Exploring the design and effects of internal knowledge markets 9783319145167, 9783319145174, 3319145169, 3319145177

Table of contents :
Research objectives --
Review of the literature.- Research methodology --
Results --
Implications for theory and practice --
Contributions to the 2020 perspective --
Conclusion.

Citation preview

SPRINGER BRIEFS IN DIGITAL SPACES

Hind Benbya

Exploring the Design and Effects of Internal Knowledge Markets

SpringerBriefs in Digital Spaces Series editor Ahmed Bounfour, Orsay, France

More information about this series at http://www.springer.com/series/10461

Hind Benbya

Exploring the Design and Effects of Internal Knowledge Markets

13

Hind Benbya Montpellier Business School Montpellier France

ISSN  2193-5890 ISSN  2193-5904  (electronic) SpringerBriefs in Digital Spaces ISBN 978-3-319-14516-7 ISBN 978-3-319-14517-4  (eBook) DOI 10.1007/978-3-319-14517-4 Library of Congress Control Number: 2014958965 Springer Cham Heidelberg New York Dordrecht London © The Author(s) 2015 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper Springer International Publishing AG Switzerland is part of Springer Science+Business Media (www.springer.com)

Contents

1 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 Review of the Literature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Traditional Approaches to Problem-Solving . . . . . . . . . . . . . . . . . . . 3 2.1.1 External Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1.2 Internal Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.2 Markets as a Key Mechanism for Problem-Solving. . . . . . . . . . . . . . 4 2.2.1 Differences Between Traditional Approaches versus Internal Knowledge Markets . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.3 Markets and Prices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3 Research Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.1 Internal Markets Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 4.1.1 Idea Markets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 4.1.2 Prediction Markets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 4.2 Mini Cases of Internal Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.2.1 Allianz “Idea Market” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.2.2 SAP “Knowledge Market”. . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4.3 Design Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.3.1 Incentives Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.3.2 Governance Mechanisms. . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.3.3 Lessons Learned. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 4.4 Internal Knowledge Markets Effects. . . . . . . . . . . . . . . . . . . . . . . . . 22 4.4.1 Case Analysis: Effects on Individual Performance . . . . . . . . 24 4.4.2 Experiments Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

v

vi

Contents

4.5 Analysis and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 4.5.1 Experimentation Phase: Learning About Markets and Communities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 4.5.2 Thrill of Competition: Markets Dominate. . . . . . . . . . . . . . . 34 4.5.3 Communalism Phase Again: Disillusionment with Inequality/Unfairness. . . . . . . . . . . . . . . . . . . . . . . . . . . 35 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5 Implications for Theory and Practice. . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.1 Implications About Incentives in Knowledge Markets . . . . . . . . . . . 39 5.2 Implications of Identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.3 Implications for Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 6 Contributions to the 2020 Perspective. . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

SpringerBriefs on Digital Spaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Introduction

The results of this report are based on several studies conducted by the author in collaboration with Tanya Menon, Marshall Van Alstyne, and Nassim Belbaly. References to these studies are detailed in the bibliography.

Many large firms have noted that they could do more to support the distributed creativity of their employees and improve their access to a common base of knowledge. First-generation attempts to manage knowledge and support innovation were mainly based on centralized approaches. Despite their potential, the top-down approach to both content elicitation and validation deployed in these systems has left a large community base with unsolved knowledge problems (Benbya 2008). The emerging literature on open innovation and Crowdsourcing demonstrates the value of openness for innovation and problem solving (Chesbrough 2003; Surowiecki 2004). One way to bring the benefits of open innovation to firms is through market mechanisms. A market is typically an IT supported platform that helps to connect users to experts and expertise within the organization. They provide opportunities for matching knowledge seekers with sources; facilitate reuse of exiting knowledge; and help find solutions to unsolved problems. Preliminary studies on this emerging area of research highlight some concrete benefits of how and when markets perform compared to established alternatives (Jeppesen and Lakhani 2010; Benbya and Van Alstyne 2011). But they also raise several new challenges related to incentives design, fairness, regulation, and governance. These preliminary studies, however, remain limited to historical data projects from external markets such as InnoCentive and TopCoder which limit further investigation of market dynamics, their benefits, and challenges in internal settings. These challenges inspire us to investigate the design and implementation of market mechanisms to explore how they can support knowledge and innovation management within firms. To pursue such exploration we rely on a multimethod design, integrating qualitative and quantitative cases with experimentation. To be able to conduct experiments we design an internal knowledge market prototype and conduct preliminary tests of its use. The remainder of this paper is organized as follows. The following chapter summarizes the research objectives and reviews the literature on this emerging vii

viii

Introduction

topic. Chapter 2 discusses the methodology used to investigate internal markets dynamics. Chapter 3 presents the study results, its managerial implications, and concludes with contributions to the 2020 perspective.

References Benbya, H. (2008). Knowledge management systems implementation: Lessons from the silicon valley. Oxford, UK: Chandos Publishing. Chesbrough, H. (2003). Open innovation: the new imperative for creating and profiting from technology. Boston: Harvard Business School Press. Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few and how collective wisdom shapes business, economies, societies and nations. New York: Random House. Jeppesen, L. B., & Lakhani, K. (2010). Marginality and problem-solving effectiveness in broadcast search. Organization Science, 21(5), 1016–1033. Benbya, H., & van Alstyne, M. (2011). How to find answers within your company. MIT Sloan Management Review, 52(2), 65–75. Benbya, H., & Menon, T. (2015). Exploring Transitions between Communal- and Market-based Knowledge Sharing: Collective Action to Resolve Uncertainty. Working paper.

Chapter 1

Research Objectives

The “Wisdom of Crowds” holds that companies can’t plan everything from the top. It assumes that firms should use external as well as internal ideas … as they look to advance their technology, products and practices. But planning from the bottom requires knowing when crowds are actually wise. Managers inside firms have to both select high value problems to be solved, and, depending on their decomposability, choose to have them solved internally or externally. Leveraging external sources of innovation whether from individuals, ­customers, suppliers or universities has attracted extensive research and has been found to influence firm performance in creating products and services. Firms can rely on a variety of modes to profit from external knowledge and ideas. Lead user workshops (Herstatt and von Hippel 1992; Urban and Von Hippel 1988) and sponsored user communities (e.g. Jeppesen and Frederiksen 2006) are common and popular in a wide range of industrial sectors. More recently, firms started to explore the potential of knowledge markets for accessing distributed knowledge and solving a variety of organizational challenges. A knowledge Market is an IT platform for matching information seekers and sources, together with the material and social incentives to encourage effective trading. The advantage of markets over other alternatives is that many ­contributors with a diverse pool of skills can enhance the chances of finding solutions to d­ ifferent sometimes unconventional and perhaps even breakthrough problems (Joshi and Sharma 2004). Organizations thus increasingly draw on markets by ­collaborating with external problem solvers to enhance their innovation c­ apabilities. InnoCentive.com, for example, routinely attracts roughly 250 individuals to d­ ifferent R&D related scientific problems on behalf of its clients. It allows corporations, government agencies, and non-profit organizations to publish a problem related to research and development to a large community of innovators around the world. The best solutions are then rewarded with prize money ranging from $5,000 to $1,000,000. Recent academic and popular literature suggests a significant potential of knowledge markets for solving a variety of organizational challenges. Jepperson and Lakhani (2010), for example, find that broadcasting problem-solving, through an external market, to a large base of participants with different perspectives is

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_1

1

2

1  Research Objectives

associated with novel solution creation. Terwiesch and Xu (2006) examine their potential in eliciting a variety of innovation outcomes. These preliminary studies, however, based on various experiments and historical data projects from external Innovation Markets such as InnoCentive and TopCoder, missed the o­ pportunity to bring these benefits within firms. They are all geared to unsticking external knowledge but leave these principles outside the company walls. The objective of this research is to explore how we can move markets inside firms. Specifically, we strive to understand the design guidelines and effects associated with internal knowledge markets. To understand how to move markets inside firms, we studied the benefits and costs of traditional versus market based approaches. We investigated firms that adopted internal markets, and we designed a working prototype to test its various effects. As a result we should get insights about how to design an efficient internal market and understand its various effects.

References Herstatt, C., & von Hippel, E. (1992). From experience: Developing new product concepts via the lead user method: a case study in a “low-tech” field. The Journal of Product Innovation Management, 9(3), 213–221. Jeppesen, L. B., & Frederiksen, L. (2006). Why do users contribute to firm-hosted user communities? The case of computer-controlled music instruments. Organization Science, 17(1), 45–64. Jeppesen, L. B., & Lakhani, K. R. (2010). Marginality and problem-solving effectiveness in broadcast search. Organization Science, 21(5), 1016–1033. Joshi, A. W., & Sharma, S. (2004). Customer knowledge development: Antecedents and impact on new product performance. Journal of Marketing, 68(4), 47–59. Terwiesch, C., & Xu, Y. (2006). Innovation contests, open innovation, and multi-agent problem solving. Management Science, 54(9), 1529–1543. Urban, G. I., & von Hippel, E. (1988). Lead user analyses for the development of new industrial products. Management Science, 34(5), 569–582.

Chapter 2

Review of the Literature

2.1 Traditional Approaches to Problem-Solving 2.1.1 External Sources Drawing on external sources (i.e., individuals, firms, suppliers or universities) to innovate has been the subject of a rich body of early research. Researchers considered the role of innovation created outside the firm by individuals. Empirical studies show that many users (up to 40 %) engage in developing or modifying products for many reasons. Sometimes, users may innovate if they want something that is not available on the market or they may value the process of innovating because of the enjoyment or learning that it brings them. Recent studies find e­ vidence that some users, “entrepreneurial users” have also commercialized their own innovations in a wide range of industries (i.e., medical devices, sporting equipment, stereo-components, new media) (Shah and Tripsas 2012). In some cases, the firm is explicitly present and collaborating in organizing or producing the innovation—with its customers or other partners—for its own financial benefit, as with co-creation (Prahalad and Ramaswamy 2004). In other cases, the firm is explicitly absent. This innovation is often referred to as user-driven innovation (von Hippel 2007). The largest category of such research emphasizes the creation of innovations by users who utilize their own scarce knowledge to provide utility for their own benefit (Bogers and Bastian 2010), which may be harnessed for the benefit of firms (Morrison and von Hippel 2000), or in direct competition with firms (Baldwin and von Hippel 2011). The importance of user innovations spans a variety of industries. While companies operating in a resource- and expertise-intensive industry can only rely on highly skilled professionals working for a research lab or a university to innovate (Terwiesch and Ulrich 2009), most other industries can benefit from user innovations. Initially, scholars found evidence of users modifying or creating physical goods, mostly in manufacturing (von Hippel 1986). More recent research finds user innovation to be also significant in and across industries services (Oliveira and von Hippel 2011). Among user innovators © The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_2

3

4

2  Review of the Literature

the concept of lead users remains a significant source of new product development. Lead users are people or firms that have advanced needs for products or ­services that are not being met by other companies. They are ahead of the majority of users in their populations with respect to an important market trend and have either to tolerate their unmet needs or innovate themselves to address them. They provide consequently the basis for products manufacturers could commercialize (von Hippel 1986, 2007). Still another approach to sense innovation opportunities and solve a variety of organizational challenges can be found in external markets.

2.1.2 Internal Sources It seems obvious that companies should use the knowledge possessed by their employees for solving organizational challenges, but few organizations actually do so (van Dijk and van den Ende 2002). Research shows that employees with promising ideas often lack the motivation or a channel to communicate them (Burt 2004). Many firms, however, recognize that innovation isn’t about the Chief Innovation Officer or even the CEO mandating what milestones R&D or Engineering must pursue or achieve. Effective innovation begins with breaking down silos that separate departments, divisions or teams—and encouraging, even welcoming participation from across the organization. Conventional participative techniques to problem-solving and innovation generation consist in running brainstorming sessions, using suggestion boxes or workshops where a diverse group of participants generate a large number of opportunities (Osborn 1963). The proliferation of internet-based technologies suggests new methods for accessing distributed knowledge as they enable the connection of a vast number of participants. One of these emerging methods is to rely on market-mechanisms. Markets provide a platform for aggregating dispersed information as they leverage the collective wisdom of the organization.

2.2 Markets as a Key Mechanism for Problem-Solving The market system has proved throughout history as a remarkably effective means for promoting efficient, cooperative interactions among entities with diverse knowledge, skills, and goals. Despite the long history of markets as an effective means of exchange, bringing market-based mechanisms inside a particular firm has not gained momentum until recently. Fueled by the promise of improving f­orecasting, decision-making, risk management, innovation and knowledge management, knowledge markets have been adopted recently in a wide range of firms. Large organizations that have a wealth of information and talent often have difficulty accessing it. The larger and more segmented the company, the harder it

2.2  Markets as a Key Mechanism for Problem-Solving

5

gets to match people to problems. Managing information flows laterally and vertically, across specialties, locations and countries, has proven particularly challenging. Managing the growing demands for expertise in diverse contexts requires decentralizing information exchange by providing access to diverse sources, not just the central repository. Meeting unanticipated needs also implies providing access to untapped knowledge. Markets provide the most common form of such decentralization. Consider the following Market benefits: • • • •

Enabling buyers and sellers to find each other Allowing individuals to make the kind of consumption choices they prefer Establishing accurate prices based on supply and demand Validating (or rejecting) product and service offerings

A knowledge Market is an IT platform for matching information seekers and sources, together with the material and social incentives to encourage effective trading. Now that the Internet has linked us all so closely, we need better mechanisms to help us find “who knows what” and to reward people who provide valuable information. The size of the problem—that of creating and moving information to the people who need it—is so large that applying market principles is one of the best ways to find solutions. Markets provide opportunities for matching knowledge seekers with sources. In addition to their role in solving traditional knowledge problems, markets provide a good match to help generate, combine, and rank ideas and also help develop new solutions, products, and services. Markets are great for facilitating the reuse of existing information, for creating new information to address previously unsolved problems, and for effectively using information resources, including peoples’ time.

2.2.1 Differences Between Traditional Approaches versus Internal Knowledge Markets We find knowledge sharing and innovation using market mechanisms to require fundamentally different design logic than traditional hierarchies (Fig. 2.1). First-generation attempts to manage knowledge and innovation relied mainly on centralized Knowledge Management Systems (KMS). In this top-down approach, a central authority prioritizes development to address mainstream problems based on two assumptions: (1) knowledge needs are predictable and (2) experts rather than peers handle knowledge elicitation and validation. Both assumptions have limitations. Unpredictable events routinely affect business. The banking crisis of 2008 brought into sharp relief problems whose answers did not yet exist. KMS can only contain historical facts. They cannot contain solutions for issues that have not occurred. When Japan experienced a magnitude 7 earthquake, Oyo Corporation used its knowledge market to rapidly solve problems by connecting people to

6

2  Review of the Literature

Fig.  2.1  Traditional versus emerging approaches to managing knowledge and innovation (Benbya and Van Alstyne 2011)

people rather than force employees to wade through a mountain of paper that would have stacked “as high as Mt. Fuji.” Relying on experts to provide all answers works when they have capacity to serve all requests—in addition to doing their own work. But, when the volume of challenges exceeds capacity, experts become bottlenecks. Problems that would admit to an adequate solution from a peer go unanswered while waiting for a perfect solution from an expert. Further, because knowledge capture is costly, organizations selectively package only a fraction of their best practices. This leaves many workers with unacknowledged needs. Not surprisingly, centralized systems produce mixed results. One audit and consulting firm we studied spent more than two years codifying best practices to place on its KMS. Yet, giving the system to more than 1,500 consultants left only 130 who could find useful content. By contrast, an internal market is the information equivalent of just-in-time inventory. The platform helps people pull information and contacts when needed. Meeting unanticipated needs implies providing access to untapped knowledge, and it means providing access to diverse sources not just the central repository. At NTT Software, a year and a half after introducing a knowledge market, queries posted to the Q&A forum received answers in as few as 5 min. Employees provided faster customer service; best practices spread across software engineers and sales reps; and redundant work fell by 9,000 h. Now, executives tend to think of knowledge markets as a powerful mechanism. But its implementation inside firms still remains tricky. Executives need to c­ onsider carefully which design features to select. Specifically, they need to c­onsider who can provide answers and comments, whether to use free or price-based ­mechanisms, and which properties the market platform should have.

2.2  Markets as a Key Mechanism for Problem-Solving

7

Some external market applications, for example, apply a fee and provide a handful of selected experts who may provide a solution to a knowledge ­problem. Take, for example, Experts-Exchange, a knowledge market through which a user pays a fee to access the knowledge base, post a technology problem and solicit solutions from experts. The experts are rewarded with points that can be ­converted to prizes and other benefits. Other applications of knowledge markets provide free services. Yahoo Answers and Answer iN, for instance, are services open to p­ articipation to any registered user and questions may be answered by both ­subject experts and ­novices. These services, at the difference of the former ­knowledge markets allow participants to exchange discussions, comments, advice and o­ pinions on a variety of topics but they lack a mechanism that filters the correct answers from the incorrect ones. Internal markets are environments where users trade their knowledge via price mechanisms. Although such markets have existed for years (Hayek 1945), their use to facilitate knowledge transfer and innovation inside organizations is ­relatively new. Early adopters such as Infosys, Siemens, McKinsey and Eli-Lilly have proven their value for managing information flows relative to traditional ­solutions. But complications have revealed implementation challenges.

2.3 Markets and Prices The importance of such incentives for sustaining knowledge sharing has long been recognized (Ba et al. 2001). Even though the literature on knowledge sharing in organizations recognizes the need to align incentives to promote participation, existing studies offer conflicting evidence on what motivates participation (Faraj and Johnson 2010). The appropriate design of incentives for knowledge sharing remains subtle and multi-dimensional (Benbya and Belbaly 2010). Market prices, in the form of monetary incentives or other proxies, stimulate participation and are often associated with greater effort and performance. However, some studies reveal mixed findings with regards to such m ­ onetary incentives and indicate that they may even have some perverse effects on ­participation (Bonner et al. 2000). In psychology, this effect has been referred to as “hidden costs of reward” (Lepper and Greene 1978), while in economics it’s recognized as “crowding-out” (Deci and Ryan 1985; Frey and Jegen 2001). According to self-determination theory (Deci and Ryan 1985), financial incentives can produce over justification, and turn intrinsic motivation into extrinsic motivation. The debate over these findings has generally shifted to the question of what specific circumstances give rise to these counterintuitive effects (Heyman and Ariely 2004). Additionally, organizations often use rigid incentive mechanisms that neglect the power of market participants to themselves determine appropriate rewards. Given these issues, a large number of companies that rely on fixed rewards when creating their internal knowledge markets have encountered mixed results. HP, for example, established a fixed incentive program based on frequent flyer

8

2  Review of the Literature

miles to stimulate participation. Posting a useful resource to the internal platform earned 2,000 miles, posting a question earned 1,000 miles, and posting an answer earned 500 miles. But 90 days after launch, only 20 % of the target audience had performed any of these activities. Siemens, similarly, relied on a fixed number of “shares” that participants could redeem for prizes. For contributing solutions and success stories, participating employees earned 20 shares while they earned 10 shares for developing customer, technology or market bids. This program significantly increased contributions; nevertheless information quality remained poor. According to price theory, however, fixed rewards are inefficient because the prices fail to adjust. When prices are fixed, information that is more valuable than the price is not created—experts won’t waste their time—while information that is less valuable than the price is created—in effect, less-expert workers volunteer just to get paid. As a result, participants can find that such systems contain irrelevant content, or excessive rewards can divert workers from more valuable activities. On the other hand, not all knowledge sharing is guided by prices (Simon 1991). Rather, knowledge sharing is also embedded within long-term relations, social ties, and trust (Uzzi 1997; Zenger et al. 2011). Research on public goods has shown that, across a wide variety of situations, people sometimes contribute even when they incur costs from their pro-social behavior (Fehr and Fischbacher 2004). Such interactions cannot be explained through economic incentives. The literature on online communities, for example, suggests the importance of social incentives for participation, such as recognition and reciprocity but highlights problems of under contribution and sustained participation (Bateman et al. 2010). In addition, participants’ social relationships can also shape their knowledge transfer dynamics and their choices as to whether to participate in markets or communal interactions. A key social psychological principle is that people compare themselves to ­others to gain information on what constitutes the “right behavior” in various contexts, especially when situations are ambiguous (Buunk and Mussweiler 2001; Suls et al. 2002). Thus, people in online settings might likewise follow norms (community vs. market) that have been established by the consensus of others. Our research will describe and document how the advantages and challenges of these models emerge in a dynamic knowledge sharing environment.

References Ba, S., Stallaert, J., & Whinston, A. B. (2001). Introducing a third dimension in information systems design: The case for incentive alignment. Information Systems Research, 12(3), 225–239. Baldwin, C., & von Hippel, E. (2011). Modeling a paradigm shift: From producer innovation to user and open collaborative innovation. Organization Science, 22(6), 1399–1417. Bateman, P. J., Gray, P. H., & Butler, B. (2010). The impact of community commitment on participation in online communities. Information Systems Research, 22(4), 841–854. Benbya, H., & Belbaly, N. (2010). Understanding developers’ motives in open source projects: A multi-theoretical framework. Communications of the Associated for Information Systems, 27(30), 589–610.

References

9

Benbya, H., & van Alstyne, M. (2011). How to find answers within your company. MIT Sloan Management Review, 52(2), 65–75. Bogers, M., Afuah, A., & Bastian, B. (2010). Users as innovators: A review, critique, and future research directions. Journal of Management, 36(4), 857–875. Bonner, S. E., Hastie, R., Sprinkle, G. B., & Youg, S. M. (2000). A review of the effects of financial incentives on performance in laboratory tasks: Implications for management accounting. Journal of Management Accounting Research, 12, 19–64. Burt, R. (2004). Structural holes and good ideas. American Journal of Sociology, 110(2), 349–399. Buunk, B. P., & Mussweiler, T. (2001). New directions in social comparison research. European Journal of Social Psychology, 31, 467–475. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-regulation in human behavior. New York: Plenum Press. Faraj, S., & Johnson, S. (2011). Network exchange patterns in online communities. Organization Science, 22(6), 1464–1480. Fehr, E., & Fischbacher, U. (2004). Social norms and human cooperation. Trends in Cognitive Sciences, 8(4), 185–190. Frey, B. S., & Jegen, R. (2001). Motivation crowding theory: A survey of empirical evidence. Journal of Economic Surveys, 15(5), 589–611. Hayek, F. (1945). The use of knowledge in society. The American Economic Review, 35(4), 519–530. Heyman, J., & Ariely, D. (2004). Effort for payment a tale of two markets. Psychological Science, 15(11), 787–793. Lepper, M., & Greene, D. (1978). The hidden cost of rewards. Hillsdale, NJ: Erlbaum. Morrison, P. D., Roberts, J. H., & von Hippel, E. (2000). Determinants of user innovation and innovation sharing in a local market. Management Science, 46(12), 1513–1527. Oliveira, P., & von Hippel, E. (2011). Users as service innovators: The case of banking services. Research Policy, 40(6), 806–818. Osborn, A. F. (1963). Applied imagination: Principles and procedures of creative problem solving (3rd ed.). New York, NY: Charles Scribner’s Sons. (Revised). Prahalad, C. K., & Ramaswamy, V. (2004). Co-creation experiences: The next practice in value creation. Journal of Interactive Marketing, 18(3), 5–14. Shah, S., & Tripsas, M. (2012). When do user innovators start firms? A theory of user entrepreneurship. In: Revolutionizing innovation: Users, communities and open innovation. Cambridge, MA: MIT Press. Simon, H. A. (1991). Organizations and markets. The Journal of Economic Perspectives, 5(2), 25–44. Suls, J., Martin, R., & Wheeler, L. (2002). Social comparison: Why, with whom and with what effect? Current Directions in Psychological Science, 11(5), 159–163. Terwiesch, C., & Ulrich, K. T. (2009). Innovation tournaments. Boston: Harvard Business School Press. Uzzi, B. (1997). Social structure and competition in interfirm networks: The paradox of embeddedness. Administrative Science Quarterly, 42(1), 35–67. Van Dijk, C., & Van den Ende, J. (2002). Suggestion systems: Transferring employee creativity into practicable ideas. R&D Management, 32(5), 387–395. von Hippel, E. (1986). Lead users: A source of novel product concepts. Management Science, 32(7), 791–805. von Hippel, E. (2007). Horizontal innovation networks by and for users. Industrial and Corporate Change, 16, 293–315. Zenger, T., Felin, T., & Bigelow, L. (2011). Theories of the firm–market boundary. The Academy of Management Annals, 5(1), 89–133.

Chapter 3

Research Methodology

To understand how firms can benefit from internal markets and their v­arious effects we conducted a mixed-method study using several complementary approaches. Our approach is based on four elements. 1. Web-analysis of practices: To identify early adopters of internal markets, the benefits and challenges associated with their implementation, we carried out a web-analysis of the top 200 firms in the Fortune Global 500. Using the firms’ websites, we sought examples, publications and communications associated with internal markets adoption. This preliminary analysis of practices allows us to draw a picture of what the world’s 200 largest firms are doing and how they are using internal markets. 2. Exploratory interviews: We conducted exploratory interviews with 20 large multinational firms in both high-tech and low-tech sectors to understand the challenges associated with internal markets adoption. Knowledge markets do not always work so to get an overview of the main challenges firms are facing we spoke to senior managers in charge of knowledge and innovation management. These interviews allowed us to draw a rough sketch of the main issues firms are facing with their implementation and start to provide a roadmap with design strategies for overcoming them. 3. Mini-case studies: Out of the exploratory interview partners, we selected 2 multinational firms: Allianz and SAP to conduct further interviews on their experience with adopting knowledge markets and their effects. We analyzed how they approached the challenges of designing internal markets and the practices they had implemented to address them. 4. An in-depth case study: To analyze the effects of knowledge market in terms of productivity we had access to data from a multinational firm named Finaco. This company was in the process of applying a virtual currency to its existing marketplace so we could not analyze the effects of incentives on participation but access to two years data of knowledge exchanged within the platform

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_3

11

12

3  Research Methodology

and individual performance data allowed us to analyze the effect of knowledge sharing on performance. 5. Experiment: To further understand the effects of internal markets we build an internal knowledge market prototype and test it in different academic settings. The objective of designing such a virtual laboratory is to test in n­aturalistic settings the effects of material and social rewards on knowledge sharing. Specifically, we conduct two experiments: (1) The first experiment is c­ onducted in the US and tests the effects of material rewards on participation. (2) The second is conducted in Europe and explores when and how participants transition from a pay-based mode of knowledge sharing towards a free mode of exchange. Empirical analysis of real-world deployment data in these two different academic settings (US and Europe) reveals different effects of internal markets. Out of this effort came a framework for internal knowledge markets with design guidelines and an analysis of some of their effects.

Chapter 4

Results

In the following section, we provide the results of our investigation. Our a­ nalysis identifies several organizational needs that can be fulfilled by using knowledge markets. We first discuss these needs and illustrate them with examples from our web-analysis of practices and exploratory interviews. We then analyze how firms approached the challenges of designing internal markets and propose strategies for solving them. We illustrate some of these challenges with two mini-cases (SAP and Allianz) and then provide a framework summarizing lessons to design an effective internal knowledge market. We then present the effects of internal markets. We start by presenting a case study on the effect of knowledge sharing on individual performance and then summarize our experiments aimed at understanding the effects of incentives on participation in internal markets.

4.1 Internal Markets Types Our analysis reveals four uniquely valuable uses of markets in companies: event forecasting via Prediction Markets, idea genesis and evaluation via Idea Markets, problem solving via Innovation Markets, and peer to peer assistance via Knowledge Markets (Table 4.1). The social computing lab at Hewlett Packard developed a prediction market to forecast the price of computer chips. Capturing information across a broad swath of employees, not just a few experts, provided price estimates up to 70 % more accurate than traditional forecasting methods. Princeton’s student government ran an idea market to generate budget spending proposals and assess their rank. Officers generated as many ideas as possible which they submitted to student vote. But, they also allowed students to submit new ideas. The result? Not only did the elected officials misjudge student priorities, but also three of the top five ideas were not on the original list! Idea markets represent a solver push model in which proposals identify new problems and opportunities. This stands in contrast to

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_4

13

Innovation markets (Problem-solving)

Ideas markets (Product development and opportunity identification)

Knowledge markets (Q&A)

Prediction markets (Forecasting)

Internal market types

Table 4.1  Internal markets types

Rationale and application Prediction markets are speculative markets for forecasting uncertain events and trading contracts that yield payments based on the outcome of those events Application: business forecasting, supply chain management, new product selection, policy analysis, and sports betting A platform for peer-to-peer assistance, pairing experts with novices, and broadcasting both news and also questions in need of answers Application: knowledge management, decentralized information sharing A forum for generating, combining, and ranking ideas based on open contributions and collectively sorted values Application: design, new product development, problem and opportunity identification Markets for R&D problem solving that act as “brokers” for matching a problem solution seeker with a network of problem solvers Application: Research and development, resolution of specific challenges 1. Bring companies with specific R&D problems together with researchers and problem solvers around the world 2. Save time and money by providing access to qualified researchers without having to pay for failed experiments

Eli Lilly, McKinsey, Pfizer, Infosys, Siemens, Bank of America, Shell, SAP, IBM, World Bank, Samsung Life Insurance, Fujitsu, Diamond Lease, NTT software

1. Match knowledge sources (consumers and producers, experts and novices) 2. Facilitate sharing across a distributed organization 3. Capture best practices. Generate faster responses 1. Involve users in idea generation improves idea diversity and ranking 2. User participation encourages transparency, acceptance and idea transfer

Boeing, DuPont, Eli Lilly, SAP, Procter and Gamble, Master Foods, Syngenta

Lego, GE, Intercontinental Hotels, Peugeot, Salesforce, Starbucks, Threadless, Cisco, IBM, Allianz, Citibank

Firm implementation Best Buy, US Department of Defense, Eli Lilly, General Electric, Google, France Telecom, Hewlett-Packard, IBM, Intel, Microsoft, Yahoo, Masterfoods, Siemens, Arcelor Mittal, Renault

Advantages 1. Improve decisions, help businesses make better investments, help governments make better fiscal and monetary policy decisions 2. Provide forecasts on a wide range of events, from presidential elections to printer sales

14 4 Results

4.1  Internal Markets Types

15

innovation markets, representing a seeker pull model in which proposals address an existing and narrowly defined problem. Using an innovation market, Syngenta issued specific agricultural challenges to its employee base. This produced a 182 % return on investment by increasing researcher output, avoiding use of outside experts, and shortening time to solutions. Finally, SAP has used a knowledge market to facilitate peer-to-peer answering of questions among enterprise software developers. This form of Q&A has improved response times from one business day to 30 min while saving SAP more than $6 million in technical support costs. Table 4.1 provides a taxonomy of internal markets and occasions when they work. In the following we introduce the main differences between two types of internal markets: idea and prediction markets.

4.1.1 Idea Markets Idea Market are electronic markets for matching solutions to problems. Members of the organization can create new ideas or submit one in response to another user’s request. A member of an organization might have innovative ideas on the organization’s operations, products or activities. She can submit her ideas on the idea market, get them examined, validated and improved by other members, and receive rewards when an idea gets promoted by senior management or even adopted. As such, idea markets support the key activities of the idea lifecycle management such as idea generation, evaluation and selection and they result in the generation of more ideas than comparable methods. Idea markets use idea stocks to represent new product ideas, let participants trade the stocks on a virtual marketplace, and use the efficiency of markets and the resulting stock prices as indicators for the possible success of the different new product ideas. Traditional methods of idea generation and selection remain limited and fail to leverage the collective wisdom of the organization. Employees often receive little feedback on their ideas. In addition, since ideas are seen only by a small number of people in the organization, there’s no spring-boarding, or building upon each other’s ideas. The ideas, once submitted, are static and do not respond to changes in the environment. Idea markets are designed to address these problems.

4.1.2 Prediction Markets Prediction markets also connect a group of participants together on a virtual marketplace and let them trade shares of virtual stocks. These stocks represent a bet on the outcome of future events. These events can be public, such as political

16

4 Results

e­ lections or sport outcomes or private within organizations, such as product sales or project success. Prediction markets were first applied to predict the outcome of the 1988 US presidential election. Participation in this market was restricted to members of the university of Iowa community. Since then, numerous election markets have been conducted both within the US and abroad. The results of predictions markets demonstrate their accuracy compared to alternative prediction methods. Since the 1980s, prediction markets have been successfully used for many different forecasting purposes and have proven to be more reliable than alternatives, such as polls, expert opinions, surveys or econometric models (Hanson 2003; Wolfers and Zitzewitz 2004).

4.2 Mini Cases of Internal Markets 4.2.1 Allianz “Idea Market” Allianz, UK a leading global financial services provider adopted an ‘ideas system’ to encourage all employees to suggest and implement ideas that would positively impact team performance and make innovation part of the day job for its 4,500 employees. Initially, an electronic suggestion box to support a series of internal communication campaigns as a call to action for ideas from across the entire organization was implemented. This served as both a launch of the innovation agenda and as a way of getting things moving. Line managers were charged with endorsing ideas, and where they could within their remit, agree to their implementation. Early on, the innovation team discovered there was much less chance that breakthrough ideas from the top would ever be implemented. The real value came from the quantity of ideas emerging from the bottom levels of the organization, though, even there was a difficulty. The ideas that came through where not always relevant to the purpose and there were simply not enough ideas being generated. The team in charge of strategic innovation had to find a way to focus innovation and make it part of the priorities of Allianz UK’s employees. Focusing innovation on continuous improvement helped the team to target the problems and priorities of workers, helping them to define a problem within their sphere of work and responsibility, to look at ways of solving the problem, in the process reflecting on how they improved the way they worked. This moved the focus of idea generation from the fuzzy to the practical and instantly the volume and relevance of ideas increased. The Innovation team had strong senior management sponsorship from the CEO to go in this direction. The support from the Innovation Steering Group, made up of Lead Innovation Champions from each department, focusing on ­making innovation relevant in context has meant meetings that occurred monthly. Since launching, over 25,000 ideas have been implemented and have delivered an annualized financial value that exceeds £12.5 m ($19 million).

4.2  Mini Cases of Internal Markets

17

4.2.1.1 Challenges and Lessons Learned • To generate large numbers of ideas make the numbers visible to all To make innovation a priority the team realized that they had to make the innovation performance visible to all so that each divisional head would begin to understand their role in achieving “innovation numbers”. By designing and producing a monthly single page innovation dashboard with clear and understandable metrics in which everyone in the organization can see their and each other’s contributions and the progress of all current innovation contributions circulating in the innovation system, each board member began to take interest to ensure their area is an active contributor. This has produced more engagement at a board level where competition between board members has led to uplift the innovation engagement. It has proven to be the vital element in the organization’s innovation strategy because without board engagement there is no possibility of any meaningful development. This dashboard approach has then been repeated across all of the divisions, highlighting superior performance across each of the business areas. • Do not target an individual idea’s financial value at the outset When working in a highly regulated and mature market such as Allianz UK does, strategic innovation may not come in the form of new product development but from the way existing products themselves are offered, sold or serviced. So it’s important to make certain idea generation is continually encouraged, whilst at the same time actively developing communication campaigns and exhorting greater thought effort around the purpose for idea generation. It’s also important not to target an individual idea’s financial value at the outset because it is impossible to predict. Most value in ideas comes from the accumulation of lots of little ideas so the volume of small ideas around any topic is absolutely crucial. • Reward people for their innovation efforts but do not focus on material rewards Allianz have an annual CEOs Award in which the most valuable innovation each year, as recognized by the CEO, receives 10 % of the value of the idea in its first year of implementation. Since its inception, an award of around £50,000 ($90,000) has been won annually by an individual or small team. This has ­created perceived tension in the organization in several ways. Firstly the question is being asked whether idea generation and implementation is part of an individual’s day job or whether the extra money is a separate reward for additional effort over and above the normal tasks of a day job. This again highlights the dilemma between moving innovation from outcome to p­ rocess. When incentives are in play, is the manager’s mind focused on thinking about how to improve the processes associated with his/her day job and producing ideas in that context or on seeking ways to create ideas for financial benefit for the manager? Secondly is the tough challenge of trying to value ideas. Many great ideas ­simply won’t be valuable, yet may very well fit into Allianz’s idea benefit ­recognition rules. Areas such as procurement and claims that save large amounts

18

4 Results

are at an advantage as they are seen as being able to generate value. So how do you balance what these areas do with other areas in the business who may not have the opportunity to create such savings for the company and thus contribute cash to reward efforts in these areas? The protocols in the idea benefit recognition rules are based on generating financial value and the challenge then is how to understand and capture the substance and the value of the idea simultaneously in teams or by the individual employee in context. • Measure innovation using quantitative as well as qualitative indicators Three key metrics are used in Allianz: (i) Number of ideas generated; (ii) Number of ideas implemented and (iii) Financial value of ideas implemented. Allianz UK therefore requires each employee or team to validate their implemented ideas by declaring a value for each idea. This requires perception and agreement from individual managers and teams and can only happen if the initial idea is visible and transparent. In addition to these numbers, qualitative indicators, in the form of a survey, captures and identifies manager’s perceptions of the system and how it works. This survey reveals time barriers and impediments to organizational creativity to the leadership team, and provides an analysis of the current environment and culture.

4.2.2 SAP “Knowledge Market” SAP (http://www.sap.com) the third largest computer company in the world used a knowledge market to support its enterprise software developers in peer-to-peer response in the form of Q&A. SAP provides a wide range of enterprise software solutions that include powerful, but complex, customizable frameworks. The application of these frameworks is a highly specialized skilled endeavor. Its knowledge market aims to support knowledge sharing and collaboration about SAP technologies via expert blogs, discussion forums, code samples, training materials and a technical library. There are over 1 million registered users and over 6,000 thousand posts per week. To motivate members to contribute, SAP based the exchange on its knowledge market on a fixed point system in which members who post threads can award points to members who help the most. Originally, developers earned points toward t-shirts and memorabilia for trouble-shooting, blogging, and answering each other’s questions. In 2008, the community wanted something more socially responsible so, based on their input, SAP offered charitable donations of €100,000 if the community could collectively generate 2.5 million points. The community rallied to a stretch goal of 3.5 million points, and SAP doubled its offer, sponsoring €200,000 worth of children’s education and health care through the United Nations. Throughout, in addition to the group-based incentives, SAP continued to offer special recognition to its highest point generating individual developers and business process experts, and consults with them on policies and product design—a clear win for everyone involved. Active Contributors enjoy enhanced recognition with a badge that shows their current level of contribution: bronze, silver, gold or platinum.

4.2  Mini Cases of Internal Markets

19

SAP have seen positive results with its knowledge market as it saved more than $6 million in technical support costs. The knowledge market is also considered as a place to “Think out loud” that exhibited comfortable expression by its members, development of social norms and self-policing to the point where it becomes an extension of the developer’s workgroup, creating a virtual organization (VO) of developers that transcends corporate organizational boundaries. 4.2.2.1 Challenges and Lessons Learned The point system used by SAP can be considered as a strong motivator for participation. SAP for instance offered temporarily to its developer’s double points to stimulate participation for a knowledge market activity relating to SAP’s new customer relationship management (CRM) products. The subsidies boosted development of CRM content that SAP itself did not have to provide. However, the point system used has both positive and negative effects. In the followings we distill some lessons learned from SAP experience. • Use material and social incentives, but let prices float SAP relies on a fixed point system to stimulate participation in its knowledge market. While fixed points increase the number of contributions their quality remains uncertain. This following example shows the high stake in the point system and the limits of a fixed point system. As a result, such systems can either became burdened with irrelevant content or divert workers from more valuable activities with excessive rewards. To address these problems, managers should recognize fixed prices in internal markets as a form of central planning and don’t fix prices for answers but rely on a floating currency where quality affects the points earned.



20

4 Results

• Encourage cooperation One social concern about using a virtual currency or a point system in markets is that they can perversely alter spontaneous interactions by converting volunteers within a company into vendors. This can be seen in a variety of categories of thread examples where: (1) helpers remind askers to reward points; (2) helpers complain if no points are awarded. There appears to be a “race” condition in a competition to score points. Helpers may be less motivated by the desire to help someone solve a problem. Instead, their motivation is to score points. In this case motivation to help solve a problem and motivation to reply to a question could be two different things. The best incentives therefore, should balance competition and collaboration. Use competitive rewards when problems require greater speed and diversity of approach, and solutions are substitutes. Alternatively, use more collaborative rewards when problems require complex coordination and solutions are complements. • Beware of manipulation Despite evidence of the development of social norms and self-regulation in knowledge markets, whenever there’s a virtual currency in stake that participants can use to earn material or social benefits, governance mechanisms to regulate participation are necessary. In the case of SAP, for instance, some users cheat to gain points and the community detects “cheaters” and report them in discussions. However, if no mechanism is in place to flag cheaters, detect and correct those behaviors, feelings of unfairness may dominate.

4.3 Design Guidelines Executives who wish to implement internal markets must consider fundamentally different design logic. We begin with two key principles then follow with implementation strategies to make such markets work.

4.3.1 Incentives Design To enable participants to exchange value, markets require a medium of exchange. Rather than rely solely on a fixed reward or rating system, firms have introduced virtual currencies to trigger quality content. An online currency is any official or de facto unit of exchange used in an online transaction. In knowledge markets, participants have usually some initial portfolio to start with which can be expanded or reduced based on their trading activities. Currency, however, can take different forms and be exchanged by any social or material benefit that people value. Time banks, for example, rely on a time-based currency to exchange services where the unit of exchange is the person-hour. Innocentive@work internal market allows participants to accumulate points and convert them to money or prizes. Participants of Newsfutures’ prediction markets can use the virtual-money they earned to buy some items in an online ­auction shop. Linden Lab, the company that created the virtual word “Second Life”

4.3  Design Guidelines

21

has issued Linden-dollars, a virtual currency that can be converted into real U.S. dollars or redeemed for goods and services that confer some kind of advantage in these virtual games. Infosys, for example, introduced Knowledge Currency Units (KCUs), which workers could convert to money and prizes. KCU scores also appeared on the internal market KShop increasing their visibility and harnessing status incentives (Garud and Kumaraswamy 2005). Within a year, people had contributed over 2,400 proposals, cases, software objects and reviews. Virtual currencies offer three major advantages over ratings alone. First, currencies are unbounded and transferable whereas ratings are bounded and non-transferable. Ratings among polite colleagues bunch at the top, limiting their value. Even armslength ratings on E-Bay are more than 98 % positive. Also, users cannot spend ratings. Imagine trading on EBay with only ratings, no posted prices, and exchanges that can only involve bartered goods. Currencies lubricate trade. Linden Labs, creator of the virtual word “Second Life,” has issued virtual Linden-dollars that can be converted into real U.S. dollars or spent on goods and services that confer advantages inside the virtual world. For internal markets, virtual currencies cost less than real money and evidence suggests they are also very good at promoting participation and providing accurate information (Servan-Schreiber et al. 2004). Second, virtual currencies promote the best use of everyone’s time. Because a person must work within a limited budget, he or she is forced to make choices about what is less and what is more important. Meanwhile, posting a question worth 5 points ­probably won’t divert executive attention from their own tasks while posting a critical problem worth 5,000 points probably will. Resource balance is the essence of price theory. Each person, acting on his or her own private information, allocates time efficiently for themselves and for others. Important or critical issues immediately get flagged as important or critical—a benefit missing from ratings that are supplied after the fact. Whether an internal knowledge market uses a virtual currency based on convertible points or real money as the unit of exchange, it’s critical that the adopted currency have two properties. Firstly, it must allow implicit prices to “float”. Fixed prices, whether in currency, points, or reputation benefits create an oversupply of low quality content and undersupply of high quality content. Secondly, introducing a currency system requires from the market maker to establish mechanisms to ensure it circulates in such a way as to avoid inflation and deflation. Executives in the C-Suite might find it tempting to flood the market with points in ever expanding pursuit of productivity. This would be unwise as it leads to clear parallels with economic inflation and instability. In this case, the knowledge market maker can conduct open market operations—such as offering services or buying points that shrink the point supply—if expansion has gone too far.

4.3.2 Governance Mechanisms Introducing an internal market opens the firm to new forms of governance and creates completely new roles for those who would manage expertise. As noted in the earlier section, the IT department can no longer behave like a central planner but must instead

22

4 Results

behave like a Market Maker or Federal Reserve. As market maker, it steps into provide liquidity when market participants have failed to match, when they lack information, or when they are too busy to provide assistance. As Federal Reserve, the IT department manages the internal economy for optimal growth. Most executives have no idea how to perform these roles. But, they are crucial to bringing markets inside the firm. This implies not only that executives change how their people interact, and change their tools, but also how executives change themselves. To run the company effectively, do they accept verdicts of their market? The true test of internal market adoption is whether firms use them for critical issues, not just fringe areas. When people pose better questions and post better answers, does the IKM empower them to influence firm choices? Management must support the change in power that follows from a change in the sources of expertise. And management must accept, or at least acknowledge the market verdict. In one Pacific Rim bank we studied, over 90 % of questions travelled up from field offices to headquarters while answers travelled down in the opposite direction. Tapping expertise only vertically misses the full benefits of openness. Truly democratizing knowledge requires an open organization where employees can deliberate, argue, vote, compete, and collaborate horizontally across fields of expertise.

4.3.3 Lessons Learned Based on our analysis, we provide lessons to design an effective internal knowledge market. Specifically, we distinguish three phases: launch, development and evolution. For each phase, we highlight both the challenges involved (e.g. lack of critical mass of content), as well as the response the firm might take (e.g. seed the market with critical knowledge). Knowledge market launch requires the market maker to seed-and-subsidize in order to provide critical mass for market germination and to match knowledge seekers with relevant sources. Sustaining the market implies that liquidity inflows should balance and slightly exceed outflows in order grow knowledge transactions. Targets can follow growth practices exactly like those of the Federal Reserve. To quantify information and IT investment value, track internal points trading while letting prices float. Market pricing will reveal direct measures of productivity. Table 4.2 summarizes mains challenges related to market adoption and solutions to address them.

4.4 Internal Knowledge Markets Effects Introducing markets inside firms produces a variety of effects some of them are positive and related to performance improvement, collaboration and openness; others need to be carefully managed such as competition, manipulation,

4.4  Internal Knowledge Markets Effects

23

Table 4.2  Mains challenges related to market adoption and solutions Market phase Launch

Challenge Critical mass problem: people abstain from using the market

Limited, non-tradeable, or inflexible incentives

Shy users. Workers self-conscious about posting issues publicly Hoarding knowledge/worker competition

Development

Contradictory or wrong answers

Limiting the strategic flow of information to competitors Trade imbalance: one group provides more market value than another

Limited use for virtual points

Devaluing community by pricing knowledge. Social substitution— reducing volunteerism to selling

Evolution

Adaption, missing features Deaf executives and governance

Collusion/manipulation

Measuring the value of investments and information

Response • Seed the market with critical knowledge • Subsidize key knowledge producers • Incentives must include both material and social rewards • Use currency or points, not just ratings, and don’t fix prices • Create small/private subgroups for sharing knowledge • Allow anonymous posting • Provide absolute (not relative) rewards • Reward in proportion to contribution • Give credit for reuse • Collect multiple answers, counterpoint and commentary. Community reconciles differences • Create market segments with different access permissions • Tie points inflows to inside and outside activity (e.g. billings) • Highlight over/under-performing groups • Create a complete market. Manage for points outflows as well as inflows e.g. redemption for real goods and services • Acknowledge and thank people for contributing. Provide recognition and status • Encourage formation of “identity” based groups • Use market to heal itself • If executives ignore market advice and decisions turn out badly, they own the consequences • Provide transactions transparency, data transparency and individual reputations • Align individual goals with firm’s goals • Use market prices to learn value. Use changes in worker productivity to check it and determine where to invest

24

4 Results

transforming spontaneous interactions to calculated exchange among others. To test some of these effects, in addition to case analysis we conducted experiments to better understand market dynamics. In the followings we summarize some of the results and lessons learned from different studies.

4.4.1 Case Analysis: Effects on Individual Performance A leader in the financial services in Japan, Finaco is the world’s top 10 financial institutions in terms of market capitalization. To maintain its competitive position Finaco needed to accelerate its customer-oriented business. This implies that it should be able to share its best practices and know-how from sales staff with the best customer knowledge and utilize that knowledge to sharpen its business capabilities. This knowledge used to be scattered across branches and departments. Each branch or department used to manage information on its respective Notes database, and information was dispersed over individual databases. That made it difficult to search across the organization for needed information or for personnel with useful knowhow. Moreover, there was almost no interactive cross-organizational communication on the Notes system. To organize and consolidate its best practices across branches and create a forum for intellectual collaboration, Finaco invested in a knowledge market. The objective of the knowledge market is to keep employees constantly informed by the headquarters and to share their information with their colleagues. It integrates a document repository summarizing processes leading up to successful deals with customers as well as a discussion board allowing frontline employees to share knowledge based on Questions and Answers (Q&A). In particular, each employee has now access to the documents made available by the headquarters, or may ask questions to other employees and provide answers to others questions. The implementation of this knowledge market allowed a more efficient communication, both vertically, between headquarter and the lower layers of the organization, and horizontally, that is, between workers. This provides us with a natural environment in which to understand how knowledge markets affect individual performance. In this environment, knowledge sharing among loan officers has several effects. First, allowing managers to share knowledge allow them to better assess the riskiness of the enterprises, or work on more projects at the same time. Second, it might allow the low ability workers to fill the gap with the most productive ones. Third, high skill workers may be required to devote a larger fraction of their time to help their colleagues, which might negatively impact their performance. Our main contribution is to disentangle the different effects that knowledge sharing has on performance. 4.4.1.1 Individual Performance Relationship managers are reviewed semi-annually to assess their performance. In order to account for branch location and task differences, there exist six main groups to which each manager may belong. These groups are: large existing

4.4  Internal Knowledge Markets Effects

25

account, small existing account, restructuring group, public sector, new strategic account, and new account. Duties correspond to those of loan officers who are responsible for the profit levels obtained for each type of loan. Two scores are used to assess the performance of relationship managers. Quantitative measure: Bank headquarters set the targets for each manager in these groups and to each relationship manager the head of the branch assigns a score of up to 80 points based on his performance in the relevant category: bank gross profit, revenue, loan profit, liquid deposit profit, loan volume, reduced disclosed debt, and reduced estimated loss. The performance of loan officers in these dimensions constitutes our objective performance measure. Qualitative measure: For increased accuracy, the branch manager may, at his discretion, assign a qualitative score up to 60 for performance in the following categories: customer-oriented process, contribution to branch operations, contribution to organization operations, and loan reinforcement. As reported by executives of the bank, the branch managers often assign this score to reward high effort that did not yield results or discount low effort that did. Our data include all targets, objective results, and subjective scores for each relationship manager in each group for every branch. We believe that the richness of our data and the fact that we need not rely on survey data to extrapolate observed performance make it highly suitable for study of performance differences across officers. In most of the analysis below, we focus our investigation on one dependent variable: total performance. This is the total score assigned to officers, representing the sum of performance in each of the different categories and including the individual subjective score. We also checked our results using only the objective performance measure, which does not include the subjective score assigned by the branch manager, and the results are unaffected. 4.4.1.2 Independent Variables: Knowledge Sharing The main variable of interest describes how relationship managers use the knowledge market to share and seek knowledge. We collected data on all accesses to the knowledge market by each manager during the period of interest. Our data include (i) the number of documents consulted by each manager, (ii) the number of questions posted, and (iii) the number of answers provided, down to the second of access and across each term. We also have information on the number of years the manager has worked for the Bank, captured by the variable tenure, and whether he came directly from school with no prior experience or transferred from another bank, captured by the dummy college. 4.4.1.3 Findings Table 4.3 reports descriptive statistics for our variables of interest and two things are worth noting. First, loan officers perform significantly differently, in fact, the mean of our main measure of performance is 52, but the standard deviation is 21.

4 Results

26 Table 4.3  Summary statistics Total performance Number documents Number questions Number answers Tenure

Mean

Median

Min

Max

52 569 77 250 10

48 522 45 238 11

8 102 4 4 0

110 1,618 585 813 32

Std. dev. 21 282 123 164 5

Then, even if we focus on the employees of the same bank, within the same country and with homogeneous tasks, we still find that their performance is heterogeneous. Second, loan officers seem to make great use of the available information within the organization. On average they have access to 569 documents, post 77 questions and provide 250 answers during a six-month period. Moreover, the standard deviation of all these three variables is quite high, ranging from 164 to 282. This will be relevant to interpreting model coefficients for our results. Our main hypothesis is that officers may access the knowledge generated within the bank in order to improve their performance. In particular, we expect low performers to ask more questions and provide fewer answers. We also analyze the total performance measure for below median number of answers (above median number of questions and below median number of questions). The loan officers who help others more often, answering their questions, on average perform better than the others. In contrast, the loan officers that ask more questions are associated with lower performance. Table 4.4 confirms this intuition showing the officers’ performance for those who have shared information more or less than the median officer. The second column, instead, shows that there exists a positive correlation between the number of answer provided and performance. The loan officers who provide a higher number of answers perform significantly better than the others. The third ­column confirms this result showing that those who ask more questions, above 11 the median, performs significantly worse than the others. These results suggest that information sharing is correlated with performance and with the managers’ innate ability. Further analysis shows that the number of document and the number of answers are positively correlated with managerial performance, while the number of questions is instead negatively correlated with their performance. Furthermore, tenure significantly affects performance, which can be the result of a longer experience within the Bank. At the same time, joining the bank right after college, without any previous experience, is positively correlated with performance. This result can be interpreted as a result of greater effort exerted by new employees. Table 4.4  Managers performance by knowledge sharing Below median Above median Difference

Number documents 52.603 (0.291) 52.025 (0.306) 0.577 (0.422)

*** denotes significance at 1 %

Number answers 50.356 (0.294) 54.438 (0.306) 4.085*** (0.420)

Number questions 53.919 (0.308) 50.727 (0.287) −3.192*** (0.421)

4.4  Internal Knowledge Markets Effects

27

The main concern with these results is that knowledge sharing and the effect on performance might be driven by other factors, such as a market turndown, a greater need of information for a specific local market or a result of a better performing branch. In order to control for all this unobserved heterogeneity, we control for time, group and branch fixed effects. Except for the effect of the number of answers, the other coefficients are still economically and statistically significant. We further control for the interaction of time and branch effects, which shows that the results are robust to this more restrictive specification. These results suggest that when managers increase the number of documents accessed, this has a positive and significant effect on the productivity of the average worker, whereas increasing the number of questions has a negative impact on his performance. The magnitude of these effects implies that when managers increase their knowledge access by one standard deviation it increases their performance by eleven percent. Similarly, an increase in the number of questions is associated with a reduction in performance of about five percent.

4.4.1.4 Summary To address the question of whether access to knowledge produced and gathered within the same organization affects the performance of knowledge workers, we examined two years of micro data from a major financial institution. Data include all accesses to a knowledge market, objective and subjective performance measures, and all promotions among more than 2,800 loan officers. We find that a standard deviation increase in the number of shared documents predicts an 11 % rise in output, in specifications with time, location and individual fixed effects. Quantile regressions estimates suggest that knowledge sharing greatly benefits the low-performance officers. In fact, questions are used more by workers of lesser ability, and when performance gains exist, they appear strongly on the left tail of the distribution, at the 10th and 25th percentiles, but do not appear on the right tail, at the 75th and 90th percentiles.

4.4.2 Experiments Results 4.4.2.1 Experiment 1: Effects of Material Rewards on Participation The objective of this experiment is to test the effects of material rewards on participation in an internal knowledge market. To explore this effect, we rely on price theory and draw upon recent investigations of the role of virtual currency (VC). Evidence suggests that VC is as good as real money in promoting participation and providing accurate information (Servan-Schreiber et al. 2004). Many firms rely on VC in the form of point systems that represent real monetary value for knowledge. Infosys, for example, introduced Knowledge Currency Units (KCUs),

28

4 Results

which workers could convert to money and prizes. KCU scores also appeared on the internal market KShop, increasing their visibility and harnessing status incentives (Garud and Kumaraswamy 2005). Similarly, InnoCentive@work participants can earn points that can be converted into money or other benefits. In these deployments, users started with certain number of points assigned to their bank accounts. We do not ask users to engage in any particular activity attentively. The knowledge market simply serves as a day-to-day tool for students to exchange knowledge, share documents, contribute ideas, post news, etc., following the underlying market rules. Incentive design, especially the translation of virtual points to meaningful rewards, is important to the deployment and experiment. The goal is to get users motivated and incentivized to participate. The auction market, “iBid”, provides a mechanism for users to exchange earned points for rewards they care about. Meanwhile, we can use the auction market to measure the value of information in dollar terms for research purpose. We designed and decided upon an incentive structure, advertised it to users at the beginning of the deployment period. For example, we posted iTunes gift cards, small consumer products, and lunch opportunities with faculty members on the “iBid” market. All user activities, transactions and market statistics have been captured by the system. Through the wide deployment of this knowledge market, we are able to collect a rich dataset on market operations and user behaviors. Analysis of User Behavior and Market Effects 1. Users are responsive to material incentives Figure 4.1 shows a dashboard of the price index of the questions and answers in the market on the U.S. site. The market maker can use it to monitor several key economic indicators of the economy. It is evident from these curves that users do respond to material incentives. In March 2011, we posted several valuable rewards for users to exchange for their virtual points, which stimulated market participation significantly.

Fig. 4.1  Price levels on the knowledge market

4.4  Internal Knowledge Markets Effects

29

Fig. 4.2  Number of questions asked categorized by the number of points offered—a all questions; b anonymous question; c non-anonymous questions

In mid-June 2011, we posted an iPad on exchange market, which also greatly stimulated market activities (and market fraud too!). 2. Market forces are significant A frequently asked question in the context of a knowledge market is if market forces are effective. To see if offered material incentives affect the quantity and quality of answers provided, we analyze data collected from the knowledge market on the US site. Figure 4.2 shows the number of questions asked on the U.S. site for each level of offered points—anonymous, non-anonymous, and the total are all displayed. Most questions were asked with around 20 points offered, while only a limited number of questions rewarded 100+ points. Figure 4.3 shows the average number of answers a question received for different levels of offered points. When the number of offered points is below 100, the average number of submitted answers received generally increases as the number of offered points arises. As the offered points exceed 100, the difficulty of the question increases more significantly, thus the number of submitted answers decreases sharply with offered points. There appears strong linearity for anonymous questions. Following other studies on knowledge markets, we use the average length of answers as a proxy for answer quality. It is evident from Fig. 4.4 that the average length of answers keeps increasing as the offered points arises for a question, and when the offered points exceed 100, the increased difficulty thwarts good quality answers. Both curves exhibit a reverse-U shape, and the local maximum in terms of both quantity and quality of answers is achieved when the number of offered points are in the 90–110 range, for both quantity and quality of answers. More interestingly, this is also where the curve in Fig. 4.4 reaches a local peak value, which implies that users asked a significant number of questions within this price range.

30

4 Results

Fig. 4.3  Average number of answers obtained, categorized by the number of points offered for questions—a all questions; b anonymous questions; c non-anonymous questions

Fig. 4.4  Average length of answers obtained, categorized by the number of points offered for questions—a all questions; b anonymous question; c non-anonymous questions

3. Evaluating information Using virtual currency to exchange information following market rules, and exchanging virtual currency for material rewards through the auction market provide a mechanism to measure the value of information created on the platform. Using the U.S. site as an example, we select 11 product items of which we can easily learn the dollar values, such as a $10 itunes card or the ipad. Virtual rewards such as iTunes gift cards, or items with undefined $ value such as “A ping-pong lesson” are not included.

4.4  Internal Knowledge Markets Effects

31

After performing regression, we can obtain the approximate exchange rate between U.S. dollar and Barter’s Yuanbao is $1 = 6:64 (Yuanbao). During the 1-year period from 04.2011 to 04.2012, the total transaction volume happening in the Q&A market is 7,476 Yuanbao—the sum of points offered and transacted for solved questions. Therefore, the value of information created in the Q&A market is $1,126 denominated in U.S. dollars. This exceeds the value of what we spent to purchase the products and stimulate the knowledge economy, which was about $750. This doesn’t include value created in other information markets such as the document and idea markets, and doesn’t count other value created such as users’ voting and tipping for helping sort and index information. 4.4.2.2 Experiment 2: Exploring Transitions Between a Pay Versus a Free Mode of Exchange The objective of the second experiment is to explore when and how participants transition from a pay mode of knowledge sharing towards a free mode of knowledge sharing. As people have positive or negative experiences within a particular relational model, they might have emotional responses to that system. For instance, as they compete in a market, they may experience the thrill of competition and the pleasure of seeing their position in a status ranking. They may feel that the system is fair and enjoy participation within it. Without governance mechanisms, however, markets can potentially be readily exploited by opportunistic players (Zingales 2000, 2010; Benbya and van Alstyne 2011). As such, people may also experience more negative emotional reactions as well. They may observe unfair processes, experience inequity, and feel outrage. When people experience inequity, they may seek to redress it (Greenberg 1987). By expressing their responses socially, they receive feedback from their peers and from the institutional authorities. They might test new relational forms and look to peers and institutional authorities, who may ignore, resist, or ­support their efforts. For instance, peers may reject a social innovation because it is not consistent with their own values or because they fail to notice it. Institutional authorities may display willingness to intervene or punish particular actions. As these psychological feelings are translated into social responses, regimes can shift. Paralleling research on revolution and social movements, these responses, if adopted by the group, can cause structural shifts. In sum, we investigate how an information technology which allows people to price their knowledge or not, and mask or reveal their identities, serves various relational goals. We can explore participants’ micro-level choices because we designed an IT tool that offers participants the option to select between market/ community and identity display options and analyze their shifts over time rather than a more static technology.

32

4 Results

Results To understand better the effects of internal markets in a naturalistic setting we conducted an experiment with professional adults (MBA working adults). We first designed a technology where participants could opt to either buy knowledge (using an incentive structure based on a virtual point system) or request it for free (Appendix 2 describes the tool and study detail). We then use a data-driven, inductive approach from 550 members and over 5,000 interactions (1,784 questions, 3,746 answers) over a period of nine months, to understand when participants opt for two IT artifacts: incentives (i.e., offering incentives for knowledge or not) and identity display (i.e., visible vs. anonymous interaction). Because the technology offered participants the freedom to choose either market or community mechanisms, we observe when people price knowledge, and when they offer it for free. We can also observe how participants seek opportunities in markets; create market efficiencies, and how they attempt to enforce norms in the system. We observe their responses to market interventions, and see them creatively generate novel patterns of both communal and market relations. As people experience the IT platform and interact with each other, their preferences for incentives and identity display change to reflect their shifts between collaborative and competitive norms. To the best of our knowledge, there has not been an in-depth investigation that assesses how individuals choose market and community models in knowledge sharing and how these choices evolve over time.

4.5 Analysis and Discussion Through passive observation and data analysis, we analyzed how participants interacted with the technology and how that technology in turn transformed ­patterns of interaction (Lee 1991). We identify three phases as participants adopted and used Wise: experimentation (as participants learned about market and communal interactions in this knowledge sharing environment), market dominance (as communal interaction transformed into market exchange), and communalism (as participants grew disillusioned with the market and abandoned it for community exchange). These three behavioral phases were manifested in how participants used the technology.

4.5.1 Experimentation Phase: Learning About Markets and Communities During the first month of using Wise, participants started to experiment with the tool and test its various features. Wise gradually became a day-to-day tool for participants to share knowledge and find answers as they engaged in work.

4.5  Analysis and Discussion

33

Participants mainly opted to reveal their identities at this initial stage. They quickly realized how markets facilitate their ability to match problems with solutions. People could be linked to other participants facing similar problems, get aggregate information from geographically dispersed peers, share what worked, and get responses to their challenges within 24 h. The average number of answers a question received was 3 (with a range from 0 to 17). The number of points offered ranged from 5 to 20, and the main questions were factual. After a few weeks, participants started exploring community exchange. It started with one participant posting a problem without offering any points in exchange. A couple of participants reacted quickly to the post, clarifying that the tool enabled exchange through a virtual currency. Q: “A simple question, do you really think people want to help you if you do not offer them any points. Think about it …” [GB] A: “Hello, The principle of this site is to monetize your knowledge. Giving points compensates for the time spent answering your question and conducting associated research (which sometimes takes very long). Have a great day! ” [GB]

The poster of the question then responded by articulating an alternative communal rationale for helping: A: “Well, I didn’t think about it, but if this is what motivates you to help others, I don’t think it’s good! But I’ll keep that in mind the next time in case there are a lot of people who think the same!!!! Thanks for the advice ” ☺ [YN]

Other participants resisted this shift, and tried to control the way the technology was being used. They tried to control the definitions and purposes of the technology, arguing that it should be focused on factual exchange rather than on general questions, which were more appropriate for other forums. However, the norms of the community had shifted and people were fluidly combining both forms of exchange. Once the participant introduced this alternative model and articulated its logic, other participants began to combine both market and community exchange. When they asked more general questions, they used social exchange, abandoning the points-based-system. General questions differed from the factual questions in that they required less research, effort, and knowledge. They were related to exams, tests, language training, professional training abroad or more organizational issues that are of concern to the community as a whole, while factual questions focused on the specifics of class content and particular problems an individual within the community is facing. General questions garnered more answers than factual questions with a minimum of 5 answers per question and a maximum of 17 answers. During this first phase personally identifiable display was the norm, during both communal and market interactions. Only 8 newcomers out of a total 130 posted questions in this phase were anonymous. Their questions focused on how to use the tool, whether it was useful, and how other participants used it. By the end of this phase, participants used community versus market-based exchange simultaneously and learned the main differences between the two. They noticed that the key outcome of accumulating points in the market-based exchange was status: ­participants earning points were ranked and became visible in the top ten

4 Results

34

best participants of the site. The experimentation phase thus consisted of not only learning the features of the technology and meeting the other participants, but learning how status could be earned within the new social system. In response to the technology, a new status order was emerging and participants next had to consider strategies to accumulate that status.

4.5.2 Thrill of Competition: Markets Dominate Participants were interested in accumulating virtual points because they sought the visibility of being recognized as one of the top ten contributors. In addition to visibility, participants inferred that these points would affect their participation grades and their winning of prizes at the end of deployment. Indeed, the faculty involved in running the distance learning program did not establish specific norms for using the technology. A presentation introducing the technology features was done at the beginning of the study and a tutorial describing these features was posted on the platform. The directions given to the participants in introducing the technology was: “Wise is an IT tool designed to overcome geographical distance and support your interactions with other participants in the program. Usage is voluntary it’s aimed at enabling you to share knowledge with others on common problems you face. The best participants can also earn prizes such as an iPad or an iPod.”

The attraction of points jeopardized market and community based interactions the co-existing as equals. As is visible in some of the questions on the site, participants’ priorities gradually shifted to point accumulation: “How can we earn more points? And what are the activities that generate more gains?” [Anonymous]

Participants started to find new ways to earn points and communal interactions gradually transformed into market exchanges, where people sought to earn more points than the others. Participants were no longer learning the dynamics of status within the system; they were actively competing for it. During this phase, the word “competition” was a recurrent theme in the comments. It was a war over virtual points, and competition among participants became the underlying engine for participation. Participants also started gaming the tool, revealing the potential for market inefficiencies and failures. They created fraudulent exchanges that led to inflation of the virtual currency. To achieve this, two participants (presumably those with prior relationships) would collude, conspiring to undermine the independence and objectivity of the market process. One participant, for example earned 200 points in a day by asking a trivial question and allocating all points to his fraud partner who just provided a naïve answer. The fraud partner in exchange, asks other questions totaling the same amount (200), and rewards his/her fraud partner. By converting banked points (which do not contribute to the ranking) into earned points, this triggered the inflation process. Participants could thus boost their rankings and

4.5  Analysis and Discussion

35

increase their visibility. One participant humorously drew an analogy between this knowledge market and the financial markets: “I regret a real inflation in the average price of questions on the market (14.66 as I write this comment …). What does the ECB [European Central Bank] do? It is time to increase the federal funds rate before the situation gets worse …” [PN]

In addition to gaming, participants also initiated other forms of market abuse. For instance, they exchanged low quality information to earn points for certain activities like leaving useless comments. All of these attempts to manipulate the market occurred via anonymous display. The community quickly detected these behaviors and aggressively policed it. They did so by comments such: “Is this a joke? Too many points for such a question?” [Anonymous] “This is the fastest question/answer I have ever seen!!” [Anonymous]

The community also voted negatively on the winning answers through manipulation. Despite the social pressure from the community, a few anonymous participants still continued to game the system. To avoid social sanctions, they gamed the market through more subtle techniques, offering smaller amounts of points (but always higher than the average amount in the market). They extended deadlines and waited to receive several answers before selecting and offering points to their fraud partner. Each week, the ranking which displayed the top ten contributors was updated and a new participant appeared in the list because of another act of fraud. Despite the subtlety of their gaming the community could still detect it: “This is a classic case of fraud with an acceptable number of points to remain untouchable. Not bad!” [Anonymous] “I confirm this is a case of fraud and despite the amount used he still made a profit. Have you noticed the flash response? One minute after the question is asked, there is one response and hop the question is solved … I have already explained several times to others that this kind of behavior is unacceptable and it was going to kill the site …” [RF]

Anonymity allowed for fraudulent behavior that could change the rules of the game and support negative behaviors. Several participants explicitly noted their fairness concerns: “There are a lot of techniques to earn points and game the system, I really hope we will not have any participation grade on Wise (though I’m in the top 10)” [Anonymous]

Other participants were more concerned about the social implications of this shift: “This tool is a disaster; it is introducing crazy competition over virtual points … why?” [BC]

4.5.3 Communalism Phase Again: Disillusionment with Inequality/Unfairness After experimenting with both market and community-based exchanges, and witnessing the hyper-competitive market environment, participants gradually shifted back to community exchange. While they continued to use the system and

36

4 Results

considered it as a powerful tool to connect to other participants in the program, to get real time information and receive fast answers to their problems; they felt that the point system created competition, could be easily manipulated, and was not necessary: “Down with the payment system, why are we paying to access knowledge? Knowledge should be free to all!” [NP]

Participants reported these problems directly to the faculty in charge of coordinating the distance-learning program. This external actor, who had initially played a laissez faire role, now intervened in the market, clarifying that participation should be based on quality and frequency of interactions rather than a ‘race over virtual points’: The faculty also explained that this “insider trading” and collusion could be easily detected and disciplined: “It’s not the quantity of points that counts but quality of participation. Points earned by manipulating the system are easily discernible, only real participation counts!” [CP]

This comment established the existence of governance. What was previously a free-wheeling market replete with fraud shifted once participants realized that fraud could be easily detected and they could perhaps be punished. At this point, if incentives were used, they were smaller (phase 3: they didn’t exceed 15 points; phase 2: they didn’t exceed 50 points). Community exchange also garnered more answers and interest than market-based exchange. Participants even displayed altruism, giving away their points to serve communal goals. One of the participants used market exchange, offering 50 points (market average is 27.81), to generate community participation (i.e. votes) to support a school charitable project so that it could win a grant. The community rallied to reach 277 votes through Wise and spread the visibility of the project. In addition, 17 participants commented that they are happy to vote and refused the points. “Hi all! I will give points to everyone who voted for Aeolus on the site below! In short, the company Total sponsors the team that collects most votes! Once you have voted, let me know by leaving a comment and I will award points. So please play the game, it takes only 1 minute, do not pretend to have voted just to get the points. Our school needs to be present in the winning teams, please do it for our school and for us! Thank you in advance!! http://www.team-total.com/fr/vote.html?bourse=Mer Ps: if you could circulate the information to everyone you know attending the program would be great!” [SB]

After the competitive frenzy over points, communal models survived and ultimately dominated at the end of the nine-months. Anonymous postings gradually increased in this phase, 64 % of postings were anonymous compared to 45 % in phase 2. Participants initially used the anonymous option to report problematic and sensitive issues but gradually, hidden identities increased and became the dominant selection. Whereas the communal stage did not involve a single fight over virtual points, there were work-related disagreements, and so participants could criticize and devalue others’ work anonymously.

References

37

References Benbya, H., & van Alstyne, M. (2011). How to find answers within your company. MIT Sloan Management Review, 52(2), 65–75. Garud, R., & Kumaraswamy, A. (2005). Vicious and virtuous circles in the management of knowledge: The case of infosys technologies. MIS Quarterly, 29(1), 25–34. Greenberg, J. (1987). A taxonomy of organizational justice theories. Academy of Management Review, 12, 9–22. Hanson, R. (2003). Combinational information market design. Information Systems Frontiers, 5(1), 107–119. Lee, A. S. (1991). Integrating positivist and interpretive approaches to organizational research. Organization Science, 2(4), 342–365. Servan-Schreiber, E., Wolfers, J., Pennock, D. M., & Galebach, B. (2004). Prediction markets: Does money matter? Electronic Markets, 14(3), 243–251. Wolfers, J., & Zitzewitz, E. (2004). Prediction markets. Journal of Economic Perspectives, 18(2). Zingales, L. (2000). The governance of the new corporation. In X. Vives (Eds.), Corporate governance theoretical and empirical perspectives. Cambridge. Zingales, L. (2010). A market-based regulatory policy to avoid financial crises. Cato Journal, 30(3), 535.

Chapter 5

Implications for Theory and Practice

We first address the implications for theory, and then consider some more practical, organizational implications.

5.1 Implications About Incentives in Knowledge Markets Our findings support and underscore the potentially pernicious effects of material incentives (Heyman and Ariely 2004). The key question that past research has raised is that paying people to do things that they would otherwise do for free can undermine intrinsic motivation (Lepper and Greene 1978; Deci and Ryan 1985; Frey and Jegen 2001) and social norms for cooperation (Gneezy and Rustichini 2000). In addition to supporting these arguments, our research additionally shows the dynamic processes by which markets and communities can transform themselves within firms. By witnessing more than 5,000 interactions over time, we are able to build on past research in two ways. First, past research focuses on motivation as an individual phenomenon. By contrast, our findings indicate how one individual’s behavior can be contagious in a society. Specifically, our on-line ethnography shows us a clear record of how people become disillusioned, how that behavior is publicly communicated, how others perceive it and how they come to either repudiate it or imitate it—thereby creating a social norm with respect to demanding or repudiating incentives. Second, and more optimistically, our findings are the first, to our knowledge, that shows the possibilities for withdrawing from market incentives and regaining communal motivation. Our participants eventually became disillusioned with the market mechanism and initiated a social movement where they ended market manipulation and began the norm of giving knowledge away for free. Importantly, this free knowledge was not lower in quantity or quality, suggesting that they were not disillusioned with participation in general, but with the market in particular. Whether it is the history of economic development (which has observed transitions from communist to market regimes), or psychological research (which shows a transition from intrinsic orientations to financial reward) we have not found © The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_5

39

40

5  Implications for Theory and Practice

any research that depicts how people can transition away from their reliance on markets and incentives. Our research documents this possibility, and offers some pre-conditions. Finally, this research takes the metaphors of knowledge community and ­knowledge market seriously. Specifically, by creating an actual knowledge market, we create various phenomena that resemble dynamics in financial markets. We see market manipulation through collusion, inflation, and market interventions to control fraudulent practices. We suggest that markets and communities are not merely metaphors for knowledge sharing, they can be created in organizations through technology.

5.2 Implications of Identity First, by analyzing 5,000 interactions, we identify several unique and contrasting reasons why people contribute anonymously. Paradoxically, those who are most likely to seek credit for themselves (i.e. Market manipulators in Phase two) and those who do not wish to claim credit (i.e. communal helpers in Phase 3) are both using anonymous interaction. This evolving significance of the signals reveals that how a technological affordance is used and perceived is socially constructed. Additionally, these findings help navigate between the conflicting evidence on when identity display is appropriate in organizations. On one hand, unethical behavior (i.e., the market manipulation, fraud, and revenge) was facilitated through anonymity. On the other hand, transitions to new social norms often began with an anonymous actor. That actor could publicly voice a subversive message, but mitigate the personal risks of revolutionary action with privacy. In organizations, such anonymity can allow people to voice conflicting opinions that can break down groupthink. By allowing people to challenge norms, anonymous interaction becomes the medium through which social change is communicated and triggers regime shifts.

5.3 Implications for Practice Our research offers several practical implications for designing IT for knowledge sharing. In particular, we assert that IT systems cannot be simply viewed as computer systems: they are also relational systems. First, our research highlights the perverse effects of using material incentives inside organizations. We suggest that people engaged in knowledge sharing should be rewarded with incentives other than money. In particular, virtual currencies could quantify reputations and reveal the status of helpers in the organization. For instance, a technology company, (SAP) has developed an innovative point system to reward its developer community for trouble-shooting, blogging, and answering each other’s questions. Participants earn reputation benefits based on their level of

5.3  Implications for Practice

41

contributions and can showcase their status with badges assigned to their name. A reputational point system would recognize people, without undermining their basic human and social motivations to help and cooperate. Still, any ranking system—reputational or financial—could create competitive dynamics. Our premise is that competition can have positive implications in terms of motivating and inspiring people to contribute more. The key issue is managing the competition to control its pernicious consequences. Indeed, as with any new social system, people are looking for status and ways to personally benefit, perhaps through manipulation. This does not invalidate the system itself, but it highlights the importance of design, governance, and regulation. For instance, a fair basis for ranking should account, not only for virtual points earned through participation, but also for community evaluations of the contributions. This incorporates quality—and not just quantity—of participation. Further, this tool does not involve governance from a centralized authority: control is “crowd sourced”. In our case, the community recognized the fraud, complained about it, but did not have a mechanism to flag it or punish it. This design element gives the system a dynamic self-corrective mechanism. The crowd can identify and correct manipulation. Further the perpetrators experience reputational consequences, which control potential contagion of bad behavior. Likewise, consider the organizational implications of allowing anonymous interaction in knowledge sharing. If anonymity facilitates revolutionary action— where a private actor could theoretically create a social movement with hundreds of supporters—its use should be carefully controlled in organizations. Rather than allowing this option to become a norm, it may be deployed in specific and controlled circumstances where decision makers seek honest feedback.

References Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-regulation in human behavior. New York: Plenum Press. Frey, B. S., & Jegen, R. (2001). Motivation crowding theory: A survey of empirical evidence. Journal of Economic Surveys, 15(5), 589–611. Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies 29(1), 1–18. Heyman, J., & Ariely, D. (2004). Effort for payment a tale of two markets. Psychological Science, 15(11), 787–793. Lepper, M., & Greene, D. (1978). The hidden cost of rewards. Hillsdale, NJ: Erlbaum.

Chapter 6

Contributions to the 2020 Perspective

The need for managing knowledge and developing innovative capabilities is consistently a top business priority among CEOs. Given the need for continuous improvements and new products developments, firms started to outsource several activities to a large, undefined group of people external to the company (i.e., customers, suppliers, and partners). One way to profit from external distributed knowledge is to rely on a market-mechanism. Several organizations rely on external markets to gather ideas for new products and services, to solve their knowledge problems or make decisions. These firms, however, missed the opportunity to bring the benefits of markets within firms. Employees may have solutions to several problems the organizations faces and be a source of ideas for innovation. They can use formal or informal channels to push their ideas through the decision makers. One such channel is an internal market. An internal market is a platform for matching users to experts and expertise within the organization. They facilitate reuse of exiting knowledge, help find solutions to unsolved problems or predict a wide range of events. These markets frequently outperform experts as they enable broader participation. They provide powerful tools to help large companies address the problems of accessing their dispersed information and talent. At the same time, they present a number of challenges. Many observers are concerned that such markets inside organizations could misdirect time, money and credit. If a company relies on real money to favor participation, employees may shirk on other tasks or choose market topics in part to reward their friends. If on the other hand, small or no incentives are offered, employees will not bother to participate. These difficulties with creating explicit reward schemes in internal markets are ubiquitous in most organizations. The objective of our research is to explore how we can move markets inside firms. Specifically, we strive to understand the design guidelines and effects associated with internal knowledge markets. To understand how firms can profit from internal markets, we studied the benefits and costs of traditional versus market based approaches. We investigated firms that adopted internal markets, and we

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4_6

43

44

6  Contributions to the 2020 Perspective

designed a working prototype to test its various effects. Our research provides rich insights on the design and effects of internal knowledge markets. First, we find four valuable uses of internal markets within firms: to support knowledge exchange “knowledge markets”, problem-solving “innovation markets”, Ideas lifecycle management “idea markets” and decision making “decision markets”. Firms initially experimented with prediction markets, their success in forecasting sales and project management has stimulated interest in their application to other business problems. The other types of markets are now widely used and often play an important role, although their popularity varies across sectors and regions. Our research reveals several benefits of internal markets. We find that access to knowledge produced and gathered within the same organization affects the performance of knowledge workers. Knowledge markets contribute also to saving costs. SAP, for example, have seen positive results with its knowledge market as it saved more than $6 million in technical support costs. In addition, they can support value creation in the form of new products and services. Introducing an internal market in Allianz for instance generated over 25,000 ideas have been implemented and have delivered an annualized financial value that exceeds £12.5 million ($19 million). However, introducing an internal market opens a company to new forms of governance and creates completely new roles for those who would manage expertise. The 2020 company is an open organization where employees can deliberate, argue, vote, compete, and collaborate horizontally across fields of expertise. A company where executives must change not only their tools and how their people interact, but also change and empower workers to influence company choices. But to make knowledge markets work managers need to provide participants with appropriate incentives. While material rewards stimulate market participation they also cause manipulation and market fraud. Our project highlights the perverse effects of using material incentives inside organizations. We see market manipulation through collusion, inflation, and market interventions to control fraudulent practices. We suggest that material rewards should be used only a short-term basis to run competitive campaigns but should be supported with an appropriate governance structure. Over the long turn, however, virtual currencies could quantify reputations and reveal the status of helpers in the organization. A reputational point system would recognize people, without undermining their basic human and social motivations to help and cooperate. The 2020 organization is also a company where fairness should dominate. Since it’s an open organization where employees participate to decision making, problem-solving and innovation, executives face a decision about if and how to provide in-process feedback to the contributors. On the one hand, such feedback could help avoid effort wasted exploring impoverished territory. On the other hand, feedback could over-determine the process, providing guidance that precludes exploration of unlikely but potentially valuable directions.

6  Contributions to the 2020 Perspective

45

Adopting clear and transparent validation mechanisms based on crowdsourcing is key to avoid people disillusion. If employees feel that the processes associated to content validation is unfair, they will publicly communicate their disillusion, others will perceive it, they will come to either repudiate it or imitate it—thereby creating a social norm with respect to participation in internal markets.

Conclusion

Internal markets provide powerful mechanisms for enhancing knowledge sharing and provide remarkable sources of ideas and innovation. To bring these virtues inside, firms ranging from Google to Microsoft and Best Buy to Peugeot are experimenting with idea markets for new product development, prediction markets for working business decisions, and innovation markets for problem solving and knowledge markets for internal knowledge exchange. Through internal markets, executives can get better forecasts, develop new products, and boost productivity. Markets can self-adapt to employee needs and help balance the best use of everyone’s time. Introducing such markets, however, depends on shifting executive mindset from central planning to market making. To succeed, we offer a blueprint with design guidelines to consider the main challenges associated with their adoption and analysis of some of their effects.

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4

47

SpringerBriefs on Digital Spaces

SpringerBriefs on Digital Spaces is an international research program—the ISD— launched in 2009 by the CIGREF Foundation (www.fondation-cigref.org). The series aims at making a set of concepts, ideas and results of projects carried out under the program available to the research, business and policy communities. Information Systems Dynamics (ISD), is a research program of public interest that works to evaluate the societal and managerial challenges related to the long-term use of information systems and digitality. Since its launch in 2009, the program has already supported more than 30 projects conducted by international teams from different academic backgrounds (Computer Science, Management Science, Economics, Sociology, Geography and Anthropology) as well as from different geographical regions (Europe, North America and Asia). The program works on the premise that the spatial dimension of the use of ­digital systems and artefacts is a critical perspective for understanding the dynamics of value creation—and more generally of socio-economizing—in our economies and societies. Understanding emerging practices in digital spaces is a key step toward delineating and conceptualizing a substantial part of the emerging paradigms of economic activities in the twenty-first century. Springer Briefs in Digital Spaces publishes research findings and monographs related to the different facets of these issues. By doing so, the series seeks to contribute to the necessary dialogue between the researchers, practitioners and public policymakers involved in these very critical and rapidly changing fields of research and action.

Editor The series is edited by Ahmed Bounfour, Professor, European Chair on Intellectual Capital Management, University Paris-Sud, and General Rapporteur of the ISD program.

© The Author(s) 2015 H. Benbya, Exploring the Design and Effects of Internal Knowledge Markets, SpringerBriefs in Digital Spaces, DOI 10.1007/978-3-319-14517-4

49