Ocasio, W. 1997. Towards an Attention-Based View of the Boyd, D. M., & Ellison, N. B. 2007. Social network sites: Firm, Strategic Management Journal, 18(S1), 187-206. Definition, history, and scholarship. Journal of Computer- Mediated Communication, 13(1): 210-230. Pascu, C., Osimo, D. Ulbrich, M., Turlea, G. and J.C. Bur- gelman. 2007. The potential disruptive impact of Internet 2 Casalo, L. V., Flavian, C., & Guinaliu, M. 2008. Promoting based technologies. First Monday, 12(3). consumer's participation in virtual brand communities: A new paradigm in branding strategy. Journal of Marketing Porter, C. 2004. A Typology of Virtual Communities: A Communications, 14(1): 19-36. Multi-Disciplinary Foundation for Future Research. Journal of Computer-Mediated Communication, 10(1): 00-00. de Valck, K., Langerak, F., Verhoef, P. C., & Verlegh, P. W. J. 2007. Satisfaction with virtual communities of interest: Porter, C. E., & Donthu, N. 2008. Cultivating trust and Effect on members' visit frequency. British Journal of Man- harvesting value in virtual communities. Management Sci- agement, 18(3): 241-256. ence, 54(1): 113-128. Deng, S.Y., He, L., Xia, W.W. 2008. A Collaborative Filter- Rheingold, H. 1993. The Virtual Community: Homestead- ing Algorithm Based on Rating Distribution. IEEE Proceed- ing on the Electronic Frontier: {Addison-Wesley Pub. Co., ings. International Symposium on IT in Medicine and Edu- Reading, Mass.}. cation. 1-2: 1118-1122. Schoberth, T., Heinzl, A., & Preece, J. 2006. Exploring Dholakia, U., Bagozzi, R., & Pearo, L. 2004. A social influ- communication activities in online communities: A longitu- ence model of consumer participation in network- and dinal analysis in the financial services industry. Journal of small-group-based virtual communities. International Jour- Organizational Computing and Electronic Commerce, 16(3- nal of Research in Marketing, 21(3): 241-263. 4): 247-265. Hagel, J., & Armstrong, A. 1997. Net Gain: Expanding Shankar, V., Smith, A. K., & Rangaswamy, A. 2003. Cus- Markets Through Virtual Communities: {Harvard Business tomer satisfaction and loyalty in online and offline environ- School Press}. ments. International Journal of Research in Marketing, 20(2): 153-175. Kardaras, D., Karakostas, B., & Papathanassiou, E. 2003. The potential of virtual communities in the insurance indus- Yang, J.M., Li, K.F., Zhang, D.F. 2009. Recommendation try in the UK and Greece. International Journal of Informa- based on rational inferences in collaborative filtering. tion Management, 23(1): 41-53. Knowledge-Based Systems, 22(1): 105-114.

A DIRECTING AUDIENCE

HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

Peter Jäger Georg von Krogh Stefan Haefliger

Chair of Strategic Management and Innovation Department of Management, Technology, and Economics (MTEC) ETH Zürich, Switzerland

ETH Zürich Working Paper 2010

Virtual online communities and their potential for value co-creation receive increasing interest from both, academia and firms. While communities of production and their contribution to new product development have been studied by innovation scholars, less is known about the value creation in communities of consumption that traditionally fall into the domain of marketing research. We inductively explore how members of a community of consumption of a Machinima company, that produces animated films in video games, contribute to innovation. This research was motivated by a lack of empirical evidence regarding the content of product feedback in a community of consumption and its innovation-relevance. We collected over 12,000 unsolicited comments by users and coded almost 7,000 comments individually to categorize and evaluate them in terms of their potential contribution to innovation, both process and product innovation. The results show that members of the community of consumption studied generated specialized, unsolicited feedback, and 10% of the members contribute more than once. The detailed knowledge of the product’s history allowed users to comment on the possible continuity of product development, which resulted in product innovation. The combination of technical expertise and knowledge about the product allowed users to generate feedback relevant for process innovation and a strong interest in detail supported the firm’s quality assurance.

67 FROM GAME TO PLAY PETER JÄGER • 2010

1. INTRODUCTION

This paper examines the relevance of feedback back, reuse, planning, and others centering on from consumers30 who are active in virtual online the task of creating specific products that fit their communities and the impact on firm innovation. needs (von Hippel, 2005). For technical or physi- Large numbers of individuals using the Internet cal products, online communities of production form various types of virtual online communities. have become established as platforms to ex- A fundamental distinction exists in the produc- change information and innovate. tion versus consumption orientation of these Consumers who form communities of consump- communities. This distinction also mirrors the tion are passionate about certain products. They interests in such communities by innovation and build and explore a lifestyle online that supports marketing researchers, respectively. Online com- and develops their strong emotions towards the munities that produce software or ideas for new existing product (Cova & Pace, 2006; Kozinets, product development have received attention 1999, 2002; Muniz & O'Guinn, 2001). Virtual from innovation scholars (Franke & von Hippel, communities of consumption, such as brand 2003; Lee & Cole, 2003), whereas communities communities, unite consumers in their consump- that consume physical or virtual products have tion of products and services and their devotion been well described by marketing scholars (Alge- to the brand (Picher & Hemetsberger, 2007) or as sheimer, Dholakia, & Herrmann, 2005; Cova & fans of TV series described as fandom (Kozinets, Pace, 2006; Kozinets, 2007; McAlexander, 2001). While interacting, these members create Schouten, & Koenig, 2002; Schau & Gilly, 2003). value for firms in different ways: they support a Both, production and consumption types of on- product or service, promote a brand and spread line communities differ significantly not only in loyalty to a product or firm, or act as a resource terms of their purpose and shared practice but for ideas (Carlson, Suter, & Brown, 2007; Kozi- also in their organization and their contribution nets, 1999; McAlexander et al., 2002; McWil- to innovation and value creation. liam, 2000; Muniz & O'Guinn, 2001; Nambisan, Research on new product development and cus- 2002; Rowley, Kupiec-Teahan & Leeming, 2007; tomer co-creation (cf. user innovation) has identi- von Hippel, 1988). fied incentives and individual motivation for (vol- Member-to-member assistance in online com- unteer) community members to engage in inno- munities has been studied both as an emergent vation activity (Hertel, Niedner, & Herrmann, phenomenon among consumers (Lakhani & von 2003; von Hippel & von Krogh, 2003). Consum- Hippel, 2003) and from the perspective of a firm ing products can lead to the desire among its sponsoring a community (Franke, Keinz & consumers to improve the product in order to Schreier, 2008; Pitta & Fowler, 2005; Wasko & better meet their needs and in other ways exploit Faraj, 2005; Wiertz & de Ruyter, 2007). Member- their contextual knowledge of how to improve to-member assistance can also play a role for the product (Baldwin, Hienerth, & von Hippel, product support or customer service. Construct- 2006; Füller, Matzler, & Hoppe, 2008; Herstatt & ing identities for online self-presentations to fel- von Hippel, 1992; von Hippel, 1994). The con- low consumers, community members typically sumers’ activities include experimentation, proto- leverage the brand or product association (as is) typing, constructing, exchange of solutions, feed- by expressing their loyalty and devotion (Cova, 1997; Picher & Hemetsberger, 2007). Consumers’ 30 The term consumer can be used interchangeably with activities also include creative endeavors such as users. While the innovation literature refers to end- described by Kozinets (1997) with the x-philes or customers as users, the marketing literature prefers the StarTrek that lead to idea generation and creative term consumer. In this vein, user innovation corresponds thinking which may in turn contribute to innova- to customer co-creation. We use the term consumer throughout this paper, except for user innovation as tion (Kozinets et al., 2008). However, the contri- established term, since the members of our focal bution happens as a by-product of the consum- community consumer media products, hence the term ers’ online activities and occurs solely among user appears somewhat inappropriate. In the sample of members and not directed towards the producer. this research, consumers are simultaneously the direct (online) customers of the firm since no intermediaries are Recently, Kozinets et al. (2008) introduced the involved. idea of creative and innovative consumers in

68 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION online communities but an in-depth analysis of inductively building theory on consumer feed- the content of contributions in view of how pro- back behavior for marketing and innovation. ducers may evaluate the ideas towards product Analyzing the details of the feedback behavior and process innovation is still lacking in the litera- towards the products published we find that ture. Kozinets and colleagues (2008) thus show community members provide feedback that falls the potential but not the actual impact of online into different categories. Even though the firm community activity on firm innovation perform- defines no specific tasks or motivates the mem- ance. As long as the content provided by consum- bers and does not solicit specific types of feed- ers during their online activity and the categories back, members not only discuss ideas among of contributions remain unexplored, the link themselves, as would be expected from existing between online consumer behavior and innova- literature, but also address the producers directly tion outcome is disputable. Applying a grounded with specialized comments. This is interesting theory approach, we inductively explore the ques- since firm members are not actively contributing tion: How does feedback from members of to the discussions. The discussion and comment- a community of consumption contribute ing behavior implies that members of communi- to firm innovation? ties of consumption that are unattended by spon- We explore the process of consumers’ comment- soring firms – hence firms do not show active ing on products while simultaneously interacting involvement in the community like setting tasks, with each other by making use of social software initiate discussions, or conduct marketing surveys features designed into the virtual community – do specialize and engage in innovation relevant infrastructure. Our study focuses on a Machinima activities giving non-solicited feedback for prod- producer: Rooster Teeth Productions, located in uct improvements (production techniques) or Austin, Texas. Machinima is the art of producing sharing ideas for new product designs (storyline animated films in video games by using the 3D changes). Even though we could show in this render technology of the game engine, which in study that consumer feedback directly impacted this process serves as a tool for the filmmakers. on firm innovation, we were not able to depict Examples of Machinima films can be watched at how the firm extracts the relevant knowledge www.machinima.com, www.machinima.org, or from the community platform. This is an impor- www.roosterteeth.com. A number of books cover tant subject for follow-up research. the topic for the general reader (Hancock & In- Beyond sharing their ideas, which is innovation- gram, 2007; Lowood, 2007; Marino, 2004; Mor- relevant information for the firm (Gurteen, 1998), ris, Kelland, & Lloyd, 2005). The paper contrib- consumers also structure this information and utes to the existing literature on 1) marketing on give advice for future innovations by directly ad- the web, 2) digital product management, as well dressing the producer in their postings. A practi- as 3) social dynamics of communities on the In- cal implication of this study is that managers ternet. We combine marketing literature regard- should skim for relevant contributions. This de- ing communities of consumption and brand mands an adequate IT infrastructure and intelli- communities (Algesheimer et al., 2005; Cova & gent filter mechanisms to isolate feedback catego- Pace, 2006; Kozinets, 2007; McAlexander et al, ries as well as consumers who show above aver- 2002; Schau & Gilly, 2003) (1) with innovation age performance in providing feedback and ideas literature focusing on communities of production for innovation (cf. von Hippel, 1988; 2005 on (Franke & von Hippel, 2003; Lee & Cole, 2003) lead users). (2) in order to better understand and position The paper is organized as follows: In the next these different types of communities along the section, we introduce the types of communities dimensions of firm involvement and creative described in the literature to motivate our induc- output (Porter, 2006). We also contribute to the tive study. Subsequently, we describe the research discussion on social dynamics in online communi- design and present the results of this study. Limi- ties by (Jeppesen & Frederiksen, 2006; Moore & tations as well as the discussion of implications Serva, 2007; Rheingold, 1993; Tafjel, 1978) (3) for research and management practice conclude this paper.

69 FROM GAME TO PLAY PETER JÄGER • 2010

2. TYPES OF COMMUNITIES

Building on Kozinets et al. (2008), we examine socialization and “relational intelligence” that the impact of members’ activities in a virtual promotes learning with partners, rather than online community of consumption on firm inno- from partners such as suppliers and customers vation. We follow the definition of Gurteen (2000: 47-48) in a community “that is governed (1998) and consider innovation-relevant informa- by a central firm that acts as the sponsor and tion as creative contributions by members that defines the ground rules for participation” (2000: may – as the basic building blocks – translate into 25). Their discussion shows the shift from merely innovation-relevant knowledge, which eventually exploiting customer knowledge to knowledge co- enables a firm to innovate. Thus creativity gener- creation, which takes place within communities in ates ideas (information) while knowledge (inter- order to spur creativity and generate new knowl- preted information) is an input to innovation (see edge. However, while the authors do not specify also Amabile, Conti, Coon, Lazenby & Herron, the organizational processes that accompany this 1996; Amabile, 1998). Organizational and indi- shift, Sawhney and Prandelli describe the forma- vidual characteristics and the collective behavior tion of communities of creation as “a reconfigu- that leads to value creation in online communities ration of cognitive labor through information are subject to an emerging stream of work technology.” (Sawhney & Prandelli, 2000: 49). (Sproull, Dutton, & Kiesler, 2007). One approach Comparing the communities that are described in to address the question how members of com- current literature (see Table 1)31 reveals differ- munities of consumption create value in form of ences along many possible dimensions such as contribution to innovation might be looking at type of activity, task orientation, member interac- the intersection of communities of consumption tion, firm involvement, and more. The activities with communities of production. and the output generated by the members vary

Table 1. Community types and cases in current literature

In an early contribution, Sawhney and Prandelli along the types of communities. Communities of (2000) located their concept of a community of creation on the intersection of communities of 31 While Sawhney & Prandelli (2000) consider Linux to be a production and consumption (see Appendix Fig- Community of Creation, we tend to understand it as a ure A1). The authors refer to the development of community of production.

70 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION consumption deliver information such as discus- tion (from an R&D perspective)? Do consumers sions of products or reports on usage, communi- translate creative ideas into innovation-relevant ties of creation generate knowledge by linking knowledge? personal experience with product tests or answer- A special focus on communities of consumption ing marketing surveys, while communities of seems critical and warranted for a number of production innovate virtual products or services reasons that cast doubt on the likely success of or – for physical products – contribute detailed finding innovation-relevant knowledge in unsolic- building instructions online. Firm involvement ited feedback from consumers. Even though takes different levels from very active participa- communities of consumption and creation have tion where firms delegate employees to contribute been observed to generate insights about products to discussions and organize brand feasts for that may help other consumers or inspire manu- community members, to passive involvement facturers to learn about trends in the market where firms just offer the online space for interac- (Nelson & Otnes, 2005; Sawheny & Prandelli, tion to their consumers, to no involvement which 2000; Rowley et al., 2007), members of commu- happens in brand or fan communities organized nities of consumption seem unlikely to engage in by consumers themselves. activities that directly transform into innovation Partly overlapping with the typology above, for the following three reasons: Kozinets et al. (2008) draw their distinction First, the consumers’ passion for a product or within online creative consumer communities product category leads them to express their pref- along the two dimensions of collective innovation erences for certain brands, features, and accesso- orientation and collective innovation concentra- ries in online communities. (Kozinets, 2002). tion referring to the context, in which the creative Their efforts focus on discussing existing products contribution is occurring, and the number of rather than ideas for new products that could be community members contributing (see Figure A2 incorporated by the sponsoring firm. Community in the Appendix). Regarding their attribution to members assist each other in the consumption of task completion, we consider “Crowds” and a product or in promoting a product rather than “Hives” as part of communities of creation while solve problems online. In fact, firm representa- “Mobs” and “Swarms” fall into the category of tives are frequently absent from the community communities of consumption. The activities of and may or may not agree with what is being swarms, “small actions by individuals – hyperlink- discussed at all (Kozinets, 1999, 2002; Muniz & ing, going to a popular Web site or portal, rating O'Guinn, 2001). From our reading of the litera- or commenting, and tagging – all have dramatic ture it appears that most, if not all research, on innovative effects when aggregated by software, communities of consumption was conducted site, and competitive marketing strategy (Kozinets form the point of view of marketing rather than et al, 2008: 350).” While implying the innovation R&D. effect, Kozinets et al. (2008) neglect to show em- Second, the community members take advantage pirical evidence for the innovative output or the of the shared context online to present them- way in which firms make use of it. selves; they understand consumption as self- Product design contests or marketing surveys defining and self-expressing (Belk, 1988; Cova, create interaction with consumers where firms 1997; Cova & Pace, 2006; Pichler & Hemets- take the initiative to solicit ideas and feedback. To berger, 2008; Schau & Gilly, 2003). According to our knowledge, no study has explored the link Goffman (1959), the presentation of self is con- between consumer creativity and its exploitation textual and targeted to a specific audience. Kozi- for firm innovation without prior, direct con- nets (2007) described this presentation of mate- sumer-firm interaction. Further validating the rial that is of interest to oneself and fellow com- model by Kozinets and colleagues (2008) thus munity members as “egoboo”. The considerable calls for a grounded theory approach, given the effort consumed by setting up and maintaining current lack of empirical work. Advancing Kozi- individual websites and selecting and displaying nets et al. (2008) study on consumer creativity in images contrasts with the work of collectively the direction of firm benefit from innovation, we solving problems online. Consumers tend to pre- explore the nature of unsolicited feedback and if sent themselves as person using the product as it is specialized or innovation-relevant to answer well as the context they use it in while the self- the research question of this study. This means in presentation of producers is incorporated in the more detail, does creativity translate into innova- created product itself. Producers thus identify

71 FROM GAME TO PLAY PETER JÄGER • 2010

with the product while consumers identify with tion and turn it into valuable knowledge?’ seem the use or application context of the product. to be understudied (Hemetsberger, 2006). Third, shared values and their diffusion, amplifi- The community sites of communities of con- cation, and promotion play an important role in sumption offer a similar infrastructure as those of communities of consumption (Dholakia, 2004; the communities of production to support inter- Muniz & O'Guinn, 2001). The narratives that action and sharing of content among community establish the values connected to a brand or members (Granitz & Ward, 1996; Kozinets, products develop from the history of the prod- 1999; McAlexander et al., 2002; Muniz & ucts. Members of the Saab community for in- O'Guinn, 2001). The social software design instance share stories about their experience with cluding available features as well as the usability the car during “an eventful, sometimes harrow- of the community infrastructure are vital for ing, but always meaningful journey in their members’ participation and interaction and in- Saabs” (Muniz & O'Guinn, 2001: 423). The teract with one another. (Dholakia, Bagozzi, & identities and values associated with a brand cul- Pearo, 2004; Huysman & Wulf, 2006; Jang, minate in an “informed celebration of mass ma- Olfman, Ko, Koh & Kim, 2008) Thus, the com- terial culture.” (Muniz & O'Guinn, 2001: 428). munity site would provide the basis for members Similar to the presentation of self, the promotion to contribute to innovation. of values places a different emphasis in commu- In the following, we use a netnographic ap- nities of consumption from that of the develop- proach, suited to the online community environment of new products. For communities of pro- ment (Kozinets, 2002; 2010), to analyze the audi- duction, values may be closely linked to the pro- ence of Machinima, an emerging genre involving duction of new technology (Ke & Zhang, 2009; filmmakers who apply engines and artwork from Stallman, 1999). the video game industry to shoot films inside Since members of communities of consumption video games. The audience forms a large com- appear to be unlikely candidates to contribute munity of consumption and contributes to the actively to the innovation of the consumed prod- development and improvement of the films they ucts, how can the community output be used and consume by feeding back their ideas to the firm be of relevance for a company to promote inno- hosting the community. This firm publishes the vation? Considering the vast amount of informa- Machinima films Considering the classification tion created by members of communities of con- introduced by Kozinets et al. (2008), the commu- sumption, questions such as: ‘How do these nity represents a “swarm” meaning many mem- communities contribute to firm innovation, if at bers contribute relatively little “as a part of the all?’ or ‘What does it take to stimulate creativity normal routine of being involved in online com- on the one hand and, how, on the other, are firms munities” (Kozinets et al., 2008: 344). We intro- able to extract the innovation-relevant informa- duce our research design next.

72 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

3. SAMPLE AND RESEARCH DESIGN

Research for this paper comprised three phases: consumers frequently possess advanced gaming selection of the case, data gathering, and data skills which enables them to give qualified feed- analysis. With very little existing theory to build back; and (5) the community is in a rather mature on that has not been empirically tested yet, we state which implies a sufficient dedication to both applied a grounded theory approach. Eisenhardt brand and product. (compare Kozinets et al., (1989: 537) suggested, “cases may be chosen to 2008 on Swarms; see also Table 1 for studies on replicate previous cases or extend emergent the- communities of consumption, especially Rowley ory”. For studying communities’ activities in an et al., 2006) online context, Kozinet’s netnographic approach (Jorgensen, 1989; Kozinets, 1998; 1999; 2010) is appropriate. We conducted a single case study 3.1 Selecting the Case (Yin, 2003), and gathered data from the online Producing animation had been restricted to me- community website of the selected Machinima dia professionals as only they could afford the company: Rooster Teeth Productions. We thor- expensive software packages necessary. These oughly explored the phenomenon on the basis of restrictions led innovative filmmakers to produce this company. Theory development on a limited animated shorts with video games, by using the sample has been discussed by March et al. (1991), underlying 3D render technology of the game Eisenhardt (1989), and Siggelkow (2007), while engines. Games like Halo, The Sims, Unreal Vaughn (1990), Lawrence et al., (2002), and Tournament, Battlefield, The Movies, World of Pervez et al. (2008) made significant contributions Warcraft, or SecondLife are relatively cheap based on a qualitative research design with one or compared to traditional production tools. In ad- few cases. Following an inductive logic we de- dition, most of the in-game assets such as charac- velop five propositions derived from the case un- ters and landscapes that resemble actors and der study to describe the impact of community scenes are already at hand. The overall produc- members’ feedback behavior on firm innovation. tion time for a film can be reduced significantly (Cohen, 1980; Eisenhardt, 1989; Eisenhardt & by reusing these in-game assets. Graebner, 2007; Glaser & Strauss, 1967; Strauss & Corbin, 1990) Having studied several different Machinima companies during the past few years yielded a Moving from data to constructs, two investigators good understanding of the Machinima scene. started with an open coding of the gathered Out of the companies studied, Rooster Teeth comments (Glaser, 1998) and triangulation for Productions seemed to be the most successful stronger substantiation (Eisenhardt, 1989). The regarding the commercialization of their various applied coding scheme is presented in Table 4. products including sponsorship subscriptions, We then grouped the codes into different catego- merchandising, and DVD sales as well as their ries. All available information was shared within reach to their consumer community. The latter, in the research team using a common database. In fact, was triggered early on when Rooster Teeth the following, we describe sampling, data gather- introduced an elaborate community platform ing and data analysis. offering the functionality of today’s social soft- Rooster Teeth Productions’ Red vs. Blue online ware sites such as Facebook, MySpce, Flickr, or brand community offered an ideal context to YouTube before these had been launched or were explore our research question because (1) the at least widely known. Rooster Teeth Productions community is hosted by a producer firm while was founded in 2003 by Burnie Burns, Matt Hul- consumers are left unattended and do not par- lum, Geoff Ramsey, Jason Saldaña, and Gus ticipate in the production process; (2) the product Sorola in Austin, Texas. Their first and most is virtual meaning that consumption and partici- widely known Machinima show is Red vs. Blue, pation in the virtual online consumer community featuring two teams of soldiers in the game Halo coincide; (3) due to the nature of the product, who are stationed in an isolated canyon where consumer feedback is comparatively easy to inte- their sole purpose is to fight each other. The grate during the production process for further popularity of the show that first aired April 1st, innovation while innovation cycles are short; (4)

73 FROM GAME TO PLAY PETER JÄGER • 2010

2003 stems from the humorous dialogues be- all 730,000 registered members (data from Sep- tween the different characters from each team. tember 2008). While the comedy was initially aimed at other Especially the latest season of Red vs. Blue “Re- gamers playing Halo, Red vs. Blue swiftly became construction” is most interesting to observe since a show that was appreciated by a much broader Rooster Teeth has introduced several major audience. To date, Rooster Teeth has released changes with its release. First and foremost, the five seasons of Red vs. Blue ‘The Blood Gulch storyline of the episodes so far is interrupted and Chronicles,’ and one season of Red vs. Blue ‘Re- taken to a new level. The narrative changed from construction,’ comprising 20 to 25 episodes each humor to drama which has taken the content as well as several spin-off mini-series. Over the further away from the game and, thus, made the years, shooting the films advanced from the game story accessible to an audience of non-gamers. Halo 1 to the latest release Halo 3 running on an Second, new characters are introduced, while a xBox 360 with overwhelming new possibilities in couple of old leading actors dropped out or were graphics and artistic composition. Most of their not introduced until several chapters had been merchandising articles are related to the Red vs. released. Third, the game Halo 3 with extensive Blue show besides which Rooster Teeth produced new production features and a major leap for- several others using different game engines in- ward in graphic quality is applied to film the cluding “The Strangerhood,” “P.A.N.I.C.S.,” or scenes. Those adaptations present a disruptive “1-800-Magic”.

Table 2. Red vs. Blue Community Statistics

Each series has its own website showing the epi- change in the product continuity. sodes, important announcements from Rooster These changes initiated diverse feedback from Teeth staff members, and offering fans the op- members of the brand community, especially portunity to engage in discussing topics around very early on. Hence, Red vs. Blue Reconstruc- the show. The fact that the discussion of the epi- tion serves as the basis for our grounded theory sodes coincides with the place where they are approach since the aforementioned changes viewed – especially while waiting for the down- promised a variety of feedback comments. In- load to finish or directly after watching them cluding the trailer, the 14 chapters cover almost online – makes this case highly valuable. Users 16,000 comments from nearly 6,000 users over a are not required to actively go to a website to period of 120 days (see Table 2 and Table 3). convey their thoughts like on almost all brand Since members tend to mainly post and read community websites – they are actually right comments on a chapter up to the point till the there before, during, and after consumption of next is released, we concentrated on that period the respective product. of usually one week thereby reducing the number Due to high customer loyalty and a steadily grow- of effective comments from 15,658 to 12,674 (see ing fan base, the Red vs. Blue community that is Table 3). After coding the comments of the first actively participating on the website channels eight chapters, we experienced that a vast major- amounts to 42,000 members who have posted ity of comments contained less than five to ten more than 400,000 comments on 165 episodes words, which we considered not enough to con- over the last three years (see Table 2). Consider- vey meaningful feedback. Thus, we set a thresh- ing the entire community, these 42,000 active old of 60 characters per comment (for more de- participants are only a small fraction of the over-

74 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

beyond the operation of the Rooster Teeth website. During phase three, extensive data consistency checks were performed and several descriptive statistics were prepared to ensure a high level of data quality and in- Table 3. Comments for Reconstruction tegrity. This analysis led to the finding that tailed information see data gathering) to filter for 42% of the users commenting on Reconstruction intelligible comments. In the end, our sample signed up after the trailer had been released contained 6,820 comments from 2,168 distinct which implies a heterogeneous sample of existing users (see Table 2 and Table 3). and new users, thus different levels of familiarity with the show and the community. 3.2 Data Gathering Data gathering consisted of three phases includ- 3.3 Data analysis ing both real-time observations and retrospective We derive the common coding scheme from the data (Brown & Eisenhardt, 1997; Jorgensen, collected comments while at the same time hav- 1989). Phase one covered several tasks conducted ing our prior work in the field of Machinima in for an earlier research project on Machinima. mind and the additional insights gained during These included desk research to build a common the data-gathering phase (see Table 4). Before understanding of the Machinima phenomenon coding the comments, one researcher reviewed all (for the use of rich information sources, see 134 Red vs. Blue videos to understand the Vaughan, 1990), entering the community during storyline as well as references to old episodes a four-day workshop for user community mem- made during the new season. The respective bers held in Amsterdam, and several rounds of chapters of Reconstruction were watched again phone and face-to-face interviews. One re- immediately before the coding was conducted. searcher visited members of Machinima companies in the US, including Rooster Teeth Produc- In the next step, a qualitative analysis of the tions. This field trip included participant observa- feedback comments was conducted. While read- tion and ten face-to-face interviews (see also ing through the comments, a code related to the Brown & Eisenhardt, 1997; Jorgensen, 1989). statement was assigned to each dataset. Com- ments that appeared to be highly valuable were During phase two in September 2008, all data marked as top feedback. Top feedback was classi- from the Rooster Teeth Red vs. Blue website con- fied based on information gathered during the 32 cerning the chapters, the users , and the com- coding such as relation to other comments, ex- ments was fetched and transferred to a local da- tremely meaningful feedback to the producers in tabase for further analysis. Storing the data lo- terms of e.g. proposed storyline changes that cally for coding also guaranteed data availability were applied later on, comments that started a discussion in the community, or comments that 32 We obtained detailed information on the users’ personal led to changes in existing episodes. (compare background by capturing the profile information as well as Park, Lee & Han, 2007 for the classification of statistical information about sign-up date, online time, last login, sponsor status, karmalevel, and several more fields feedback quality) Coding in categories allowed that allow insights into the users’ personal preferences and for later sorting and evaluating the information interest, for future evaluation. based on the individual codes. Coding of the first

75 FROM GAME TO PLAY PETER JÄGER • 2010

Table 4. Coding Scheme hundred comments of the Trailer and Chapter 1 During the coding of the first eight chapters, a to 3 was conducted by two independent re- vast number of comments were classified as no searches working in parallel and the results were feedback since they contained less than five to ten checked against each other. Checking for the words or consisted mainly of emoticons. Ac- individual codes, roughly two thirds matched – counting for this, a threshold33 of 60 characters when looking at the aggregated level of special- per comment avoided coding of noise. This left ized product feedback, general, or no feedback 4,033 (55%) comments for the first eight chapters (see Table 4 last column, Table 5) almost all cor- or 6,820 (54%) comments altogether (see Table responded. Applying an iterative process (Eisen- 3). hardt, 1989; Glaser & Strauss, 1967), the coding Next, the codes were evaluated statistically started with seven different codes and got thereby checking for the individual codes as well amended by another three later on to achieve as the aggregated categories on the level of better accuracy of discrimination in the aggre- comments and users. Only 681 of the 2,168 users gated categories of general and no feedback. If posted more than one specialized comment (see Table 5). Moving from data to constructs by evaluating the qualitative information of the top-rated comments as well as the quantitative data based on the coding statistics, we derived five propositions that show how members of a community of consumption contribute to product innovation. To enhance construct validity, all propositions were related to and discussed in the light of existing literature.

Table 5. Comments contributed by unique users necessary, episodes from prior seasons were revis- 33 Since the coding already covered more than 7,000 coded ited or additional data from the forum or from comments, checking at which threshold what number of comments from which category would be omitted was later comments were inspected to verify the valid- feasible. This analysis led to a value of more than 60 ity of arguments in the comments, thereby trian- characters per comment thereby dropping 25% of gulating critical data. specialized feedback, 57% of general feedback, and 62% of no feedback comments but less than 3% of the top rated comments.

76 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

4. TOWARDS A MODEL OF INNOVATION RELATED CONSUMER FEEDBACK IN ONLINE COM- MUNITIES

Analyzing the comments made on a season of Storyline feedback are comments which ask for animated episodes airing online between June changes in the storyline or which approve the and September 2008, first, a distinction between current development of the storyline. While innovation related feedback and noise can be character and storyline feedback deal with the drawn and, second, a categorization of speciali- content, respectively the writing of the episodes, zation within the qualified feedback. The exten- technical feedback and feedback on flaws cover sive data from the Rooster Teeth community comments on the making, respectively the direct- shows that members of a community of con- ing of the show. Detailed descriptions of the indi- sumption do offer specialized and product and vidual feedback categories are presented in the process feedback targeted at the producers inno- remainder of this chapter. vations without being directly stimulated to do. Technical feedback mostly recognizes and appre- We first present an overview of the specialization ciates the game-related technical aspects that are categories illustrated with corresponding com- used to produce a show. Both this and the direct- ments and, second, three incidents that demon- ing skills are honored in the following comment: strate the scope of feedback activity, both in terms of the knowledge required to express the “Man, you guys/gals are doing an awesome job! I like a feedback and in terms of the impact the feedback lot of the little touches you put in, like the hog jumping can generate for product development. over the camera and leaving dirt on the lens. I played RvB Episode 1 and wow, what a difference. I remember you talking about doing that opening shot by using the 4.1 Feedback Falls into Specialized Categories tank gun as a camera crane. You are really doing an awesome job on the cinematography. Keep up the great work!” Members of the community frequently tend to (cbetts (id 1134) comment #593 Chapter 13) comment on individual characters of the show as illustrated by the following comment from chap- Giving hints on how to use specific features of the ter 10: Halo game engine can be experienced as well although this seldom occurs (see the next chapter “I like Griff and I'm proud of his new rank but he for details on this). The last category regarded as CAN'T BE SARGENT! It wouldn't "fit" well with his specialized feedback is comments written on roll [sic]. […] Griff is a lazy slob who dosn't [sic] care flaws, which are the most rarely given of all what anyone else thinks about him. That's why we love comment types. The following quote mentions a him!” (LordVencar (id 682384) comment #386 Chap- minor visual discontinuity error: ter 10) “Sorry, great episode as usual, but I got to point out a Character feedback comprises all feedback relat- countinuity [sic] error. In some of the close-ups with ing to one character or a defined group of char- Sarge, He's weilding a Battle Rifle instead of a shotgun, acters. This can either be commenting on the which he was weilding [sic] seconds earlier. I didn't even absence of a specific character due to a storyline notice the first couple times.” (Lim_Arcadia (id shift or on the personal traits of a character. 690356) comment #368 Chapter 10) The category of storyline feedback is closely tied Flaws, inconsistencies or storyline discontinuities to character feedback. The shift from comedy to uncover editing, plot, or visual errors in the show. drama in Red vs. Blue Reconstruction triggered Some of them are of minor relevance while oth- significant storyline feedback during the first ers lead to changes later applied to the episode chapters when people asked to switch back to the (see chapter technical feedback and flaws on this). old style of humor as can be seen in the comment What is interesting about Lim_Arcadia’s com- below: ment above is that the user had watched the same “I don't think its good at all, why kill the comedy. If it’s chapter several times to actually recognize the not funny it’s not worth watching” (jakesh (id 602887) flaw, indicating a high level of interest in the comment #94 Reconstruction Trailer) technical details of the show (understood as one kind of specialization).

77 FROM GAME TO PLAY PETER JÄGER • 2010

Table 6. Comments per category

Table 6 shows how the detailed codes and the tures; we take requests from them on new features, and aggregated categories amount to the total of things like that. We try to keep the site current, techno- 6,820 coded comments. Considering all 12,674 logically speaking, and asides from that, that’s our whole comments posted in the period between two job.” (Burnie Burns, Rooster Teeth Productions) chapters, about 45% can be considered meaning- The interview statement shows, that the interest ful feedback, whereas less than 30% are actually of Rooster Teeth in their firm hosted online specialized feedback with the potential to lead to community is primarily of marketing and con- innovation. Out of the ten possible codes (see sumer relations kind. Thus, making it a commu- Table 4), four subsume under the category spe- nity of consumption without active firm involve- cialized feedback. These are character (12%), ment such as task setting. storyline (13%), technical (4%), and flaw (0.5%). Considering the different categories to which Rooster Teeth staff members posted as few as five consumers post their comments as well as the comments, when including the moderators and attitude of the sponsoring firm, we propose that site admins a total of 16 comments were posted (1): Members of communities of consump- by members related to the firm. None of the five tion generate specialized innovation- comments referred to any innovation-relevant relevant34 feedback (that can be distin- feedback written by a community member. This guished from noise) without active firm indicates, that the producers do not show any involvement. direct presence by posting comments. Burnie Burns confirmed Rooster Teeth’s passive community involvement during an interview: 4.2 Commenting Activity Varies Across Mem- “they built us the community site. After that it took on a bers life of its own, it really did. […] in order to keep them According to the comment statistics (Table 5), not [the members] interested what we do, is we make the all members tend to show the same level of activ- content at the top of the page and that draws in more ity and dedication to meaningful feedback. 62% people, and occasionally if people come in they watch a of all active members do not contribute any in- video, they leave, they come in, they leave, they come in, novation-relevant comment – instead generating and suddenly they stay. […] That’s all that we really noise (according to an innovation perspective). add into that process of the community, is occasionally Only 12% of the community members contrib- we are the thing that draws people in […] those new ute two or more specialized comments while less people come in to the community, they meet the other peo- than 3% contribute more than ten comments. ple then that’s it. That’s our whole purpose in that. The rest of the community is all theirs; they go crazy with it as they want. Now we have to police sometimes, because, 34 We provide evidence for the innovation-relevance later on you know people on the internet, they’re anonymous and in that we show how the sponsoring firm included the everything they can do with that. We give them new fea- feedback into their products (see next propositions).

78 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

The high diversity in amounts of comments ing amount of comments asking for an immedi- posted per member leads us to propose (2): Dis- ate change back to the old style of comedy: tinguishing meaningful feedback from “I think they should keep the entire cast. Each character noise, we can expect a small but signifi- complements each other very well. If they brought back cant number of members of communities just Caboose it wouldn't be as funny if they didn't bring of consumption to contribute content with Church back as well to yell at him all the time. And you potential innovation-relevance more than can't have Sarge without Grif lol. Plus RVB is too good once. at being funny to become totally serious.” (AIUnknown A more general observation points to the high (id 675796) comment #27 Reconstruction Trailer) variance in commenting behavior among members of communities of consumption. In her comment, AIUnknown reflects on the history of the series to propose a change for the continuing storyline. The requested storyline 4.3 Storyline Feedback for Product Innovation change regarding the characters thereby resembles a product alteration to return to the familiar The overwhelming amount of feedback (13%) is whereas the statement on the funny vs. serious given for the storyline, which shows the storyline switch refers to more general values attached to as the foundation of the show attracts the most the brand. consumer interest. Regardless of the quality of Chapter 2 then made a sudden turnaround and the feedback – whether it’s just praising the show, reintroduced some of the old characters as well as disliking it, or giving constructive feedback – the the old kind of humorous dialogues. The subse- sheer amount of comments points out how the quent chapters kept balancing humor and drama, community of users perceives an episode. Burnie while continuously bringing back the old cast. Burns, the lead writer and director of Red vs. Over time, users seemed to adapt to the new style Blue at Rooster Teeth, illustrated the importance of the show. While the amount of comments of the community feedback for the development demanding the old storyline back steadily de- of the storyline during an interview: creased, more and more comments contributed “Yeah, well, the community site does help and we can tell to the further development of the actual storyline not only by when they post something. When they don’t with its dramatic pieces and the integration of post about something that could be a joke that we really the old characters and their distinct kind of hu- like and no one is that interested in it or like say you put mor. This can be seen in the following comments: up an episode and people will say if they don’t like it, they don’t care about hurting your feelings, they would “I really love that this series is back and the direction that really say a lot, and that’s okay. And you know, its okay, it is going. I felt that near the end of the first 5 seasons, being indifferent to something is really one of the worst the story was actually getting too good for the ridiculous- things that can possibly happen. I mean I rather have ness that usually happens in the show. Frankly, this new someone react very positively or very negatively to some- story driven Reconstruction is awesome. It feels like our thing cause that is a reaction, the last thing you want is beloved characters are thrown into an action no reaction. So if we put up an episode and it only gets drama/thriller and the end result is classic. At the begin- like, I don’t know, 200 comments where people just ning, I never thought that this series would ever grip my aren’t even interested in commenting on that, then we attention with its story and actually have me waiting to know that’s a problem and we have to sit back and say see what happens next, but it has.” (SiLenT366 (id okay what is that. So sometimes no feedback is like the 683734) comment #904 Chapter 4) strongest feedback of all.” (Burnie Burns, Rooster Teeth “This is possibly the best series created by man. Even an Productions) improvement from the original RvB. The storyline is so In order to write sophisticated comments on the deep, it makes you actually start to know the characters, storyline, a user has to either have experience in hwo [sic] they feel, how they would react, their pasts, etc. writing and directing or a very high knowledge of The comedy is there as well. Personally, I would have to say that Church is my favorite character for comic relief the previous storyline of Red vs. Blue. What is in this series, unless Doc comes back, then he takes over. very interesting in this case is that both the Overall, I don't mind waiting for the next episode for Trailer for Reconstruction as well as Chapter 1 about a week, because i [sic] know that these series are foreshadowed a dramatic storyline rather than a actually written with care, a good structure. RT has ma- humorous one as could have been expected from tured alot [sic] over these past few years. It shows in these the previous seasons. This led to an overwhelm- episodes. Maturity in the sense of creating good structured

79 FROM GAME TO PLAY PETER JÄGER • 2010

episodes. The immature jokes are still there. But I really in their comments, we propose (4a): Product don't mind. :D” (KlassyReborn (id 280924) comment and external knowledge in combination #799 Chapter 6) enables members of communities of consumption to formulate specialized feed- The last comment shows how feedback on the back on the production process. storyline partly overlaps with feedback on the characters. Both require a deep understanding (4b): Specialized feedback on the produc- and constant observation of the show as well as tion process may lead the sponsoring firm the forum activities where Rooster Teeth mem- to process innovation. bers comment from time to time on current development issues. We thus propose (3a): Con- stantly monitoring the progress of a prod- 4.5 Feedback on Flaws for Quality Assurance uct enables members of communities of Flaws were mentioned rather infrequently by consumption to provide detailed feedback users but if mentioned were sometimes even the on the product’s possible continuity cause for a change to the episode. This shows the (thereby referring to the product’s his- most direct and immediate sort of feedback gen- tory). erated by the community where a sudden change (3b): Specialized feedback on the product’s indicates a response by Rooster Teeth which continuity may lead the sponsoring firm shows their continuous interest in the comments. to benefit for product innovation. Burnie Burns described an incident when back in Season 5 a sponsor user mentioned a problem in an episode that triggered a change: 4.4 Technical Feedback for Process Innovation “We just had an episode where somebody pointed out to Technical feedback largely deals with the possi- us that there was an inconsistency with the storyline that bilities that the game engine offers. While seasons was in the episode, and that was the sponsor’s point that one to five of Red vs. Blue “The Blood Gulch we released on Friday and so between Friday and when Chronicles” were set in Halo 1 and Halo 2, ‘Re- we released it on Monday to the public we fixed the prob- construction’ makes use of the latest version of lem, we just … all we had to do was make one character the game, Halo 3. The latter now incorporates say the line versus another character. Very simple change, Forge, which is a kind of a level editor that gives it wasn’t a huge deal, but it was big enough to where we the producer of the Machinima film much more wanted to make the change, and it just goes to show how freedom when deciding on the scenery and the on the Internet you get instant feedback and they help us character prop interaction. Another feature is a produce the episode.” (Burnie Burns, Rooster Teeth Pro- change in the use of the cloaking equipment to ductions) turn invisible. Halo 3 added a new mode of ac- Reconstruction Chapter 8 showed an identical tive camouflage to have the user appear either incident where several users pointed out an in- ghostly or glass-like. This is highly relevant for the visible person standing next to an unconscious Red vs. Blue production, since one character can character in one of the key scenes of the season. appear as a ghost. The following comment shows This immediately led to rampant discussions in how a user suggested another way of rendering the comments about the further development of the actor partly invisible as well as a comment the storyline: two chapters later indicating that Rooster Teeth had applied the new way of production. “Did anyone else see the glimmer at the top of the screen at 4:35ish?As Maine walks up to Caboose, it shows “Ghost church [Church is the name of one of the main Maine's feet. After that, it flips angles. When it does, characters of the show] must be so much easier to do in watch the top of the screen, there is a very clear invisible halo 3, what with the weak camo. Did Bungie do that glimmer that takes a step to the right, moving behind the just for you guys?” (Xtremephoen (id 592100) comment metal. I suspect this is cannon, because it's not going to #512 Chapter 11) be a camera character or anything, as there are saved “I noticed that you took that guy’s advice and made films which eliminate that need. Accidental invisible church a bad cloak.” (Eschex (id 690399) comment character moving in the shot? Doubtful, in my eyes.” #460 Chapter 13) Tural (id 47055) comment #50 Chapter 8 Given the consumers’ background and knowl- “That's why I don't see the shimmer!!!!!! I was wonder- edge about video games, which they incorporate ing what everyone was talking about but I didnt [sic] get

80 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION to watch it until the videos came back online! The video For some flaws, members state to have re- is also called Hi-Res FIX. Maybe that shimmer wasn't watched the episode several times before discov- meant to be there.” EatMoreRame (id 290720) com- ering it. Considering the effect of feedback on ment #806 Chapter 8 flaws we propose (5a): Observing the product “Hopefully this will clear up the "invisible [sic] person" thoroughly due to an affection with the situation. Yes there was "cloaked feet standing nearby product leads members of communities Caboose which moved around the corner but the video of consumption to contribute specialized was changed and reposted with that scene slightly zoomed feedback on flaws and inconsistencies. in. Looks to be a mistake, probably someone to knock (5b): Specialized feedback on flaws and Caboose out or set his fallen body. Post edited 7/23/08 inconsistencies helps the sponsoring firm 1:21AM” Der1cho (id 607024) comment #899 to maintain the continuous quality of the Chapter 8 product. Altogether, the data shows that some members of communities of consumption offer specialized After about 500 comments, first users stated that innovation-relevant feedback while others pro- the episode was no longer accessible. Later on, duce primarily noise with respect to innovation. users mentioned that the observed shimmer The noise however, may be important for their originating from the assumed invisible person had own personal needs and interests. Separating the vanished and, apparently, Rooster Teeth had specialized comments, a signal to noise ratio of uploaded a new version (comment #806). An- less than 30% significant comments are observ- other user then tried to explain the reason for the able (see also Table 6). Sponsors, representing phenomenon to their peers in the community 32% of the users writing specialized feedback, (comment #899). This again shows how mem- contributed 35% of the feedback comments. bers not only give specialized feedback to the Thus, the fact that members pay a subscription producers of the series in their dialogues (who fee does not result in an above average level of may notice the feedback and implement changes) activity in the community. but also to their fellow community members.

81 FROM GAME TO PLAY PETER JÄGER • 2010

5. DISCUSSION

In the following, we discuss how each proposition withdraw or change communication behavior in our model contributes to existing knowledge (Jones, Ravid, & Rafaeli, 2004). While withdrawal on online communities in innovation. means ending communication, a change in be- Proposition 1. Our results show that consumers havior could include two strategies: responding to give unsolicited, specialized feedback on new and and responding with simpler messages (Jones et existing products. They target producers with al., 2004) or specialization (Hiltz & Turoff, 1985). their feedback and the blurring of the boundary The latter response means introducing individual between communities of consumption and com- filters and structures to control content that help munities of production can be affirmed since reduce the amount of cognitive resources needed consumers’ feedback represents an important to process the information (March, 1978; Simon, step towards learning and, thus, productivity. 1955). These findings contribute to the emerging litera- Kozinets et al. (2008) introduce the concept of ture on online communities (Kozinets et al., 2008; hierarchies of expertise when talking about col- Rowley et al., 2007) in that they present empirical lective content creation in online communities: evidence and lead to a number of future research “collective creativity develops and produces ‘con- questions while the study also shows several limi- tent’ that exceeds the value that could be pro- tations. vided by single individuals.” (Kozinets et al., A community of consumption that provides spe- 2008: 343). Content creation is limited to an es- cialized, innovation-relevant feedback blurs the timated one percent of all site users (McConnell boundaries usually drawn in the literature be- & Huba, 2006), which we can confirm with our tween communities of consumption and commu- data. nities of production. Members of communities Proposition 3-5. Specialized feedback takes the of consumption explore specific lifestyles and form of a contribution towards the value created express themselves and values connected to the by the online product from the perspective of products consumed (Kozinets, 2002; McAlexan- value coproduction (Ramirez, 1999). The Rooster der et al., 2002; Muniz & O'Guinn, 2001) and Teeth community would be considered a swarm therefore appear to be an unlikely place to tackle given the Web 2.0 setting. “Within […] Swarms, problems and tasks related to innovation (espe- the value-added of most individual contributions cially if these tasks are not predefined or pre- may be quite low, but the aggregate value of the assigned). However, the community members high collective quantity and quality of contribu- who make up the audience of the Machinima tions will be very high” (Kozinets et al., 2008: films in this sample engage in terms of giving 350) The consumer adds their domain knowledge specialized feedback to the producers, although from e.g. gaming or film production to the collec- tasks are not coordinated or pre-given. Only the tive discussion and consumption of the films and, large-scale investigation of comments could trace thus, creates value. The feedback is directly these emergent forms of specialization: we con- linked with the product through inseparable web tribute to the literature a precise analysis of the presence (on the same page), forms part of the content of the feedback and an empirically consumption experience and translates creative grounded categorization relevant to product and ideas into product or process-relevant innova- process innovation. tions. From this perspective, the community Proposition 2. Large online communities can gen- member steps over the line that, traditionally, erate a level of communication activities that separates the consumer from the producer and leads to information overload for individual community types might be in need of reconsid- community members (Hiltz & Turoff, 1985). In- eration. Four implications for future research in dividuals must read a large amount of messages, innovation derive from this observation, which we understand the context and content of these, and discuss in the next section, generate responses. Research in large, online The innovation aspect of communities of con- interaction spaces has identified two basic options sumption has been neglected so far in terms of for individuals who face information overload: empirical evidence for innovative output.

82 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

Sawheny and Prandelli (2000) aimed at closing firms need to extract the meaningful information, the gap by introducing the concept of communi- make sense of it in terms of knowledge genera- ties of creation. Even though being oriented tion and evaluate its innovation relevance. The more towards communities of consumption, captured knowledge might eventually be applied members of communities of creation are pro- to increase the products’ performance or design vided with tasks which offer the members a con- entirely new products. text to generate ideas. Without this context given,

83 FROM GAME TO PLAY PETER JÄGER • 2010

6. CONCLUSIONS, IMPLICATIONS FOR RESEARCH AND MANAGEMENT

This research was motivated by a lack of empiri- on motivation or deterrence and intimidation of cal work on the feedback activity in communities members by the member-to-member comment of consumption, particularly the content of the rating system, especially when entering the com- feedback and the relevance for innovation. We munity, is of interest. In addition, the users coor- collected over 12,000 unsolicited comments by dinate among themselves using features of the users and coded almost 7,000 comments indi- social software platform to develop a common vidually to categorize and evaluate them in terms understanding of the product they desire, which of their potential to generate process and product may eventually overlap with the company’s adap- innovation. The results show that members of the tations of a product based on the users’ sugges- community of consumption studied generated tions. Future research should focus on social soft- specialized, unsolicited feedback, and 10% of the ware for consensus building and interaction in members contribute more than once. The de- communities of consumption and creation. tailed knowledge of the product’s history allowed Third, the distinction between communities of users to comment on the possible continuity of production and communities of consumption product development, which may result in prod- should be revisited and refined. We presented a uct innovation. The combination of technical first impression what firms can expect from expertise and knowledge about the product pre- communities of consumption. When does a shift pared users to generate feedback relevant for in cognitive labor emerge that triggers specialized process innovation and an apparent interest in feedback and problem solving by community detail supported the sponsoring firm’s quality members? How does the presence of company assurance. members impact the activity of the community? These results lead to a number of questions for Two important aspects to consider here are the future research. First, future research on service task assignment and whether the community is innovation should consider the role of online left unattended or managed by a sponsor or host- communities of consumption since specialization ing firm. in comments and feedback link the provider with Our work also relates to research on open innova- the customer: how can an online product be dis- tion, which encompasses knowledge co-creation tinguished from an online service if consumers’ of firms and community members. Pisano and feedback creates part of the product value? A Verganti (2008) discuss advantages of firms that waning distinction between online products and manage to leverage a network of outside partici- online services has far-reaching consequences for pants in innovation. However, discussions about the marketing and development of online prod- the types of governance applicable to communi- ucts and services, such as customized media (on ties need to take a step back and the tasks and demand). activities community members adopt and solve Second, social software features a number of should be considered first. The interaction of filtering, rating, and interaction tools that enable governance (decision rules, institutional context, community members to build and maintain rela- membership criteria, etc.) and community activ- tions online (Pascu, Osimo, Ulbrich, Turlea, & ity deserves close attention in future research Burgelman 2007). The connection between social because only few community members are very software and innovation processes is vague. How active. can social software impact innovation practices The current study suffers from a few limitations. on the level of the innovation process? The find- Specialization could not be measured on the level ings of this study suggest that the role of a com- of individuals as intended at first but had to be munity of consumption and the use community evaluate on the level of comments. Specialization members make of social software features de- is sometimes understood in relative terms, that is, serves close attention by innovation scholars. Be- how much specialized work an individual per- cause community members provide valuable forms in relationship to non-specialized work. For feedback, sorting noise from feedback and ena- this study, the amount of comments per member bling user-generated rankings and ratings be- was so small that the discussion of a ratio was comes a managerial priority. A longitudinal study

84 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION pointless. Future research should define individ- communities, firms need to pay special attention ual specialization in order to characterize indi- to several aspects that influence the later per- vidual learning over time. For this, a longitudinal formance of the members and the community’s study of individual users seems to be most appro- intended output. While the design of online priate. Findings on member specialization might communities is fundamental to the members’ also yield insights, which help developing intelli- attachment, loyalty, and commitment to the gent filtering mechanisms for innovation-relevant community, the members strongly influence the feedback. organization structure and evolution of the Observing and coding a mature community of community (Ren, Kraut, & Kielser, 2007). consumers limits the implications for manage- Whether innovation-relevant activities can be ment practice in that it focuses on a successful designed into an online community remains un- example. Comparing the Rooster Teeth commu- clear. nity with a less successful online community may Managers may want to experiment with software reveal more about the self-enforcing dynamics of tools that allow the filtering of noise from feed- a growing community. In addition, extended cod- back. The analysis suggests that character limits ing within the current data could lead to a com- may be a simple automatic attempt but that the parison with the starting phase of Rooster Teeth value of the contributions lies in their detailed when the online infrastructure was first made accounts that require reading and understanding. available to the public. Then again, the fact that On the upside, a few very active community 42% of the contributing members signed up dur- members submit a high number of specialized ing the season also shows the integration of less comments. Monitoring these members may pro- experienced consumers. vide high-quality feedback. Separating communi- A first takeaway for managers is to expect more ties of production from communities of con- from your community of consumption than just sumption may, thus, not be practically feasible pure consumption. Community members may and possibly not necessary because members of get involved beyond sharing values and self- online communities tend to show a deep interest expression to offer specialized feedback that re- in the products and, if given the opportunity, will quires a deep understanding of the history and offer valuable feedback that falls into specialized intricacies of the product. When creating online categories.

85 FROM GAME TO PLAY PETER JÄGER • 2010

7. ACKNOWLEDGEMENTS

The authors thank our interview partners and the well as the participants at the Morteratsch Re- community of Machinimators for their time and search Seminar for their valuable feedback. This valuable input that made this research possible, project received support from the Swiss National Philip Reichen, Richard Einstein, and Florian Science Foundation grant number: 105512- Egloff for their excellent research assistance, as 106932.

86 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

APPENDIX

Figure A1. The Community of Creation as an Overlap between a Community of Production and a Community of Con- sumption (Sawhney & Prandelli, 2000: 48)

Figure A2. Typology of Online Creative Consumer Communities (Kozinets et al., 2008: 345)

Figure A3. Extended Typology of Communities considering Task Setting

87 FROM GAME TO PLAY PETER JÄGER • 2010

REFERENCES

Algesheimer, R., Dholakia, U. M., & Herrmann, A. 2005. Eisenhardt, K.M. 1989. Building Theories from Case Study The social influence of brand community: Evidence from Research. Academy of Management Review. 14(4): 532-550. European car clubs. Journal of Marketing. 69(3): 19-34. Eisenhardt, K.M. & Graebner, M.E. 2007. Theory Building Amabile, T., Conti, R., Coon, H., Lazenby, J. & Herron, M. from Cases: Opportunities and Challenges. Academy of Man- 1996. Assessing the Work Environment for Creativity. Acad- agement Journal. 50(1): 25-32. emy of Management Journal. 39(5): 1154-1184. Franke, N. & von Hippel, E. 2003. Satisfying heterogeneous Amabile, T. 1998. How to Kill Creativity. Harvard Business user needs via innovation toolkits: The case of apache secu- Review. 76(5): 76-87. rity software. Research Policy. 32(7): 1199-1215. Amblee, N. & Bui, T. 2008. Can Brand Reputation Improve Franke, N. and Shah, S. 2003. How communities support the Odds of Being Reviewed On-Line? International Journal of innovative activities: An exploration of assistance and shar- Electronic Commerce. 12(3): 11-28. ing among end-users. Research Policy. 32(1): 157-178. Anonymous. 1998. To Reveal or Not to Reveal. A Theoreti- Franke, N., Keinz, P. & Schreier, M. 2008. Complementing cal Model of Anonymous Communication. Communication Mass Customization Toolkits with User Communities: How Theory. 8(4): 381-407. Peer Input Improves Customer Self-Design. Journal of Product Innovation Management. 25(): 546-559. Bagozzi, R.P. & Dholakia, U.M. 2006) Antecedents and purchase consequences of customer participation in small Füller, J., Jawecki, G., & Mühlbacher, H. 2007. Innovation group brand communities. International Journal of Research in Creation by Online Basketball Communities. Journal of Marketing. 23(): 45-61. Business Research. 60(1): 60-71. Baldwin, C., Hienerth, C., & von Hippel, E. 2006. How user Füller, J., Matzler, K., Hoppe, M. 2008. Brand Community innovations become commercial products: A theoretical Members as a Source of Innovation. Journal of Product Innova- investigation and case study. Research Policy. 35(9): 1291-1313. tion Management. 25: 608-619. Belk, R.W. 1988. Possessions and the Extended Self. Journal Gächter, S., von Krogh, G., Haefliger, S. 2010. Initiating of Consumer Research. 15(2): 139-168. private-collective innovation: The fragility of knowledge sharing. Research Policy. Forthcoming. Berry, L.L. 1995. Relationship marketing of services: Grow- ing interest, emerging perspectives. Journal of the Academy of Glaser, B. & Straus, A.L. 1967. The Discovery of Grounded Marketing Science. 23(4): 236-245. Theory. Chicago, IL: Aldine. Brown, S.L. & Eisenhardt, K.M. 1997. The art of continu- Glaser B.G. 1998. Doing Grounded Theory - Issues and Discus- ous change: Linking complexity theory and time-paced sions. Sociology Press. evolution in relentlessly shifting organizations. Administrative Science Quarterly. 42: 1-34. Goffman, E. 1959. The Presentation of Self in Everyday Life. New York, Doubleday. Brown, S., Kozinets, R.V. & Sherry, J.F. Jr. 2003. Teaching Old Brands New Tricks: Retro Branding and the Revival of Granitz, N.A. & Ward, J.C. 1996. Virtual Community: A Brand Meaning. Journal of Marketing. 67(): 19-33. Sociocognitive Analysis. In Advances in Consumer Research, Vol 23. Eds: Corfman, K.P. & Lynch, J.G. Provo, UT: Associa- Carlson, B.D., Suter, T.A., & Brown, T.J. 2007. Social versus tion for Consumer Research. 161-166. psychological brand community: The role of psychological sense of brand community. Journal of Business Research. 61(4): Gruen, T.W., Osmonbekov, T. & Czaplewski, A.J. 2006. 284-291. eWOM: The impact of customer-to-customer online know- how exchange on customer value and loyalty. Journal of Cohen, B.P. 1980. Developing Sociological Knowledge: Theory and Business Research. 59(): 449-456. Method. Englewood Cliffs, NJ: Prentice Hall. Gurteen, D. 1998. Knowledge, Creativity and Innovation. Cova, B. 1997. Community and Consumption. Towards a Journal of Knowledge Management. 2(1): 5-13. Definition of the “Linking Value” of Product or Services. European Management Journal. 31(3/4): 297-316. Haefliger, S., von Krogh, G., & Spaeth, S. 2008. Code Reuse in Open Source Software Development. Management Science. Cova, B. & Pace, S. 2006. Brand community of convenience 54(1): 180-193. products: new forms of customer empowerment - the case "my Nutella The Community". European Journal of Marketing. Haefliger, S., Reichen, P., Jaeger, P. & von Krogh, G. 2009. 40(9-10): 1087-1105. Modding as rating behavior in virtual communities: The case of Rooster Teeth Productions. Online Communities and Dholakia, U.M., Bagozzi, R.P. & Pearo, L.K. 2004. A social Social Computing, Lecture Notes in Computer Science. 5621(): 197- influence model of consumer participation in network- and 206. small-group-based virtual communities. International Jour- nal of Research in Marketing. 21(): 241-263. Hancock, H. & Ingram, J. 2007. Machinima for Dummies. For Dummies. Dichter, E. 1966. Ho Word-of-Mouth Advertising Works. Harvard Business Review. (Nov-Dec.): 147-166. Hemetsberger, A. 2006. Understanding Consumers’ Collec- tive Action on the Internet: A Conceptualization and Em- pirical Investigation of the Free- and Open-Source Move-

88 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION

ment. Research Synopsis. University of Innsbruck Habilita- Kozinets, R.V. 2002. The Field behind the Screen: Using tion. Netnography for Marketing Research in Online Communi- ties. Journal of Marketing Research. 39(Feb): 61-72. Henning-Thurau, T. & Walsh, G. 2004. Electronic Word-of- Mouth: Motives for and Consequences of Reading Custo- Kozinets, R.V. 2007. Inno-tribes: Star Trek as Wikimedia. in mer Articulations on the Internet. International Journal of Consumer tribes, B. Cova, R.V. Kozinets & A. Shankar, 194- Electronic Commerce. 8(2): 51-74. 211. Oxford and Burlington, MA: Butterworth-Heinemann Herstatt, C. & von Hippel, E. 1992. From Experience: De- Kozinets, R.V., Hemetsberger, A. & Schau, H.J. 2008. The veloping New Product Concepts Via the Lead User Method: Wisdom of Consumer Crowds. Collective Innovation in the A Case Study in a "Low Tech" Field. Journal of Product Inno- Age of Networked Marketing. Journal of Macromarketing. vation Management. 213-221. 28(4): 339-354. Hertel, G., Niedner, S., & Herrmann, S. 2003. Motivation Kozinets, R.V. 2010. Doing Ethnographic Research of software developers in Open Source projects: an Internet- Online. Sage Publications Ltd. based survey of contributors to the Linux kernel. Research Policy. 32: 1159-1177. Lakhani, K. R. & von Hippel, E. 2003. How open source software works: "free" user-to-user assistance. Research Policy, Hienerth, C. 2006. The commercialization of user innova- 32(6): 923-943. tions: the development of the rodeo kayak industry. R&D Management. 36(3): 273-294. Lawrence, T.B., Hardy, C., Phillips, N. 2002. Institutional effects of interorganizational collaboration: The emergence Hienerth, C. & Lettl, C. 2009. Exploring how peer commu- of proto-institutions. Academy of Management Journal. 45(1): nities enable lead user innovations to become the industry 281-290. standard: Community pull effects. Journal of Product Innovation Management. (forthcoming). Lee, G. K. & Cole, R. E. 2003. From a firm-based to a community-based model of knowledge creation: The case of Hiltz, S.R., & Turoff, M. 1985. Structuring Computer- the Linux kernel development. Organization Science. 14(6): Mediated Communication Systems to Avoid Information 633-649. Overload. Communications of the ACM. 28(7): 680-689. Leimeister, J.M., Huber, M., Bretschneider, U. & Krcmar, H. Jang, H., Olfman, L. Ko, I, Koh, J. & Kim, K. 2008. The 2009. Leveraging Crowdsourcing: Activation-Supporting Influence of On-Line Brand Community Characteristics on Components for IT-Based Ideas Competition. Journal of Community Commitment and Brand Loyalty. International Management Information Systems. 26(1): 197-224. Journal of Electronic Commerce. 12(3): 57-80. Lowood, 2007. High-performance play: The making of Janzik, L., Raasch, C. & Herstatt, C. 2009. Online commu- Machinima. In: Clarke, A. & Mitchell, G. (Eds.) Videogames nities in mature markets: Why join, why innovate, why and art. Bristol: Intellect Books. share? Hamburg University of Technology Working Paper. Luedicke , M. 2006. Brand Community Under Fire: The Jeppesen, L.B. & Frederiksen, L. 2006. Why Do Users Con- Role of Social Environments for the HUMMER Brand tribute to Firm-Hosted User Communities? The Case of Community. Advances in Consumer Research. 33(): 486-493. Computer-Controlled Music Instruments. Organization Sci- ence. 17: 45–63. Luthje, C., Herstatt, C., & von Hippel, E. 2005. User- innovators and “local” information: The case of mountain Jones, Q., Ravid, G., & Rafaeli, S. 2004. Information Over- biking. Research Policy. 34: 951-965. load and the Message Dynamics of Online Interaction Spaces: A Theoretical Model and Empirical Exploration. MacCormack, A., Rusnak, J., & Baldwin, C.Y. 2006. Explor- Information Systems Research. 15(2): 194-210. ing the structure of complex software designs: An empirical study of open source and proprietary code. Management Jorgensen, D.L. 1989. Participant observation: a methodol- Science. 52: 1015–1030. ogy for human studies. Newbury Park, CA: Sage. March, J. G. 1978. Bounded Rationality, Ambiguity, and Ke, W. & Zhang, P. 2009. Motivations in Open Source Engineering of Choice. Bell Journal of Economics, 9(2): 587- Software Communities: The Mediating Role of Effort Inten- 608. sity and Goal Commitment. International Journal of Electronic Commerce. 13(4): 39-66. March, J.G., Sproull, L.S., & Tamuz, M. 1991. Learning from Samples of One of Fewer. Organization Science. 2(1): 1- Kozinets, R.V. 1997. ‘I want to believe’: a netnography of 13. the X-philes’ subculture of consumption. In Advances in Consumer Research, M. Brucks & D.J. MacInnis. 24(): 470-475. Marino, P. 2004. 3D game-based filmmaking: The art of Association for Consumer Research, Provo, UT. Machinima. Paraglyph Press Kozinets, R.V. 1999. E-tribalized marketing? The strategic McAlexander, J.H., Schouten, J.W., & Koenig, H.F. 2002. implications of virtual communities of consumption. Euro- Building Brand Communitiy. Journal of Marketing. 66(Jan): pean Management Journal. 17(3): 252-264. 38-54. Kozinets, R.V. 2001. Utopian Enterprise: Articulating the McConnell, B. & Huba, J. 2006. Citizen marketers: when people Meanings of Star Trek’s Culture of Consumption. Journal of are the message. Chicago: Kaplan. Consumer Research. 28: 67-88. McWilliam, G. 2000. Building stronger brands through online communities. Sloan Management Review. 41(3): 43–54.

89 FROM GAME TO PLAY PETER JÄGER • 2010

Morris, D., Kelland, M., & Lloyd, D. 2005. Machinima. Rowley, J., Kupiec-Teahan, B. & Leeming, E. 2007. Cus- Course Technology PTR tomer Community and Co-creation: A Case Study. Marketing Intelligence & Planning. 25(2): 136-146. Muniz, A. M. & O'Guinn, T. C. 2001. Brand community. Journal of Consumer Research. 27(4): 412-432. Sassenberg, K. 2002. Common Bond and Common Identity Groups on the Internet: Attachment and Normative Behav- Nambisan, S. 2002. Designing Virtual Customer Environ- ior in On-Topic and Off-Topic Chats. Group Dynamics: Theory, ments for New Product Development: Toward a Theory. Research, and Practice. 6(1): 27-37. Academy of Management Review. 27(3): 392-413 Sawhney, M. & Prandelli, E. 2000. Communities of Crea- Nambisan, S. & Baron, R.A. 2007. Interactions in Virtual tion: Managing Distributed Innovation in Turbulent Mar- Customer Environments: Implications for Product Support kets. California Management Review. 42(4): 24-54. and Customer Relationship Management. Journal of Inter- active Marketing. 21(2): 42-62. Schau, H. J. & Gilly, M. C. 2003. We are what we post? Self- presentation in personal Web space. Journal of Consumer Nambisan, S. & Nambisan, S. 2008. How to Profit from a Research. 30(3): 385-404. Better ‘Virtual Customer Environment’. MIT Sloan Man- agement Review. 49(3): 53-61. Schmidt, J. 2007. Blogging practices: An analytical framework. Journal of Computer-Mediated Communication, 12(4): article Nelson, M.R. & Otnes, C.C. 2005. Exploring cross-cultural 13. http://jcmc.indiana.edu/vol12/issue4/schmidt.html ambivalence: a netnography of intercultural wedding message boards. Journal of Business Research. 58: 89-95. Siggelkow, N. 2007. Persuasion with case studies. Academy of Management Journal. 50(1): 20-24. Ogawa, S. & Piller, F. T. 2006. Reducing the risks of new product development. MIT Sloan Management Review. 47(2): Simon, H.A. 1955. A Behavioral Model of Rational Choice. 65-71. Quarterly Journal of Economics. 69(1): 99-118. O'Mahony, S. & Ferraro, F. 2007. The Emergence of Gov- Spaeth, S., Haefliger, S., von Krogh, G., & Renzl, B. 2008. ernance in an Open Source Community. Academy of Manage- Communal resources in open source software development. ment Journal. 50(5): 1079-1106. Information Research. 13(1). Park, D., Lee, J. & Han, I. 2007. The Effect of On-Line Spaeth, S., Stuermer, M. & von Krogh, G. 2010. Enabling Consumer Reviews on Consumer Purchasing Intention: The knowledge creation through outsiders: towards a push model Moderating Role of Involvement. Interntional Journal of Elec- of open innovation. International Journal of Technology Manage- tronic Commerce. 11(4): 125-148. ment. (forthcoming) Pascu, C., Osimo, D. Ulbrich, M., Turlea, G., & J.C. Bur- Sproull, L., Dutton, W., & Kiesler, S. 2007. Introduction to gelman. 2007. The potential disruptive impact of Internet 2 the Special Issue: Online Communities. Organization Studies. based technologies. First Monday. 12(3). 28(3): 277-281. Pervez, N.G., Tarnovskaya, V. & Elg, U. 2008. Market driv- Stallman, R. 1999. “The GNU Operating System and the ing multinationals and their global sourcing network. Interna- Free Software Movement,” in Open Sources: Voices from the Open tional Marketing Review. 25(5): 504-519. Source Revolution, M. Stone, C. DiBona, and S. Ockman, (eds.), Sebastopol, CA: O’Reilly and Associates, Inc., pp. 53- Picher, E.A. & Hemetsberger, A. 2007. “Hopelessly Devoted 70 to You” Towards an Extended Conceptualization of Con- sumer Devotion. Advances in Consumer Research. 34: 194-199. Strauss, A.L. & Corbin, J. 1990. Grounded Theory Re- search: Procedures, Canons and Evaluative Criteria. Pichler. E.A. & Hemetsberger, A. 2008. Driven by Devotion Zeitschrift für Soziologie. 19(6): 418 ff. – How consumers Interact with their Objects of Devotion. Advances in Consumer Research. 35: 439-443. Stuermer, M., Spaeth, S. & von Krogh, G. 2009. Extending private-collective innovation: a case study. R&D Management. Pisano, G.P. & Verganti, R. 2008. Which Kind of Collabora- 39(2): 170-191. tion is Right for You? Harvard Business Review. 86(12): 78-86. Terwiesch, C. & Xu, Y. 2008. Innovation contests, open Pitta, D.A. & Fowler, D. 2005. Online Consumer Communi- innovation, and multiagent problem solving. Management ties and their Value to New Product Developers. Journal of Science 54(9) 1529-1543. Product & Brand Management. 14(5): 283-291.

Porter, C.E. 2006. A Typology of Virtual Communities: A Multi-Disciplinary Foundation for Future Research. Journal Vaughan, D. 1990. Autonomy, interdependence, and social of Computer-Mediated Communication. 10(1): 00-00. control: NASA and the space shuttle Challenger. Administra- tive Science Quarterly. 35(2): 225-257. Ramirez, R. 1999. Value Co-Production: Intellectual Ori- gins and Implications for Practice and Research. Strategic von Hippel, E. 1988. The Sources of Innovation. Oxford Uni- Management Journal. 20: 49-65. versity Press. Raymond, E.S. 1999. The Cathedral & the Bazaar. O'Reilly von Hippel, E. 1994. Sticky Information and the locus of problem solving: Implications for innovation. Management Ren, Y.Q., Kraut, R., & Kiesler, S. 2007. Applying common Science. 40(4): 429–439. identity and bond theory to design of online communities. Organization Studies. 28(3): 377-408. von Hippel, E. & von Krogh, G. 2003. Open source software and the "Private-Collective" innovation model: Issues for organization science. Organization Science. 14(2): 209-223.

90 A DIRECTING AUDIENCE • HOW SPECIALIZED FEEDBACK IN A VIRTUAL COMMUNITY OF CONSUMPTION STIMULATES NEW MEDIA PRODUCTION von Hippel, E. 2005. Democratizing innovation. Cambridge. MA: Interviews MIT Press. Burns, Burnie. Rooster Teeth Productions (USA). Face to von Krogh, G., Spaeth, S., & Lakhani, K. R. 2003. Com- face interview on April 3rd, 2007 munity, joining, and specialization in open source software innovation: A case study. Research Policy. 32(7): 1217-1241. Ramsey, Geoff. Rooster Teeth Productions (USA). Face to face interview on April 3rd, 2007 Wasko, M. M. & Faraj, S. 2005. Why should I share? Exam- ining social capital and knowledge contribution in electronic Sorola, Gus. Rooster Teeth Productions (USA). Face to face networks of practice. MIS Quarterly. 29(1): 35-57. interview on April 2nd, 2007 Wiertz, C. & de Ruyter, K. 2007. Beyond the Call of Duty: Lowood, Henry. Stanford University (USA). Telephone Why Customers Contribute to Firm-hosted Commercial interview on March 13th, 2007 Online Communities. Organization Studies. 28(3): 347-376. Hullum, Matt. Rooster Teeth Productions (USA). Face to Yamauchi, Y.; Yokozawa, M.; Shinohara, T. & Ishida, T. face interview on April 4th, 2007 2000. Collaboration with Lean Media: how open-source Marino, Paul. Academy of Machinima Arts & Sciences software succeeds. In: Proceedings of the 2000 ACM confer- (USA). Face to face interview on March 31st, 2007 ence on Computer supported cooperative work, ACM New York, NY, USA. 329–338. Yin, R. K. 2003. Case study research: design and methods. Thou- sand Okas, CA: Sage.

91 FROM GAME TO PLAY PETER JÄGER • 2010

CURIOUS LURKERS

LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

Peter Jäger

Chair of Strategic Management and Innovation Department of Management, Technology, and Economics (MTEC) ETH Zürich, Switzerland

ETH Zürich Working Paper 2010

Online communities can positively affect the economic performance of firms. Knowing about members’ motivations to contribute helps community managers to better specify their offering. Unfortunately, lurkers, the passive members of online communities, dominate at levels of up to 90 per cent. Lurkers can be perceived as a loss since – in contrast to new members – lurkers are already familiarized with the community regarding its content and norms, which in turn assumes lower costs of incentivizing lurkers compared to acquiring outsiders. Lurkers have drawn increasing interest from researchers of various disciplines. While (short-term) motives for lurking were identified in several studies, no research has been conducted on linking the personality traits of lurkers with their actual contribution behavior in online communities. However, we believe that curiosity, the stable emotional tendency that drives individuals’ information seeking behavior, will help to establish a process model of lurkers’ online behavior. In understanding the different stages of the process, we try to explain an individual’s propensity to de-lurking. Generally being an enduring personality trait, the level to which curious behavior is activated varies across individuals resulting in different momentary emotional- motivational states stimulated by various levels of wanting and liking new information. We theorize that curiosity as a feeling of informational deprivation (D-type epistemic curiosity: need for information) can offset “joining costs” and thus accounts for members’ active contribution behavior in online communities while curiosity as a feeling of interest (I-type epistemic curiosity: pleasure of acquiring new knowledge) keeps members engaged in browsing online community content yet does not suffice to engage them in active participation. Empirical research testing our model would stimulate new approaches to incentivizing community members with low levels of D-type trait curiosity to participate while members with low levels of I- type trait curiosity might be persuaded to stay.

93 FROM GAME TO PLAY PETER JÄGER • 2010

1. INTRODUCTION

Members of online communities, their behavior, pleasure of new discoveries. According to their and their motivation and intention to join and study individuals differ in two distinct ways – contribute draw increasing interest from re- their predisposition to become curious and their searchers of various fields (Jeppesen and tendency to experience curiosity as feelings of Frederiksen, 2006; Moore and Serva, 2007; interest or deprivation. Koo and Choi (2010) Rheingold, 1993; Tafjel, 1978). Retaining existing found I-type and D-type EC to moderate the link members is crucial since they show a higher pro- between an individuals’ motives to using a knowl- pensity compared to newcomers to answer ques- edge search service and the forming of the intentions, contribute content, care about and enforce tion – the users’ predisposition – towards actually norms of appropriate behavior, and support the using it but did not observer continued motiva- community in various other ways behind the tion based on curiosity. Litman (2010) proposed a scenes (Ren et al., under review). Knowing about model for information-seeking behavior that re- community members’ desires such as access to lates I-type and D-type EC to different levels of quality content or creating incentives for site visi- an individual’s wanting and liking of new infor- tors to stay and participate in community activi- mation thereby referring to an appetite for and ties like fostering members’ embeddedness is es- the pleasure in acquiring new information (Ber- sential for community success (Porter and ridge, 1999). Hereby, a high level of wanting new Donthu, 2008; Ren et al., under review). Alarm- information combined with an expected pleasure ingly, several studies exhibited that 90 per cent of – thus the relief of uncertainty in acquiring this community members are passive participants – so information relates to D-type EC (“thorough called lurkers. Lurkers do not interfere with the investigation”) while I-type EC considers the community in a negative way but turning these pleasure in absence of a specific need for infor- passive participants to active members appears mation (“just browsing”). more reasonable than recruiting new members Community members who intend to actively since lurkers are already familiar with community participate in a community such as describing content and culture. Few studies exist on lurking current problems or posting new questions are (Kozinets, 1999; Nonnecke and Preece, 2000; usually in need of new information (Hemets- Rafaeli et al., 2004); none of these studies linked berger, 2002; Lakhani and Wolf, 2005; Fueller et lurking – and de-lurking respectively – to individ- al., 2007). Participation and with it contribution ual levels of curiosity. However, if lurking behav- behavior varies depending on the community ior can be linked to an individual’s curiosity, de- type and context. While members of communi- lurking could be predicted on a personal level via ties of production such as open source software e.g. questionnaires35 assessing the individual’s development or sports enthusiasts’ communities current needs and general predisposition towards collaborate online to solve problems, innovate information seeking. Implementing appropriate products, or generate new product ideas driven measures might later incentivize members to by a specific need36 (Franke and von Hippel, contribute. 2003; Lee and Cole, 2003; Jeppesen and Epistemic curiosity (EC), the individual desire to Frederiksen, 2006) members of online communi- gain new knowledge, close information gaps, and ties of consumption predominantly unite to dis- tackle intellectual problems, is one such personal- cuss their experience with existing brands or ity trait (Berlyne, 1954; Litman and Spielberger, products, their performance, and potential future 2003; Litman, 2008). Litman and Jimerson development for the joy they derive from the (2004) distinguish between interest type (I-type) online activity itself (Algesheimer et al., 2005; and informational deprivation type (D-type) epis- Cova and Pace, 2006; Kozinets, 2007; McAlex- temic curiosity. D-type EC reflects the desire to ander et al., 2002; Schau and Gilly, 2003). Online eliminate undesirable states of uncertainty, where health communities and similar types play a spe- elimination can be seen as rewarding and thus cial role since members support each other in pleasurable while I-type EC directly relates to the

36 Compare von Hippel (1988; 2005) on user innovation to 35 Triggered when a formerly active member turns passive solve personal needs especially lead users who experience for a predefined time. needs ahead of trends

94 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

giving advice and providing suggestions for fur- 2004). However, this model being true would ther action based on individual problems. New- predict that a certain proportion of members comers often find valuable archive information with low D-type trait curiosity cannot be lured online but have to actively participate by asking into continuous active participation regardless of questions if no identical case has been discussed the incentives provided by community managers. yet. Their liking of new information (I-type curiosity) We thus observe an opportunity to close a gap does not suffice and need states are difficult to “what might incentivize lurkers to de-lurk” in maintain over an extended period of time by current literature and explain de-lurking in the pure external stimulation. This finding implied light of curiosity as personality trait (Litman and that from an economical perspective, the costs of Jimerson, 2004) and momentary emotional state incentivizing members to participate increase (Litman, 2010; Naylor, 1981; Spielberger, 1975). heavily after the threshold of members with suffi- Linking I- and D-type curiosity according to Lit- ciently high trait curiosity has been reached. man’s (2010) wanting-liking model of informa- Following this introduction, we review the core tion-seeking to members’ tenure and lurking be- characteristics of online communities, lurking havior is expected to explain an individual’s pro- behavior, and members’ motivation to join and pensity to de-lurking. With D-type curiosity re- contribute to online communities, which leads us flecting an unsatisfied need-like state that moti- to a gap in current research where de-lurking is vates exploration (Litman, 2005), members show- understudied. Section 3 discusses curiosity as a ing this emotional-motivational tendency are personality trait and forms the basis of our re- expected to be more likely to engage in commu- search question. We develop the resulting process nity activity (at all) while members showing a model of lurker curiosity with related hypotheses high level of trait curiosity are expected to engage in section 4 and finally, in section 5 provide con- in enduring, sustained community activity. So far, clusions, implications for research and manage- research on lurking had basically concentrated on ment, the limitations that apply to this model as why lurkers lurk to speculate on potential remedy well as suggestions for a future research agenda. (Nonnecke and Preece, 2001; 2003; Rafaeli et al.,

95 FROM GAME TO PLAY PETER JÄGER • 2010

2. ONLINE COMMUNITIES, MEMBER CHARACTERISTICS, AND MOTIVES FOR PARTICIPATION

Forming communities online has been a desire of scribed as fandom (Kozinets, 2001). While inter- internet users from the very inception of the me- acting, these members create value for firms in dium (Rheingold, 1993; Tafjel, 1978). Whether it different ways: they support a product or service, be blogs, wikis, communities of consumption or promote a brand and spread loyalty to a product production, bulletin boards, discussion lists, or or firm, or act as a resource for ideas (Carlson et social networking sites, large numbers of indi- al., 2007; Kozinets, 1999; McAlexander et al., viduals gather online to form virtual communi- 2002; McWilliam, 2000; Muniz and O'Guinn, ties. Facebook, being one of the most prominent 2001; Nambisan, 2002; Rowley et al., 2007; von social networking sites today, just topped 500 Hippel, 1988). million users37 in July 2010. Most users who en- Online communities thus impact positively on gage in online social bonding share a common firm performance (Armstrong and Hagel, 1996). trait, their interpersonal curiosity, thus their de- Algesheimer and Dholakia (2006) found a 50% sire for new information about other people’s increase in revenues for firms that have managed lives or emotions (Litman and Pezzo, 2007). Be- their communities well. In addition members of sides interpersonal curiosity, other types of curi- online communities remain twice as loyal to and osity exist such as sensory curiosity (Litman et al., buy almost twice as often from the sponsoring 2005b), perceptual curiosity (Collins et al., 2004), firm. With customer loyalty being linked to the or epistemic curiosity (Berlyne, 1954; Litman, economic performance of the firm (Armstrong 2008). The latter is relevant for this study since and Hagel, 1996), community management be- epistemic curiosity drives community members’ comes an invaluable asset (Williams and Cothrel, intentions to acquire new knowledge online (Koo 2000). Online communities can further shift bar- and Choi, 2010). The process of online knowl- gaining power from suppliers to customers (Kozi- edge acquisition is special in that it is often pre- nets, 1999); spread positive word-of-mouth (Dho- ceded by actively inquiring about a topic, e.g. lakia et al., 2004; Hadida, 2009); help firms learn posting questions or stating current problems in a from customers (Kardaras, et al., 2003); increase new forum thread of the online community si- website traffic (Hagel and Armstrong, 1997); raise multaneously making it a contribution. We first entry barriers for competitors (Hagel and Arm- discuss online communities, member characteris- strong, 1997); facilitate product development tics, and motivation to join and contribute before efforts (Nambisan, 2002); and increase customer linking these constructs to the literature on curi- satisfaction and loyalty (Shankar, et al., 2003). osity. Thus, knowing about community members’ desires such as access to quality content or creating incentives for site visitors to stay and participate 2.1 Online Communities in community activities like fostering members’ Online communities that produce software or embeddedness positively impacts on firm per- ideas for new product development have received formance (Porter and Donthu, 2008). attention from innovation scholars (Franke and While the type and purpose of online communi- von Hippel, 2003; Lee and Cole, 2003; Jeppesen ties differ strongly, their members show certain and Frederiksen, 2006), whereas communities similarities in their online social behavior. This that consume physical or virtual products have behavior is stimulated by interest or uncertainty been well described by marketing scholars (Alge- where curiosity resembles the affect as common sheimer et al., 2005; Cova and Pace, 2006; Kozi- denominator. Drawing on such similarities allows nets, 2007; McAlexander et al., 2002; Schau and for research on the core necessities of community Gilly, 2003). Online communities of consump- members, especially in their social interaction tion, such as brand communities, unite consum- and knowledge acquisition behavior. Since cogni- ers in their consumption of products and services tion resembles the link between affect and behav- and their devotion to the brand (Pichler and ior, community design plays a vital role in influ- Hemetsberger, 2007) or as fans of TV series de- encing members’ behavior. Thus understanding members’ needs enables community hosts to 37 Facebook Tops 500 Million Users. By Jenna Wortham, better structure their offering and thus make the New York Times. July 21, 2010 most out of their members’ potential (Dholakia et

96 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

al., 2004). The particular reasons why visitors to report lower figures, a fact that does not collide stay, become members, and eventually contribute with our conception of lurker personality since to the community in form of asking –hence post- we expect members to join OSS developer coming– questions in order to receive answers or munities driven by different motives than mem- commenting on existing threads are therefore bers who join problem solving communities or crucial (Cothrel, 2000). Contribution behavior brand communities. and its motivation vary across online communi- However, drawing a closer look identifies lurkers ties according to their scope as described before to be more than just free-riders as Kollock and while the need for and the liking of new informa- Smith (1996) described their behavior. Lurkers tion experienced by individuals remain the same. lurk while getting familiar with the community or Even though attracted once, retaining active when they intend to leave a group (Nonnecke and members is a serious problem among communi- Preece, 2003). Familiarizing themselves with new ties, which, if not resolved effectively, might lead topics or community norms represents “an im- to their failure (Ren et al., under review). New portant transition mechanism for novices to community members show a strong tendency to learn” (Ahuja and Galvin, 2003; Nonnecke and post once and then either leave the community or Preece, 2001: 3). Community members who – become lurkers resulting in a high member turn- primarily– lurk mentioned reasons such as: they over (Butler, 1999; Joyce and Kraut, 2006). This wished to preserve their anonymity, were shy effect has been studied in the context of Usenet about posting, encountered time constraints, or groups where Arguello et al. (2006) found 68% of experienced the volume of messages posted by new users only posted once, in World of Warcraft fellow community members too low or too high with 25% of guild members leaving after one (Butler, 2001; Rafaeli et al., 2004; Preece et al., month (Ducheneaut et al., 2006), or Perl OSS 2004; Katz, 1998; Wasko and Faraj, 2000). developers with 54% “post-and-vanish” behavior Considering these motives brings along a differ- (Ducheneaut, 2005). Considering these problems ent perspective: lurkers should be valued since in conjunction with the huge gains firms draw they work at knowing the group prior to full en- from online communities, research started to try, try not to add to the chaos, extend the group focus on the question why lurkers lurk (Nonnecke into real life by reporting online content to their and Preece, 2001; 2003; Rafaeli et al., 2004). friends or family, side-post to other communities, and make a commitment in joining a group than just browsing it (Nonnecke and Preece, 2003). 2.2 Lurking Behavior Experienced members often regard messages Over time, community members evolve and take posted by newcomers immediately after sign-up different roles (Wenger, 1998). Starting as visitors, as spam. Hence, not adding to the chaos reflects new members continue their lifecycle as new- the fact that newbies, so called “noobs”, who comers and finally take the role of experts (Kim, decide to start off as lurkers reduce the overall 2000). Each role brings along new levels of skills message load. As mentioned before, spam deters and experiences for the member which impacts members from contributing who prefered lurking on their reputation and status (Wenger, 1998; instead when faced with the high volume of – Kim, 2000; Iriberri and Leroy, 2008). meaningless– messages. Side-posting to other Irrespective of their membership stage, users communities or reporting online community show a strong tendency towards lurking. Lurkers content to real-life friends can be considered do not actively participate in community activities word-of-mouth activities that raise awareness of but take the role of passive observers (Kozinets, and might attract potential new members into the 1999; Nonnecke and Preece, 2000; Rafaeli et al., existing community (Butler et al., 2007). Hence, 2004). For lurkers, browsing the online commu- the community benefits from the lurkers’ mediat- nity content appears to be sufficient to gather the ing activities. While several studies have been desired information, a fact that explains more conducted on Internet users’ motivations to join than 50% of lurking behavior (Preece et al., and contribute to online communities (Moore 2004). Various research indicates lurking levels to and Serva, 2007), less is known about incentives be around 90% depending on the community and stimuli that motivate members to de-lurk type and purpose (Katz, 1998; Mason, 1999; thus to turn to active participation (Rafaeli et al., Nonnecke and Preece, 2000; 2001; 2003). Open 2004). A question, we address throughout this Source Software development communities tend paper.

97 FROM GAME TO PLAY PETER JÄGER • 2010

2.3 Motivation to Join and Contribute Shah, 2006 for similar findings). This compares to the crowding-out effect, which explains a de- One major framework to explain individual mo- crease in activity of an intrinsically motivated tivation is the self-determination theory intro- community after extrinsic motives were introduced by Deci and Ryan (1985). Knowing the duced (Alexy and Leitner, 2008; Frey and Ober- source of motivation forms a better understand- holzer-Gee, 1997; Frey and Jegen, 2001;). For ing of its interrelation with initiated action. Self- blog participation, Hsu and Lin (2008: 67) identi- determination theory (SDT) draws a basic dis- fied “knowledge sharing factors such as expected tinction between intrinsic and extrinsic motiva- reciprocal benefit, reputation, expected relation- tion (Deci and Ryan, 1985). Either one eventually ships, trust and altruism” to motivate members. drives action. While an extrinsically motivated Yu et al. (2007) studied knowledge contribution in action is performed to obtain a separable output, problem solving virtual online communities intrinsically motivated action is performed for the where they identified eight motivations similar to inherent interest or joy of it (Deci and Ryan, Hsu and Lin (2008) (see also Constant et al., 1985). Action motivated intrinsically is proposed 1994). For a classification of online community to be driven by the inherent intrinsic needs: com- types and the related motivational factors see petence, autonomy, and relatedness (Deci and Moore and Serva (2007) who aggregated 14 mo- Ryan, 2002). Examples of internal motives (in- tivational categories relative to four types of on- trinsic motivation) are fun and enjoyment, com- line communities. munity identification, or altruism. Pure extrinsic motives comprise career or pay (for an extensive Motivation to join and contribute to an online review of the motivation literature see von Krogh community can be influenced or stimulated by et al., 2010). several external factors. von Krogh and colleagues (2010) analyzed the relationship between Advancing the classic SDT perspective, Roberts developers’ motivations and institutional ar- and colleagues (2006) described reputation, recip- rangements and found an interrelation between rocity, learning, and own-use value as internalized individual motivation and community sponsor- extrinsic motivations, which they studied in the ship or license restrictions. Shah (2006) studied context of Open Source Software development. the influence of governance structures, especially These, by definition extrinsic motivations, regu- regarding the evolution of motives. Applying the late developers’ behavior and are directly related collective effort model to reduce social loafing to competence acquisition, one of the intrinsic (Karau and Williams, 1993; 2001) Ling and col- needs identified by Deci (1975). Internalized ex- leagues (2005) identified that explaining to users trinsic motives are thus not motivated by pure what value they had added by contributing has a interest but by personal goals an individual de- positive effect on future contribution behavior sires to achieve, like for example reputation that especially when the value is created for a defined has a positive impact on career goals. subgroup of the entire community, which the In the context of online communities, internal individual identifies with (Ling et al., 2005. motives such as knowledge sharing and relation- Rashid et al., 2006; see also Preece, 2004). Thus, ship building appear to dominate. This has been highlighting the benefits one adds to their group studied in the context of health care communi- can stimulate altruistic motives. ties, where individuals seek information related to Advancing the discussion of motives, Joyce and their disease or available cures (Preece et al., Kraut (2006) decomposed community participa- 2004), in communities of hobbyists who innovate tion into two distinct components that are stimu- on tangible products like Lego and Playmobil lated by different motives: 1) joining and 2) con- (Janzik et al., 2010) or intangible products like tributing to an online community (see also Open Source Software development (Jeppesen McGrath and Hollingshead, 1994; Moreland and and Frederiksen, 2006), and in professional user Levine, 2001; Straus and McGrath, 1994; Wit- communities where internalized extrinsic motiva- tenbaum et al., 1999). Janzik and colleagues tion dominates, like in communities of practice (2009) observed these motivational dynamics (Brown and Duguid, 2001; Lave and Wenger, more closely. In their study on privately operated 1991). Jeppesen and Frederiksen (2006) identified online communities dedicated to tangible con- hobbyists’ intrinsic motivations to foster commu- sumer products, they showed how members’ mo- nal innovation behavior in contrast to extrinsic – tives evolve over time. Motives to sign up to on- economically driven– motives of professional line communities differed from motives to con- users who contributed far less than hobbyists (see

98 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

tribute in later stages when members’ acquired a members feel more comfortable and are more certain level of experience. While in the early likely to post to online communities after having stage a strong need for information and support reached a certain expert level (Wasko and Faraj, was recognized (Hemetsberger, 2002; Lakhani 2005). This finding especially holds true for pro- and Wolf, 2005; Fueller et al., 2007), flow- fessional communities or communities of practice experience (Czikszentmihalyi, 1990) played an where enhancing members’ professional reputa- important role in later stages. The flow-model tion is of relevance (Brown and Duguid, 2001; predicts members to return to an activity due to Lave and Wenger, 1991). The fear of providing growing abilities and the experience of participa- wrong answers can – at the same time – be a tion itself – members get absorbed in their activ- barrier to participation when members lack a ity (Nakamura and Czikszentmihalyi, 2002). sufficient expert status. To counter this, trust Thus, personal needs and the identification with needs to be established among members of the online community are of major importance. communities (Ardichvili et al., 2003; Leimeister et Substantiating this, Ren et al. (under review) al., 2005; Porter and Donthu, 2008; Ridings et found community commitment to increase over al., 2002; von Krogh, 2002). time due to a shift from members’ identity-based In a nutshell, most of the motives identified to to bond-based commitment. drive participation in online communities are of In later stages, altruistic motives like the desire to intrinsic or internalized extrinsic nature (cp. Jan- give something back to the community domi- zik et al., 2009; von Krogh et al., 2010) including nated (Hall and Graham, 2004; Janzik et al., social bonding but above all the desire to learn 2009; Ren et al., 2007; von Krogh, 1998; Wasko and the need for new information. Describing and Faraj, 2000). Philosophically speaking, the current problems to ask specific questions very conversion from “I-intention” to “we-intention” often precedes acquiring new information, which has completed (Bagozzi and Dholakia, 2006; is provided in forum threads of the online com- Gilbert, 1989; Nonaka, 1991). Additionally, munity that follow up on the initial question.

99 FROM GAME TO PLAY PETER JÄGER • 2010

3. CURIOSITY AS PERSONALITY TRAIT

“Curiosity, the intrinsic desire to seek out and acquire recognizing the importance of the internal states new information (Berlyne, 1966; Loewenstein, 1994), that eventually motivate specific or diversive ex- comprises two basic goals: The first involves obtaining ploration, Berlyne did not develop measures to information expected to stimulate positive feelings of in- assess individual differences in epistemic and terest (I-type), while the second is concerned with reduc- perceptual curiosity as personality traits. ing undesirable states of informational deprivation (D- type) (Litman, 2008). I-type curiosity is activated when individuals recognize opportunities to discover something 3.2 Curiosity as Emotional-Motivational State entirely new, whereas D-type curiosity is stimulated when and Personality Trait people lack information they intend to incorporate into an Cattell and Schreier (1960), while studying anxi- existing body of knowledge (Litman, 2008; Litman, ety, first recognized and empirically studied indi- Hutchins, & Russon, 2005; Litman & Jimerson, 2004; Litman & Spielberger, 2003; Loewenstein, vidual differences in how often internal states 1994).” (Litman, 2010: 397) associated with curiosity as motivator were experienced. Later Spielberger (1975) advanced the idea of the state-trait distinction thereby laying 3.1 Revisiting Berlyne: Epistemic Curiosity out the foundation for the State-Trait Curiosity In- ventory (STCI) (Spielberger and Butler, 1971; Curiosity as conceptualized by Berlyne (1949; Spielberger et al., 1976; 1981). “The STCI is 1950) is the individual’s reaction to novel stimuli comprised of scales that assess the intensity of that cause feelings of interest or uncertainty, two curiosity at a particular time as an emotional- internal states which motivate the individual to motivational state (S-Curiosity), and how often acquire new knowledge (Berlyne, 1954). Distin- curiosity is generally experienced as a personality guishing between perceptual and epistemic curiosity, trait (T-Curiosity).” (Collins et al., 2004: 1128). the two emotional states38 related to interest and Trait curiosity thereby refers to individual differ- uncertainty, allowed Berlyne to predict different ences in the potential for curiosity experiences (if) action. Epistemic curiosity is stimulated by com- while state curiosity refers to individual differ- plex problems like scientific theories that call for ences in curiosity experiences themselves (how) asking specific, reflected questions or testing hy- (Naylor, 1981). Measures of the STCI like “I feel potheses to arrive at answers and thus gaining mentally active” provide a close link to Berlyne’s new knowledge (Berlyne, 1954). Perceptual curi- concept of epistemic curiosity but do neither osity is triggered by sensory stimulation like assess perceptual curiosity nor distinguish be- sounds or sights and motivates the individual to tween diversive and specific exploration. For a inspect and observe the stimulus in order to ac- thorough review of different scales to assess curi- quire new information (Berlyne, 1957; 1958). osity, novelty and sensation seeking see the work Advancing his theory, Berlyne (1960; 1966) later by Collins and colleagues (2004) in which they introduced the concept of specific and diversive established perceptual curiosity as meaningful exploratory behavior where both, epistemic and personality construct besides epistemic (Berlyne, perceptual curiosity, stimulate specific explora- 1954; Litman, 2008), interpersonal (Litman and tion. He assumed this to happen in situations Pezzo, 2007) and sensory curiosity (Litman et al., characterized by rich, new stimulation. In situa- 2005b). tions when stimulation is low or absent, individuals tend to get bored and start exploring their surroundings in order to find stimulation, a be- 3.3 Curiosity as Feelings of Interest and Depri- havior he termed diversive exploration. Once vation exposed to new stimulation, individuals engage in acquiring new information about the stimulus Lowenstein (1994) linked uncertainty that results driven by specific exploratory behavior. Though from perceived knowledge gaps with curiosity, which manifests itself in unpleasant feelings of deprivation (D). Thus, information-seeking behav- 38 The differentiation of curiosity as emotional state vs. ior to acquire new knowledge is motivated by the personality trait as conceptualized in Spielberger’s work is individual’s desire to abolish the discomfort of discussed in the next chapter. uncertainty (Lowenstein, 1994). Contrary to the

100 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

work by Spielberger and Starr (1994), Lowenstein more related to levels of state-curiosity with a did not regard information-seeking behavior mo- greater impact on information-seeking behavior tivated by pure feelings of interest (I) as curiosity. than I-type curiosity. Combining both ideas to assess individual differences in curiosity, Litman and Jimerson (2004) took the parallel to eating when they integrated 3.4 An Integrated Wanting-Liking Model of Lowenstein’s reduction-oriented and Spielberger Information-Seeking and Starr’s induction-oriented model of curiosity. Litman (2005) associated these emotional- They posit that curiosity as feeling of interest motivational states associated with I- and D-type (CFI: induction-oriented) and curiosity as feeling curiosity and their subsequent intention for in- of deprivation (CFD: reduction-oriented) can co- formation-seeking to varying degrees of wanting exist (Litman and Jimerson, 2004) by comparing and liking (Berridge, 1999). Wanting and liking curiosity to hunger, which may be stimulated by are both neural processes that affect appetitive pure physiological aspects but as well by the motivation and the resulting feelings of pleasure pleasant smell of food with both giving pleasure by dopamine and opioid activation (Berridge, upon food consumption. 1999; Berridge & Robinson, 1998). While want- Relating Litman and Jimerson’s work to the SDT ing resembles the intensity of “appetite” for new framework (Deci and Ryan, 1985; 2002; Ryan information, liking resembles the “joy”, hence the and Deci, 2000), I-type curiosity is largely consis- degree of pleasure derived from the acquisition tent with pure intrinsic motivation while D-type of new information (Berridge, 2003; 2004; Lit- curiosity with its reduction of unpleasant states of man, 2005; 2010). Wanting and liking being two uncertainty more relates to internalized extrinsic cooperative but dissociated processes (Berridge, motivation (Litman, 2008; Roberts et al., 2006; 2001; Berridge and Robinson, 1998) are activated Ryan and Deci, 2000). I-type curiosity is thus a contiguously meaning that stimuli that are pleasurable feeling towards knowledge acquisition wanted are also expected to be liked (Berridge, in a playful manner. Enjoyment from engage- 1999). However, since both systems operate inde- ment, entertainment, or aesthetic pleasure is as- pendently they can as well be activated to differ- sociated with learning and new information (Lit- ent degrees. Addicts, for example, who crave for man, 2005; 2008). Accordingly, individuals are in drugs (high wanting) might no be experiencing a positive mood when experiencing I-type curios- much pleasure (low liking) from taking the drug ity, which predicts a negative correlation with (Robinson & Berridge, 2003). The resulting want- states of anger, anxiety, and depression – scales ing-liking relationships with their related emo- that are used to assess individual differences in I- tional-motivational states are illustrated in figure type curiosity (for empirical studies on I-type 1. curiosity see: Litman and Jimerson, 2004; Litman I-type curiosity experienced by the individual as and Spielberger, 2003; Litman, 2010). Contrast- “subrational liking” is characterized by relatively ing this state, individuals who are seeking new high liking in relation to relatively low wanting information or knowledge due to an unmet need (Litman, 2005; 2010). This compares to why food or uncertainty tend to experience feelings of an- may be enjoyed (high liking) while the individual ger or anxiety (Litman, 2005; Litman and Jimer- does not experience any significant hunger (low son, 2004; Litman and Spielberger, 2003, Lit- wanting) (Berridge, 2003; 2004). Regarding in- man, 2010). Thus, for D-type curiosity, new information-seeking behavior, individuals enjoy formation is rewarding since it reduces negative browsing and finding new pieces of information, feelings attributed to uncertainty (Litman, 2010). which stimulate their curiosity to further engage The degree to which each type of curiosity is in the topic while they do not have any concrete experienced and ultimately drives action varies questions they pursue (Litman, 2005). Measures across individual differences in relatively stable I- for I-type curiosity hence comprise items like and D-type curiosity tendencies (Litman, 2010). (Litman, 2008): “Enjoy exploring new ideas”, Both “assess positively correlated but psychomet- “Find it fascinating to learn new information”, rically distinct dimensions of trait-curiosity (Lit- “Enjoy learning about subjects that are unfamil- man, 2008; Litman and Jimerson, 2004; Litman iar to me”, “Enjoy discussing abstract concepts”, & Silva, 2006)” (Litman, 2010: 397). However, and “Learn something new, like to find out more Litman (2005) reported D-type curiosity to be about it”

101 FROM GAME TO PLAY PETER JÄGER • 2010

ual experiences “general indifference” (Litman, 2005). Winkielman and Berridge (2003) described the forth and last wanting-liking combination as “irrational wanting” which resembles a need for closure (NFC) (Web- ster and Kruglanski, 1994) where individuals seek any information to resolve uncertainty independent of how valuable this information is. Such individuals prefer simple answers to intellectually complex ones (Webster and Kruglanski, 1994). Figure 1. Hypothesized emotional-motivational tendencies relevant to informa- Another example might be mor- tion-seeking associated with relatively high or low levels of wanting and liking. bid curiosity, a yet understudied (Litman, 2010: 398) concept, which leads individuals to investigate –per se– unpleas- Individuals who lack essential information that ant things like scenes of violence or death (Zuck- they need to complete a task or information con- erman & Litle, 1986). On the NFC – D-type cerning their health are exposed to feelings of curiosity scale, the need for closure directs indi- uncertainty. Since arriving at correct answers viduals to arrive at the simplest possible answer constitutes relief, which can be seen as pleasur- even if erroneous while D-type curiosity drives able experience, these individuals show intensive individuals to finding the right answer by thor- craving resulting from the high wanting and an- ough investigation (Litman, 2010). ticipated high liking of new information. This emotional-motivational sate is thus consistent Considering the evolution trajectories of com- with D-type curiosity (Litman, 2005; 2010). munity members as discussed in section 2, we Measures for D-type curiosity comprise items like now relate the information-seeking model to the (Litman, 2008): “[Spend] Hours on a problem different stages of community membership because I can’t rest without answer”, “Brood for thereby paying special attention to lurking behav- a long time to solve problem”, “Conceptual prob- ior in the respective stages. The question we ad- lems keep me awake thinking”, “Frustrated if I dress in our model is: How are members who can’t figure out problem, so I work harder”, and de-lurk influenced by different levels of “Work like a fiend at problems that I feel must be wanting and liking, thus their momentary solved” emotional states and enduring personal- Low levels of wanting and liking can be linked to ity traits of I- and D-type curiosity. Litman ambiguity tolerance (AT) (Litman, 2005) a state in (2010: 401) proposed future research “to examine which the individual neither experiences much the relationships among these variables in a fol- pleasure from new information nor the desire to low-up study that also assesses transient states as acquire it. “AT reflects a motive to accept rather well as providing opportunities for participants to than resolve uncertainty, suggesting that AT tenden- seek (or choose not to seek) new information.” cies may direct individuals to passively avoid op- Members of online communities do have this portunities to resolve uncertainty and learn new choice in that they can easily switch between information.” (Litman, 2010: 398). The individ- lurking and active contribution.

102 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

4. TOWARDS A PROCESS MODEL OF LURKER CURIOSITY

Internet users who form the intent to join an from interacting with the community or their online community often seek specific information observation of the community as lurkers. to reduce uncertainty or close knowledge gaps We assume, that “early-lurkers” are in a transi- (Hemetsberger, 2002; Lakhani and Wolf, 2005; tional state of I-type curiosity with low individual Fueller et al., 2007). D-type trait curiosity. While gathering more in- According to the Litman (2010) model of infor- formation about a certain topic members might mation-seeking, new members with their strong not obtain the desired answers but ponder on appetite for new information are expected to their original question or even come across new show distinct states of D-type curiosity which questions. Hence, stimulated by increasing uncer- eventually motivate them to ask respectively post tainty imposed by additional questions, their need questions to the online community. On the other for information together with the expected pleas- hand, members experience discomfort in posting ure from relieving this feeling of uncertainty and to the community since they feel insecure, have thus their D-type state curiosity rises to the point language constraints, or fear negative feedback where they may turn active members. This be- (Wasko and Faraj, 2005). Based on the motivation havior is illustrated in figure 2: transition from 1 literature, joining and posting could thus be con- to 2. Members who contribute to online commu- ceptualized as various types of costs for the indi- nities of production might thereby be better able vidual. However, if the discomfort imposed by to formulate their immediate need as question to uncertainty does not reach the threshold of the the community than members who just browse discomfort related to de-lurking (Wasko and the content communities of consumption because Faraj, 2005), members tend to first gather information and familiarize themselves with the online community before posting or asking questions (Non- necke and Preece, 2003). Put differently, how much curiosity does a lurker need in order to offset the cost of joining or posting. For instance, members of health communities who are rather unfamiliar with online interaction might experience posting a question as relatively high cost compared to a software developer who is used to communicate problems via online forums. On the other Figure 2. Lurker dynamics in the Litman (2010) information-seeking model hand, uncertainty related to a personal health issue can be expected to create they like to explore new information about a unpleasant need-states far beyond those of a product or brand. The latter group might first software developer who intends to obtain a quick have to gather enough information before they solution compared to superfluous own work. In- are able to contribute, stumble over an intriguing dividuals in either case are inclined to actively topic, which raises their interest, or experience an contribute to the online community (need-state actual problem with the product for which they stimulating curiosity) if the value of the expected desire a solution from the online community. information (liking) exceeds the cost of inquiring Based on these considerations we hypothesize: (joining, posting). Members therefore might assess 1) the likelihood of obtaining relevant informa- H1: Internet users who join but lurk before contributing tion and 2) the potential quality of the obtained to an online community motivated by the pleasure information based on either their past experience of new information (high liking, insufficient want-

103 FROM GAME TO PLAY PETER JÄGER • 2010

ing) show emotional-motivational tendencies of I- tion 3 to 2 to 3 in figure 2). While the hypotheses type curiosity. that predict transitions in our model relate to tendencies (states) of I-/D-type curiosity, behav- H2: Internet users who join and immediately contribute ior in one stage is predicted by trait curiosity. We to an online community motivated by a high need thus expect: for information (high wanting) show emotional- motivational tendencies of D-type curiosity. H4: Community members with high D-type trait curiosity contribute more frequently to the community in While spending time in the online community, later stages when stimulated by new information. members close their knowledge gaps and eventually reach an expert level (Kim, 2000). Acquiring H5: Community members with low D-type trait curios- new information is of less importance in this ity but high I-type trait curiosity maintain their evolutionary stage while the wish to contribute membership in later stages but show a higher pro- back to the community increases (Hall and Gra- pensity to lurk. ham, 2004; Janzik et al., 2009; Ren et al., 2007; von Krogh, 1998; Wasko and Faraj, 2000). For Examples for community members with high D- instance, members of a community of filmmak- type trait curiosity might be found in online ers first learn about the new technology of using communities of production where lower levels of video games as tools for animation production lurking are reported while members with low D- and later, when starting businesses based on the type but high I-type trait curiosity are more ex- technology and hence turn experts, still contrib- pected to be found in online communities of ute to the community how the professionalized consumption where participation is predomi- method works for them (Haefliger et al., 2010). nantly driven by the liking of new information Another motive to continue participation is the related to a product or brand and need-states individuals’ desire to be stimulated by new ideas only occur when individuals face problems and from fellow community members (Berlyne, 1960; desire immediate support. 1966; Czikszentmihalyi, 1990; Janzik et al., 2009; When both, the level of appetite as well as the Litman and Jimerson, 2004; Nakamura and pleasure from acquiring new knowledge decrease, Czikszentmihalyi, 2002). Sports enthusiasts often membership comes to an end (Shah, 2006). Indi- use communities to obtain information about and viduals might shift their attention to new topics discuss new products, technologies, techniques, or when they cease to perform a specific sport or tricks related to their favorite sports as has been hobby activity or switch to products of a compet- described for rodeo kayaking (Baldwin et al., ing brand. This represents a state of ambiguity 2006), kite surfing (Tietz et al., 2005), or moun- tolerance with a general indifference towards the tain biking (Lüthje et al., 2005) communities’ activities (Litman, 2005). “People Reduced appetite for but high pleasure with new tend to leave if there isn’t new material to keep information thus indicates a shift from D- to I- them engaged” (Nonnecke et al., 2004: 6). Before type curiosity in later stages of membership terminating their membership, members usually (compare figure 2: transition from 2 to 3). This move into a state of fade-out where they take the leads us to hypothesize: role of lurkers (Nonnecke and Preece, 2003). They still retain their membership but are con- H3: Community members with higher tenure who turn stantly loosing interest in the community’s activi- experts and are motivated by fun and personal ties (compare figure 2: transition 3 to 4). We needs (high liking, low wanting) show emotional- therefore hypothesize: motivational tendencies of I-type curiosity. H6: Community members who intend to leave the group With their primary needs satisfied (low wanting), experience low levels of wanting and liking new we expect community members’ further contribu- information and thus show motivational-emotional tion behavior to depend on their D-type trait cu- tendencies of ambiguity tolerance. riosity. This might be compared to classrooms where some students continuously ask questions Figure 2 illustrates how these six hypotheses relate (Spielberger et al., 1981) while others only pipe to the information-seeking model (Litman, 2010). up when they are exposed to the fear of not keep- Considering the fact, that lurking is expected to ing pace any longer (this is reflected by the fre- happen in stage 1, 3 or 4 according to our quency (Litman and Spielberger, 2003) of transi- amended model, we state two further hypotheses:

104 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

H7: Taking a passive role motivated by low discomfort We do not expect community members to show a with uncertainty, lurkers show low levels of want- motivational-emotional tendency of need for ing. closure since such individuals prefer to seek the easiest obtainable answers. These individuals are H8: Since members lurk before and after contributing, therefore not expected to sign up to online com- lurking behavior is unrelated to the level of liking. munities.

105 FROM GAME TO PLAY PETER JÄGER • 2010

5. DISCUSSION AND CONCLUSIONS

Given the measurable personality traits related to complements them in relating known reasons for ambiguity tolerance, I- and D-type curiosity (see lurking and motivations for joining which can be Spielberger, 2010), the lurking respectively de- conceptualized as cost to curiosity as personality lurking behavior of community members should trait that might offset these costs. become more predictable. We theorize that contribution to an online community is very much dependent on high levels of either D-type trait or 5.2 Implications for Community Managers D-type state curiosity (for state-trait curiosity see: Spielberger and colleagues (1981) identified col- Collins et al., 2004; Spielberger, 1975; for feelings lege students with high trait curiosity scores to ask of deprivation to motivate curiosity see Lowen- more than twice as many questions as students stein, 1994). Additionally, I-type trait curiosity with low trait curiosity (T-Curiosity) sores. Like in keeps the individual engaged in the process of classrooms where few students with high trait exploring community content which again can curiosity scores post the majority of the questions, stimulate sufficient levels of D-type state curiosity comparatively few community members will (for individuals with a tendency of relatively show the adequate level of D-type trait curiosity lower D-type trait curiosity) to contribute. In (the wanting of new information) to contribute short, D-type curiosity drives contribution while continuously. Regardless of the provided incen- I-type curiosity keeps members engaged in ob- tives – these incentives can only stimulate state serving the community. Joining and posting can curiosity (S-curiosity) – a significant proportion of therefore be conceptualized as cost for the indi- members will contribute relatively less new con- vidual, which might be offset by curiosity. tent and some will always lurk. Thus, from an economical perspective, the costs of incentivizing members to participate increase heavily after a 5.1 Implications for Research certain threshold has been reached. “Our review takes a first step toward mining social sci- In order to sustain a healthy community, mem- ence theories to inform community design. We also argue bers with I-type curiosity need to be stimulated more generally that we can study theory in social psychol- with new and interesting information or quality ogy, sociology, and economics to help us take a more prin- content as Porter and Donthu (2008) described it. cipled approach to understanding online communities.” Providing content that is “liked” thus attaches (Ren et al., 2007:402) members to the community. As soon as these We extend current literature on online commu- members get “bored” and do not find new stimuli nity members’ lurking behavior in applying cur- (Berlyne, 1960; 1966), they reach a state of ambi- rent research on curiosity from personality psy- guity tolerance and fade-out (Litman, 2005). chology to explain lurking behavior in online Should ambiguity tolerance among community communities thereby following Ren and col- members reache a certain threshold, chances are leagues’ suggestion combined with the recom- high, that the entire community will fail since mendation by Nonnecke and Preece (2003) “to “experts” who are the ordinary contributors (Hall cover non-public interaction, and non-online and Graham, 2004; Janzik et al., 2009; Ren et al., interaction will [to] reveal even more about lurk- 2007; von Krogh, 1998; Wasko and Faraj, 2000) ers’ community involvement.” leave and novices are not attracted into the online Most studies on lurkers focused on external mo- community any more – potentially resulting in a tives for lurking behavior like restrictions by the downward spiral. employer, fear of reputation loss, missing incen- We theorize that the majority of members who tives, or time constraints (Butler, 2001; Rafaeli et join an online community are most likely moti- al., 2004; Preece et al., 2004; Katz, 1998; Wasko vated by situational needs (D-type state curiosity) and Faraj, 2000). In concentrating on personality and not by inherent, hence lasting, desires (D- traits of online community members, we advance type trait curiosity) to acquire new information, current research by explaining lurking behavior which we expect to be responsible for the high with low levels of individual D-type curiosity. fraction of lurkers (Nonnecke and Preece, 2000; This does not render existing studies obsolete but Rafaeli et al., 2004). As soon as new members

106 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

obtain their needed information, their wanting, network sites. Apart from social networking sites or appetite as Litman (2005) put it, is satisfied and where the prime interest lies in social bonding, they convert to lurkers. This is consistent with the this might especially be true for health support observation that most members follow a “post- communities where empathy (Preece, 1999a; and-vanish” behavior (Butler, 1999; Joyce and 1999b) is highly prominent. Kraut, 2006). Re-motivating these members Our research clearly lacks from empirical sub- would require inducing a need-state equivalent to stantiation. In a next step, we recommend testing the original experience when the member posted our model and the related hypotheses (see figure initially. 3) in a real online community setting. Applying a large sample study using online questionnaires appears to be most appropriate (Andrews et al., 2003). A study of this kind is not supposed to be biased towards active community members since Andrews and colleagues (2003) identified lurkers to largely respond to such questionnaires. Traits are generally understood to be enduring behavioral patterns of an individual. Since wanting and liking new information change over time, which in turn triggers different behavior, curiosity is seen to be a relatively stable personality trait as well as a temporary emotional state. D-type curiosity is on a relatively higher level regarding wanting and thus motivation for action that can, Figure 3. Identified variables and dependencies to test if triggered, “overrule” I-type curiosity. Accord- our hypotheses ing to the state-trait concept by Spielberger (1975), curiosity varies across individuals and over time. Studying the same individuals across different communities, hence alternative contexts 5.3Limitations and Future Research would yield important insights into how much Our theory does not reflect the fact that commu- curiosity states depend on the community charac- nity members with higher tenure tend to contrib- teristics. This might deliver insights on how to ute driven by altruistic motives (Hall and Gra- create incentive structures for individuals with ham, 2004; Janzik et al., 2009; von Krogh, 1998; rather low D-type trait curiosity. Another ap- Wasko and Faraj, 2000). Then again, these expert proach to identify stimuli that keep individuals members are not actively seeking new knowledge engaged would be a longitudinal study of few (D-type curiosity) but enjoy discovering new ideas individuals in one or many communities. This “en passant”. They are thus knowledge-seeking might result in new findings on how frequent D- wise in an I-type curiosity state. We respond to type curiosity is activated based on the personal- this in speculating, that experts with I-type curiosity traits and the external stimulation. Again, ity do not start new threads in online communi- insights for community design and operation ties but merely reply to existing threads where could be gathered. they have an answer readily available. However, Since D-type (state) curiosity can be stimulated, experts who reply to complex questions would future research should address the question what need to first reflect on the problem to provide an incentives or stimuli are needed and how often, accurate and satisfying answer, which in turn depending on the personality traits, these states implies motivation by D-type curiosity. of high D-type curiosity can be activated. There- A further explanation lies in the nature of this fore it is important to observe whether the aver- study which concentrates on explaining knowl- age level of state or trait curiosity of individuals is edge seeking behavior. When social bonding as dependent on the community. For instance, do occurs in later stages of community membership health support communities attract members becomes apparent, interpersonal curiosity (Lit- with higher (D-type) state curiosity since indi- man and Pezzo, 2007) might be a more viable viduals are desperate for cures –once– (high predictor of continued commitment like in social wanting) while innovation communities attract members with higher (I/D-type) trait curiosity

107 FROM GAME TO PLAY PETER JÄGER • 2010

who enjoy playing and experimenting and thus We conceptualize joining and posting as cost for yield continuous interest for the community? the member with curiosity offsetting these costs. When constructing his wanting-liking model of However, we did not distinguish whether the information seeking, Litman (2010) did not ex- various costs of joining, depending on the moti- plicitly emphasize the differences between state vation, are different for individuals. Further stud- and trait curiosity. We believe this to be an impor- ies could advance our understanding of the dif- tant aspect that needs to be integrated into the ferent costs of joining, how they differ across model for a better understanding of the shifts individuals, and whether they depend on the that take place between lurking and contributing. community type. A question we deter to future research.

108 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

ACKNOWLEDGEMENTS

I would like to thank my colleagues Michaela This project received support from the Swiss Kolbe and Georg von Krogh for their invaluable National Science Foundation (grant 100014 support. 125513/1)

109 FROM GAME TO PLAY PETER JÄGER • 2010

REFERENCES

Ahuja, M.K. & Galvin, J.E. 2003. Socialization in Virtual Berridge, K.C. & Robinson, T.E. 1998. The role of dopa- Groups. Journal of Management. 29(2): 161-185. mine in reward: Hedonics, learning, or incentive salience? Brain Research Reviews. 28(): 308-367. Alexy, O. & Leitner, M. 2008. Norms, Rewards, and their Effect on the Motivation of Open Source Software Develop- Berridge, K.C. 1999. Pleasure pain desire and dread: Hid- ers. Technical report, Technical University of Munich. den core processes of emotion. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: Foundations of hedonic Algesheimer, R.; Dholakia, U.M. & Herrmann, A. 2005. psychology (pp. 527-559). New York: Russell Stage Founda- The social influence of brand community: Evidence from tion. European car clubs. Journal of Marketing. 69(3): 19-34. Berridge, K.C. 2001. Reward learning: Reinforcement, Algesheimer, R. & Dholakia, P.M. 2006. Do customer com- incentives and expectations. In D.L. Medin (Ed.), Psychology munities pay off ? Harvard Business Review. 84(11): 26- of learning and motivation (Vol. 40, pp. 223-278). Orlando, +. FL: Academic Press. Andrews, D.; Nonnecke, B. & Preece, J. 2003. Electronic Berridge, K.C. 2003. Pleasures of the brain. Brain and Survey Methodology: A Case Study in Reaching Hard-to- Cognition. 52(): 106–128. Involve Internet Users. International Journal of Hu- man-Computer Interaction. 16(2); 185-210. Berridge, K.C. 2004. Pleasure, unfelt affect and irrational desire. In A.S.R. Manstead, A.H. Fischer & N. Frijda (Eds.), Ardichvili, A., Page, V. & Wentling, T. 2003. Motivation and Feelings and Emotions: The Amsterdam Symposium (pp. barriers to participation in virtual online knowledge-sharing 43-62). Cambridge, UK: Cambridge University Press. communities of practice. Journal of Knowledge Man- agement. 7(1): 64-77. Brown, J.S. & Duguid, P. 2001. Knowledge and organization: a social-practice perspective. Organization Science. Arguello, J.; Butler, B.S.; Joyce, L.; Kraut, R.E.; Ling, K.S.; 12(2): 198–213. Rosé, C.P. & Wang, X. 2006. Talk to me: Foundations for successful individual-group interactions in online communi- Butler, B. 1999. The dynamics of cyberspace: Examin- ties. In: Proceedings of the 2006 ACM Conference on Hu- ing and modeling online social structure. Unpub- man Factors in Computing Systems, New York: ACM Press. lished doctoral thesis, Carnegie Mellon University, Pittsburg, 959-968. PA. Armstrong, A. & Hagel, J III. 1996. The Real Value of On- Butler, B.S. 2001. Membership Size, Communication Activ- Line Communities. Harvard Business Review. May- ity, and Sustainability: A Resource-Based Model of Online June: 134-141. Social Structures. Information Systems Research. 12(4): 346-362. Bagozzi, R.P. & Dholakia, U.M. 2006. Open source software user communities: a study of participation in Linux user Butler, B.; Sproull, L.; Kiesler, S. & Kraut, R. 2007. Com- groups. Management Science. 52(7): 1099-1115. munity effort in online groups: Who does the work and why? In S. Weisband. (Ed.), Leadership at a distance. (pp. 171- Baldwin, C.; Hienerth, C. & von Hippel, E. 2006. How user 194). Hillsdale, NJ.: Lawrence Erlbaum Associates. innovations become commercial products: A theoretical investigation and case study. Research Policy. 35 (9): Carlson, B.D.; Suter, T.A. & Brown, T.J. 2007. Social versus 1291–1313. psychological brand community: The role of psychological sense of brand community. Journal of Business Re- Berlyne, D.E. 1949. Interest as a psychological concept. search. 61(4): 284-291. British Journal of Psychology. 39(): 184–185. Cattell, R.B. & Scheier, I.H. 1960. Stimuli related to stress, Berlyne, D.E. 1950. Novelty and curiosity as determinants of neuroticism, excitation, and anxiety response patterns: illus- exploratory behavior. British Journal of Psychology. trating a new multivariate experimental design. Journal of 41(): 50–68. Abnormal & Social Psychology. 60(): 195–204. Berlyne, D.E. 1954. A theory of human curiosity. British Collins, R.P.; Litman, J.A. & Spielberger, C.D. 2004. The Journal of Psychology. 45(): 180-191. measurement of perceptual curiosity. Personality and Berlyne, D.E. 1957. Determinants of human perceptual Individual Differences. 36(): 1127-1141. curiosity. Journal of Experimental Psychology. 53(): Constant, D.; Kiesler, S. & Sproull, L. 1994. What’s Mine Is 399–404. Ours, or Is It? A Study of Attitutdes about Information Berlyne, D.E. 1958. The influence of complexity and nov- Sharing. Information Systems Research. 5(4): 400-421. elty in visual figures on orienting responses. Journal of Cothrel, J.P. 2000. Measuring the success of an online com- Experimental Psychology. 55(): 289–296. munity. Strategy & Leadership. 17-21. Berlyne, D.E. 1960. Conflict, arousal, and curiosity. Cova, B. & Pace, S. 2006. Brand community of convenience New York: McGraw-Hill. products: new forms of customer empowerment - the case Berlyne, D.E. 1966. Curiosity and exploration. Science. "my Nutella The Community". European Journal of 153(): 25–33. Marketing. 40(9-10): 1087-1105.

110 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

Csikszentmihalyi, M. 1990. Flow - the Psychology of share? International Journal of Innovation Manage- Optimal Experience. New York: HarperCollins. ment. forthcoming Deci, E.L. 1975. Intrinsic motivation. New York: Plenum Jeppesen, L.B. & Fredriksen, L. 2006. Why Do Users Con- Publishing Co.. tribute to Firm-Hosted User Communities? The case of Computer-Controlled Music Instruments. Organization Deci, E.L. & Ryan, R.M. 1985. Intrinsic Motivation and Science. 17(1): 45-63. Self-determination in human Behavior. New York: Plenum Press. Joyce, E. & Kraut, R.E. 2006. Predicting Continued Partici- pation in Newsgroups. Journal of Computer-Mediated Dholakia, U.M.; Bagozzi, R. & Pearo, L.K. 2004. A social Communication. 11(): 723-747. influence model of consumer participation in network- and small-group-based virtual communities. International Karau, S.J. & Williams, K.D. 1993. Social loafing: A meta- Journal of Research in Marketing. 21(): 241-263. analytic review and theoretical integration. Journal of Personality and Social Psychology. 65(4): 681-706. Ducheneaut, N. 2005. Socialization in an open source software community: A socio-technical analysis. Computer Karau, S.J. & Williams, K.D. 2001. Understanding individ- Supported Cooperative Work. 14(4): 323-368. ual motivation in groups: The collective effort model. In M. E. Turner (Ed.), Groups at Work: Theory and Research (pp. Ducheneaut, N.; Yee, N.; Nickell, E. & Moore, R.J. 2006. 113-141). Mahwah, NJ: Lawrence Erlbaum Associates. Building an MMO with mass appeal a look at gameplay in world of warcraft. Games and Culture. 1(4): 1-38. Kardaras, D.; Karakostas, B. & Papathanassiou, E. 2003. The potential of virtual communities in the insurance indus- Franke, N. & von Hippel, E. 2003. Satisfying heterogeneous try in the UK and Greece. International Journal of user needs via innovation toolkits: The case of apache secu- Information Management. 23(1): 41-53. rity software. Research Policy. 32(7): 1199-1215. Katz, J. 1998. Luring the Lurkers. Available at Frey, B.S. & Jegen, R. 2001. Motivation Crowding Theory. http://slashdot.org/features/98/12/28/1745253.shtml Journal of Economic Surveys. 15(5): 589-611. Kim, A.J. 2000. Community Building on the Web. Frey, B.S. & Oberholzer-Gee, F. 1997. The Cost of Price Secret Strategies for Successful Online Communi- Incentives: An Empirical Analysis of Motivation Crowding- ties. Berkeley: Addison-Wesley Longman. Out. American Economic Review. 87(4): 746-755. Kollock, P. & Smith, M. 1996. Managing the virtual com- Fueller, J.; Jawecki, G. & Mühlbacher, H. 2007. Innovation mons: cooperation and conflict in computer communities, Creation by Online Basketball Communities. Journal of in: S. Herring (ed) Computer-Mediated Communication: Business Research. 60(1): 60-71. Linguistic, Social, and Cross-Cultural Perspectives. Amster- Gilbert, M. 1989. On Social Facts. Routledge, London, dam, Philadelphia, p.109-128. UK. Koo, D.-M. & Choi, Y.-Y. 2010. Knowledge search and Hadida, A.L. 2009. Motion picture performance: review people with high epistemic curiosity. Computers in Hu- and research agenda. International Journal of Man- man Behavior. 26(): 12-22 agement Reviews. 11(3): 297-335. Kozinets, R.V. 1999. E-tribalized marketing? The strategic Haefliger, S.; Jäger, P. & von Krogh, G. 2010. Under the implications of virtual communities of consumption. Euro- radar: Industry entry by user entrepreneurs. Research pean Management Journal. 17(3): 252-264. Policy. 39(9): 1198–1213. Kozinets, R.V. 2001. Utopian Enterprise: Articulating the Hagel, J., & Armstrong, A. 1997. Net Gain: Expanding Meanings of Star Trek’s Culture of Consumption. Journal Markets Through Virtual Communities. Harvard of Consumer Research. 28: 67-88. Business School Press. Kozinets, R.V. 2007. Inno-tribes: Star Trek as Wikimedia. in Hall, H. & Graham, D. 2004. Creation and recreation: Consumer tribes, B. Cova, R.V. Kozinets & A. Shankar, 194- motivating collaboration to generate knowledge capital in 211. Oxford and Burlington, MA: Butterworth-Heinemann online communities. International Journal of Informa- Lakhani, K.R. & Wolf, R.G. 2005. “Why Hackers Do What tion Management. 24(): 235-246. They Do: Understanding Motivation and Effort in Hemetsberger, A. 2002. Fostering cooperation on the Inter- Free/Open Source Software Projects,” in Perspectives on net: social exchange processes in innovative virtual consumer Free and Open Source Software, J. Feller, B. Fitzgerald, S. communities. Advances in Consumer Research. 29(): Hissam, and K. R. Lakhani, (eds.), Cambridge MA: MIT 354-356. Press, pp. 3-22. Hsu, C.-L. & Lin, J.C-C. 2008. Acceptance of blog usage: Lave, J. & Wenger, E. 1991. Situated Learning: Legiti- The roles of technology acceptance, social influence and mate Peripheral Participation. Cambridge University knowledge sharing motivation. Information & Manage- Press, Cambridge, UK. ment. 45(): 65-74. Lee, G.K. & Cole, R.E. 2003. From a Firm-Based to a Iriberri, A. & Leroy, G. 2009. A Life Cycle Perspective on Community-Based Model of Knowledge Creation: The Online Community Success. ACM Computing Surveys. Case fo the Linux Kernel Development. Organization 41(2): 1-29. Science. 14(6): 633-649. Janzik, L. & Raasch, C & Herstatt, C. 2010. Motivation in Leimeistier, J.M.; Ebner, W. & Krcmar, H. 2005. Design, innovative online communities: Why join, why innovate, shy Implementation, and Evaluation of Trust-Supporting Com-

111 FROM GAME TO PLAY PETER JÄGER • 2010

ponents in Virtual Communities for Patients. Journal of Online Communities: A Proposed Framework. SIGMIS Management Information Systems. 21(4): 101-135. CPR 2007: Proceedings fo the 2007 ACM SIGMIS CPR Conference Global Information Technology Workforce. 153- Ling, K.; Beenen, G.; Ludford, P.; Wang, X.Q.; Chang, K.; 158. Li, X.; Cosley, D.; Frankowski, D.; Terveen, L.; Rashid, A.M.; Resnick, P. & Kraut, R. 2005. Using social psychology Moreland, R.L. & Levine, J.M. 2001. Socialization in orga- to motivate contributions to online communities. Journal nizations and work groups. In: M.E. Turner (Ed.), Groups at of Computer-Mediated Communication. 10(4). work: Theory and research (pp. 69-112). Mahwah, NJ: Lawrence Erlbaum. Litman, J.A. & Spielberger, C.D. 2003. Measuring epistemic curiosity and its diversive and specific components. Journal Muniz, A. M. & O'Guinn, T. C. 2001. Brand community. of Personality Assessment. 80(): 75-86. Journal of Consumer Research. 27(4): 412-432. Litman, J.A. & Jimerson, T.L. 2004. The measurement of Nakamura, J. & Csikszentmihalyi, M. 2002. The Concept of curiosity as a feeling of deprivation. Journal of Personal- Flow, in C.R. Snyder; S.J. Lopez (Eds.): Handbook of Posi- ity Assessment. 82(): 147-157. tive Psychology, p.89-107. Oxford: Oxford University Press. Litman, J.A. 2005. Curiosity and the pleasures of learning: Nambisan, S. 2002. Designing Virtual Customer Environ- Wanting and liking new information. Cognition and ments for New Product Development: Toward a Theory. Emotion. 19(6): 793-814. Academy of Management Review. 27(3): 392-413 Litman, J.A.; Hutchins, T. L. & Russon, R.K. 2005a. Epis- Naylor, F.D. 1981. A state-trait curiosity inventory. Austra- temic curiosity, feeling-of-knowing, and exploratory behav- lian Psychologist. 16(2): 172-183. iour. Cognition and Emotion. 19(): 559–582. Nonaka, I. 1991. The knowledge-creating company. Har- Litman, J.A.; Collins, R.P. & Spielberger, C.D. 2005b. The vard Business Review. 69(6): 96-104. nature and measurement of sensory curiosity. Personality Nonnecke, B. & Preece, J. 2000. Lurker Demographics: and Individual Differences. 39(): 1123-1133. Counting the silent. Proceedings of the CHI 2000. The Litman, J.A. & Silvia, P.J. 2006. The latent structure of trait Hague. ACM. curiosity: Evidence for interest and deprivation curiosity Nonnecke, B. & Preece, J. 2001. Why Lurkers Lurk. Ameri- dimensions. Journal of Personality Assessment. 86(): cas Conference on Information Systems. 2001. 318–328. Nonnecke, B. & Preece, J. 2003. Silent participants: Getting Litman, J.A. & Pezzo, M.V. 2007. Dimensionality of to know lurkers better. In: Lueg, C. & Fisher, D. From Use- interpersonal curiosity. 43(): 1448-1459. net to Cowebs: Interacting with social information spaces. Litman, J.A. 2008. Interest and deprivation factors of epis- Computer Supported Cooperative Work Series. 2003: 110- temic curiosity. Personality and Individual Differ- 132. ences. 44(): 1585-1595. Nonnecke, B.; Preece, J.; Andrews, D. & Voutour, R. 2004. Litman, J.A. 2010. Relationships between measures of I- Online lurkers tell why. Paper presented at AMCIS 2004. and D-type curiosity, ambiguity tolerance, and need for New York, NY. closure: An initial test of the wanting-liking model of infor- Pichler. E.A. & Hemetsberger, A. 2008. Driven by Devotion mation-seeking. Personality and Individual Differ- – How consumers Interact with their Objects of Devotion. ences. 48(): 397-402. Advances in Consumer Research. 35: 439-443. Loewenstein, G. 1994. The psychology of curiosity: A re- Porter, C.E. & Donthu, 2008. Cultivating Trust and Harvest- view and reinterpretation. Psychological Bulletin. 116(): ing Value in Virtual Communities. Management Science. 75-98. 54(1): 113-128. Lüthje, C.; Herstatt, C. & von Hippel, E. 2005. User- Preece, J. 1999a. Empathic communities: balancing emo- innovators and “local” information: The case of mountain tional and factual communication. Interacting with Com- biking. Research Policy. 34(): 951-965. puters. The Interdisciplinary Journal of Human– Mason, B. 1999. Issues in virtual ethnography. In Proc. Computer Interaction. 12(1): 63–77. Ethnographic Studies in Real and Virtual Environments: Preece, J. 1999b. Empathy online. Virtual Reality. 4(): 1– Inhabited Information Spaces and Connected Communities. 11. Edinburgh. 1999: 61-69. Preece, J. 2004. Etiquette Online: From Nice to Necessary. McAlexander, J.H.; Schouten, J.W. & Koenig, H.F. 2002. Communications of the ACM. 47(4): 56-61. Building Brand Communitiy. Journal of Marketing. 66(Jan): 38-54. Preece, J.; Nonnecke, B. & Andrews, D. 2004. The top 5 reasons for lurking: Improving community experiences for McGrath, J.E. & Hollingshead, A.B. 1994. Groups Inter- everyone. Special Issue of Computers in Human acting with Technology: Ideas, Issues, Evidence, Behavior: An Interdisciplinary Perspective. 20(2). and an Agenda. Newbury Park, CA: SAGE Publications. McWilliam, G. 2000. Building stronger brands through Rafaeli, S.; Ravid, G. & Soroka, V. 2004. De-Lurking in Virtual Communities: A Social Communication Network online communities. Sloan Management Review. 41(3): Approach to Measuring the Effects of Social and Cultural 43–54. Capital. HICSS vol. 7, pp.70203, Proceedings of the 37th Moore, T.D. & Serva, M.A. 2007. Understanding Member Annual Hawaii International Conference on System Sci- Motivation for Contributing to Different Types of Virtual ences (HICSS'04) - Track 7, 2004

112 CURIOUS LURKERS • LINKING I/D-TYPE CURIOSITY TO CONTRIBUTION BEHAVIOR IN ONLINE COMMUNITIES

Rashid, Al M.; Ling, K.; Tassone, R.D.; Resnick, P.; Kraut, Spielberger, C.D. & Starr, L.M. 1994. Curiosity and explora- R. & Riedl, J. 2006. Motivating Participation by Displaying tory behavior. In H.F. O’Neil Jr., & M. Drillings (Eds.), Mo- the Value of Contribution. Proceedings of the SIGCHI tivation: theory and research. Hillsdale, NJ: Lawrence Erl- conference on Human Factors in computing systems. ACM, baum Associates. New York. 955-958. Straus, S.G. & McGrath, J.E. 1994. Does the Medium Mat- Ren, Y; Kraut; R. & Kiesler, S. 2007. Applying Common ter? The Interaction of Task Type and Technology on Identity and Bond Theory to Design of Online Communi- Group Performance and Member Reactions. Journal of ties. Organization Studies. 28(03): 377–408. Applied Psychology. 79(): 87–97. Ren, Y.; Harper, F.M.; Drenner, S.; Terveen, L.; Kiesler, S.; Tafjel, H. 1978. Differentiation between social groups. Riedl, J. & Kraut, R.E. XXXX. Increasing Commitment to London: Academic. Online Communities: Designing Theory. Under Review at MIS Quarterly. Tietz, R.; Morrison, P.D.; Lüthje, C. & Herstatt, C. 2005. The process of user-innovation: A case study on user innova- Rheingold, H. 1993. The Virtual Community (1st. ed.). tion in a consumer goods setting. International Journal Addison-Wesley Pub. Co. of Product Development. 2(): 321-338. Ridings, C.M.; Gefen, D. & Arinze, B. 2002. Some antece- von Hippel, E. 1988. The Sources of Innovation. Oxford dents and effects of trust in virtual communities. Journal of University Press. Strategic Information Systems. 11(): 271-295. von Hippel, E. 2005. Democratizing innovation. Cam- Roberts, J.A.; Hann, I-H. & Slaughter, S.A. 2006. Under- bridge. MA: MIT Press. standing the Motivations, Participation, and Performance of Open Source Software Developers: A Longitudinal Study of von Krogh, G. 1998. Care in knowledge creation. Califor- the Apache Projects. Management Science. 52(7): 984- nia Management Review. 40(): 133-154. 999. von Krogh, G. 2002. The communal resource and informa- Robinson, T.E. & Berridge, K.C. 2003. Addiction. Annual tion system. Journal of Strategic Information Sys- Review of Psychology. 54(): 25-53. tems. 11(): 85-107. Rowley, J.; Kupiec-Teahan, B. & Leeming, E. 2007. Cus- von Krogh, G.; Haefliger, S.; Spaeth, S. & Wallin, M. 2010. tomer Community and Co-creation: A Case Study. Market- Open Source Software: A Review of Motivations to Con- ing Intelligence & Planning. 25(2): 136-146. tribute. ETH Zurich Working Paper. Ryan, R.M. & Deci, E.L. 2000. Intrinsic and extrinsic moti- Wasko, M.M. & Faraj, S. 2000. “It is what one does”: why vations: Classic definitions and new directions. Contempo- people participate and help others in electronic communities rary Educational Psychology. 25(): 54-67. of practice. Journal of Strategic Information Systems. 9(): 155-173. Ryan, R.M. & Deci, E.L. 2002. An overview of self- determination theory. In E. L. Deci & R. M. Ryan (Eds.), Wasko, M.M.& Faraj, S. 2005. Why Should I Share? Exam- Handbook of self-determination research (pp. 3-33). Roches- ining Social Capital and Knowledge Contribution in Elec- ter, NY: University of Rochester Press. tronic Networks of Practice. MIS Quarterly. 29(1): 35-57. Schau, H. J. & Gilly, M. C. 2003. We are what we post? Self- Webster, D.M. & Kruglanski, A.W. 1994. Individual differ- presentation in personal Web space. Journal of Con- ences in need for cognitive closure. Journal of Personality sumer Research. 30(3): 385-404. and Social Psychology. 67(): 1049-1062. Shah, S.K. 2006. Motivation, Governance, and the Viability Wenger, E. 1998. Communities of Practice – Learning as a of Hybrid Forms in Open Source Software Development. Social System. Systems Thinker. 9(5): 3-7. Management Science. 52(7): 1000-1014. Williams, R.L. & Cothrel, J. 2000. Four Smart Ways To Run Shankar, V.; Smith, A.K. & Rangaswamy, A. 2003. Cus- Online Communities. Sloan Management Review. tomer satisfaction and loyalty in online and offline environ- Summer: 81-91. ments. International Journal of Research in Market- Winkielman, P. & Berridge, K.C. 2003. Irrational wanting ing. 20(2): 153-175. and subrational liking: How rudimentary motivational and Spielberger, C.D. & Butler, T.F. 1971. On the relationship affective processes shape preferences and choices. Political between anxiety, curiosity, and arousal: a working paper. Psychology. 24(): 657-680. Unpublished manuscript. Florida State University. Wittenbaum, G.M.; Hubbell, A.P. & Zuckerman, C. 1999. Spielberger, C.D. 1975. Anxiety: state-trait-process. In C.D. Mutual enhancement: Toward an understanding of the Spielberger & I.G. Sarason (Eds.), Stress and anxi- collective preference for shared information. Journal of ety.Washington: Hemisphere/Wiley. Personality and Social Psychology. 77(): 967-978. Spielberger, C.D.; Peters, R.A. & Frain, F. 1976. The State- Yu, J.; Jiang, Z. & Chan, H.C. 2007. Knowledge Contribu- Trait Curiosity Inventory. Unpublished manual, Univer- tion in Problem Solving Virtual Communities: the Media- sity of South Florida, Tampa. tion Role of Individual Motivations. Proceedings of the 2007 ACM SIGMIS CPR conference on Computer person- Spielberger, C.D.; Peters, R.A. & Frain, F. 1981. Curiosity nel research: The global information technology workforce. and anxiety. In H. G. Voss & H. Keller (Eds.), Curiosity 144-152. research: Basic concepts and results. Weinheim, Federal Republic of Germany: Beltz.

113 FROM GAME TO PLAY PETER JÄGER • 2010

Zuckerman, M. & Litle, P. 1986. Personality and curiosity about morbid and sexual events. Personality and Indi- vidual Differences. 7(): 49-56.

114 DIPL.-ING. PETER MICHAEL JÄGER • CURRICULUM VITAE

Date of birth March 24, 1979 Place of birth Überlingen, Germany Nationality German

Contact Englischviertelstr. 52 • 8032 Zürich • Switzerland m +41 79 434 43 02 e [email protected]

Personal Profile and Objectives Strengths Young professional with sound experience, enthusiastic, committed, eager to learn, deci- sive, yet helpful, caring and cooperative. Excellent communication, presentation, writing and project management skills. Strong analytical competencies. Fast grasp, thorough technological understanding. Team oriented, motivating, prepared to take the lead. En- trepreneurial thinking and experience. Interests Photography (Freelancer: spiegelschlag.ch), Travelling, Cooking, Social Media, Snow- boarding, Skiing, Sailing, Climbing, Swimming, Inline-Skating

Education 10.2006 – 12.2010 ETH Zurich. Chair of Strategic Management and Innovation, Prof. Georg von Krogh. Zu- rich, Switzerland: Doctorate Research on strategy and innovation, user entrepreneurship and community management; publications in international journals and presentations at international conferences – fostering analytical thinking, information refining, and communication skills 10.1999 – 8.2005 University of Stuttgart. Stuttgart, Germany: Engineering Diploma in Mechatronics Two semester thesis on software strategy and development. Diploma thesis on strategies for technology transfer and management; major in Technology Management and Controls Engineering – establishing a sound background in engineering combined with management science

Practical Experience 10.2010 – today Global Procurement Analyst Trinity Procurement GmbH / SABMiller plc. Zug, Switzerland Support project execution within Business Development as key analyst 6.2006 – 8.2010 Research Assistant ETH Zurich. Chair of Strategic Management and Innovation, Prof. Georg von Krogh. Zurich, Switzerland Lecturing and tutoring; IT coordinator; responsible for corporate design; administration on chair and department level; coordination and communication with various stakeholder groups as assistant to study delegate and head of department – expanding project management, communication and coordination skills; engaging in stakeholder management; establishing an extensive network of former students and colleagues 11.2005 – 4.2006 Post Graduate Intern Consultant Roland Berger Strategy Consultants. Munich, Germany Data gathering, analysis and interpretation; communication with customers; identification of optimization potential; development of implementation concepts – excellent insights into various corporate structures; knowledge about professional service firms; delivering conceptual work under extremely stressful conditions in a team effort 1.2005 – 9.2005 Freelance Field Engineer for Wagner Systems, Inc. Phone support, field service, and commissioning in Eastern Europe – exploitation of acquired skills; expanding independent work experience and relationship management 1.2004 – 12.2004 Controls Application Engineer Wagner Systems, Inc. Carol Stream, USA System design of large industrial coating equipment; programming; field support; commissioning; training of customers and employees; gatekeeper to and technology transfer from European headquarter – international project management involving US, D, CH, H, UK, China; autonomous working; resource allocation; assembling teams and taking the lead in critical situations when decisions are vital for project success; exposition to various cultures 9.2003 – 12.2003 Intern Heidelberger Druckmaschinen AG. Wiesloch, Germany Redesign of technology introduction and quality management processes for new production unit; reorganization of rework evaluation to allocate costs; design and implementation of a software tool to manage data integrity across departments to economize procurement – enhancing communication skills; expertise in overcoming resistance to change 12.2000 – 9.2003 Junior Research Assistant Fraunhofer Institute for Manufacturing Engineering and Automation IPA. Stuttgart, Ger- many Design of electrical and mechanical components for service robots; assistance with system integration – gaining insights into applied research and prototype development; experience in writing product documentation 2.2000 – 3.2000 Intern Rohwedder AG. Markdorf, Germany Acquisition of basic production skills; inspiring others to collaborate on joint projects 1.2000 – today Co-founder and Managing Partner ViaVideo GbR – Video Production. Markdorf, Germany Initiation of foundation, establishment of the company; executive producer; customer relationship management; procurement; in-house R&D for production equipment – deal- ing with legal issues; performing financial reporting and budgeting; motivating others to support own ideas; assembling and managing production crews with voluntary outsiders; cooperating in a team of partners 1997, 1998, 1999 Vacation Job MTU Friedrichshafen GmbH. Friedrichshafen, Germany Shipping and handling of large engines – further insights into logistics; interpreter Eng- lish and French; coordinating work with others; acting with increasing responsibility 11.1995 – 6.1998 Part-time Employee J. Wagner GmbH. Markdorf, Germany Several duties in logistics department – first exposition to corporate culture; managing stock; readiness to take on responsibility; enjoying to work with people of various background and education

Language Skills German (mother tongue) English (fluent) French (basic)

IT Skills OS MAC OS X (professional); Windows (advanced) Office / DTP Keynote, Word, Excel (pro); PowerPoint, InDesign, Photoshop, Illustrator (advanced) ERP / CAD SAP, AutoCAD (basic) Programming Siemens Simatic S5/S7 (professional); C, Visual C++, Delphi, Modula-2, Lisp, Pascal (basic) Database Filemaker Pro, MySQL, Access (advanced) Web WCMS (Typo3, SilvaCMS), HTML, CSS (advanced) Publications P. Jäger “From Game to Play. Curiosity, Innovation & Entrepreneurship in Online User Communities.” ETH Zu- rich Dissertation (under review) P. Jäger “Curious Lurkers. Linking I/D-type Curiosity to Contribution Behavior in Online Communities”. ETH Zurich Working Paper S. Haefliger, P. Jäger, G. von Krogh “Under the Radar: Industry Entry by User Entrepreneurs in the Case of Ma- chinima”. Research Policy 39: 1198-1213 H.C. Steckling & P. Jäger “Switzerland – Intellectual Capital Center of the Future?”. Swiss-American Chamber of Commerce Yearbook 2010/2011 P. Jäger, R. Einstein, P. Hertz, G. von Krogh “Game on: Die Videospiel-Industrie startet durch”. io new management 0910, Sept. 2009 S. Haefliger, P. Reichen, P. Jaeger, G. von Krogh “Modding as rating behavior in virtual communities: The case of Rooster Teeth Productions”. Online Communities and Social Computing, Lecture Notes in Computer Science, Vol. 5621, Springer, 2009 P. Jäger, G. von Krogh, S. Haefliger “A Directing Audience: How Specialized Feedback in a Virtual Community of Consumption Stimulates New Media Production”. ETH Zurich Working Paper Various contributions to OBISness, the quarterly magazine of the Organization of Business and Industrial Engineering Students at ETH Zurich

Presentations and Speeches P. Jaeger, S. Haefliger, G. von Krogh. “Online communities of consumption or production? The case of Rooster Teeth Productions”, presentation at the Academy of Management Meeting (AoM) on 8-11 August, 2009 in Chicago, IL G. von Krogh, S. Haefliger, P. Jaeger “Learning In Online Communities. The Case Of Machinima”, Presentation at the European Academy of Management Meeting (EURAM) on 11-14 May, 2009 in Liver- pool P. Jaeger, S. Haefliger, G. von Krogh “Emergent Specialization in a Large Online Customer Community. The Case of Rooster Teeth Productions”, presentation at the DIG Seminar Series at Politecnico di Milano, April 21, 2009 in Milan S. Haefliger, P. Reichen, P. Jaeger, G. von Krogh “Modding as rating behavior in virtual communities: The case of Rooster Teeth Productions”, presentation at the Human Computer Interaction Confer- ence (HCII) on 19-24 July, 2009 in San Diego, CA G. von Krogh, S. Haefliger, P. Jäger “User Innovation Beyond Market Barriers: The Case Of Machinima”, presentation at Academy of Management Meeting (AoM) on 11-13 August, 2008 in Anaheim, CA G. von Krogh, S. Haefliger, P. Jäger “User Innovation Beyond Market Barriers: The Case Of Machinima”, presentation at the 6th Annual International Open and User Innovation Workshop. Harvard Business School and Massachusetts Institute of Technology on 4-6 August, 2008 in Bos- ton, MA G. von Krogh, S. Haefliger, P. Jäger “User Innovation Beyond Market Barriers: The Case Of Machinima”, presentation at the European Academy of Management (EURAM) on 14-17 May, 2008 in Ljubl- jana G. von Krogh, S. Haefliger, P. Jaeger “User-innovation beyond market boundaries: The case of Machinima”, presentation at the Strategic Management Society Conference (SMS) on 14-17 October, 2007 in San Diego, CA Academic Services Memberships Academy of Management (AoM) European Academy of Management (EURAM) Strategic Management Society (SMS) Honorary Member Organization of Business and Industrial Engineering Students (OBIS) Ad-hoc Reviewer Management Information Systems Quarterly (MISQ) Journal of Strategic Information Systems (JSIS) Die Betriebswirtschaft (DBW) AoM Division of Technology and Innovation Management

Teaching Experience ETH Zurich Strategic Management, Master Level Industry and Competitive Analysis, Master Level User Innovation, Master Level Corporate Strategy, Master Level Marketing, Master Level Basics of Scientific Work, Master Level Tutoring Master Thesis (25) Tutoring Bachelor Thesis (2) Tutoring Semester Thesis (1)

Supervised Thesis Master Thesis Document and Content Management Systems. A case study. September 2010 How to improve customer value in public transports through information systems. June 2010 Innovation Behavior in the Swiss MedTech Industry. May 2010 Dominant Designs and the Role of the Market Leader in Upcoming, Complementary Indus- tries – A Case Study on Guided Surgery for Dental Treatment Application. May 2010 New Marketing Concept for Customized Production Communication Software. June 2010 Entwicklung eines Modells zur Performancemessung einer Professional Service Firm. July 2009 Lean Administration as a Resource for Temporary Competitive Advantage. July 2009 RAMS Methodologies and Innovation Process. July 2009 How to improve knowledge sharing in technology alliances between companies. July 2009 Informationsbedürfnisse von Kunden im Zusammenhang mit Leistungssystemen. June 2009 Businessplan-Modellanalyse von Klein- und Mittelgrossunternehmen. May 2009 Hybrid Virtual Communities: Classic Virtual Communities with Social Network Site-Features – Exploring Mass Communication Patterns and the Impact of User-to-User Interaction on Participation. March 2009 Analyse der Outsourcingpotentiale von Transportdienstleistungen und Ausarbeitung von Optimierungskonzepten bei einem Zementhersteller. March 2009 Strategischer Nutzen und Positionierung eines Industrieunternehmens in der Bekämpfung des Klimawandels. March 2009 Fortune 500 companies in Second Life – Activities, their success measurement and the satisfaction level of their projects. December 2008 Szenariomodellierung zur Selektion der optimalen Kooperationspartner für die Lenzerheide Bergbahnen AG. October 2008 Coping with Radical Change in the Elevator Industry. September 2008 Konzept für die Einführung eines integrierten Technology Intelligence Ansatzes. July 2008 Developing a model to detect focus areas for innovation within the industry of private banking. June 2008 Erzielung von OEM-Bindungen in der Kompressortechnik. June 2008 Improvement of Supplier’s Product Reliability in the Product Creation Process. March 2008 Game-Based Learning and its Application in Business Environments. February 2008 Repositioning the Offer of an Electrical Product and System Supplier to the Oil and Gas In- dustry. December 2007 Studie „Recyclingbeton im Raum Zürich“. Von der Marktanalyse über Unternehmensstrate- gie bis zur Analyse des Beschaffungsprozesses. July 2007 Diploma Thesis Kriterien für die Kommerzialisierung von Machinima-Produkten. April 2007 Bachelor Thesis User Innovation in the Video Game Industry. Strategies of Firms to target added value to the Product by Users and to build up a Source of Innovation. February 2010 Entwicklung eines SOP-Konzeptes für die Produktion von Primärpackmitteln und die Eig- nung von SOPs zum Wissenstransfer in Unternehmen. August 2008 Semester Thesis Computed Chairside Implantology CCI – Marktanalyse, Akzeptanzstudie und Eintritts- strategie. April 2009