INTRA-CORPORATE CROWDSOURCING (ICC): LEVERAGING UPON RANK AND SITE MARGINALITY FOR INNOVATION J. Andrei Villarroel, and Filipa Reis Catholic University of Portugal Palma de Cima 1649-023 Lisbon
[email protected] [email protected] ABSTRACT Our study of a crowdsourcing innovation initiative inside a large multi-business firm shows that: (1) rank marginality —defined as being lower positioned in the corporate hierarchy—, and (2) site marginality —defined as being spatially distant relative to the corporate epicenter—, are positively associated with better innovation performance. Prior research showed the positive association of marginality, technical and social, with better innovation performance in crowdsourcing initiatives involving participants external to the firm [6]. Our findings contribute to the management of innovation literature by identifying two new dimensions of marginality that may help create an innovation performance advantage for the firm implementing intra-corporate crowdsourcing initiatives.
Categories and Subject Descriptors K.4.3 [Computers and Society]: Organizational Impacts – computer-supported collaborative work, reengineering.
well as issues related to the transfer of tacit knowledge impose limitations on the approach [14]. Seeking to alleviate concerns regarding the transfer of knowledge and the associated intellectual property rights, several innovation services companies developed crowdsourcing products for client firms to deploy inside their own corporate walls: e.g. InnoCentive@Work, Imaginatik, Exago Markets, etc. By implementing internal crowdsourcing innovation initiatives, companies are relieved from their concern with appropriability of the ideas generated [11]. Furthermore, open collaboration can be encouraged to diminish the barriers of information stickiness [14]. Our study explores the effectiveness of an internal crowdsourcing innovation initiative in a large multi-business firm. To the best of our knowledge, there is little research addressing this phenomenon, and current understanding is limited regarding the underlying sources of innovation advantage in intra-corporate crowdsourcing initiatives of this kind.
2. A STOCK MARKET FOR INNOVATION General Terms Management, Theory.
Measurement,
Performance,
Human
Factors,
Keywords Corporate Crowdsourcing, Marginality Advantage, Innovation Management, Firm Boundaries.
1. INTRODUCTION An acknowledged example of crowdsourcing is InnoCentive, a for-profit company that makes crowdsourcing innovation services its core business. The company offers an online platform that client companies use to broadcast their problems to an online community of solvers distributed around the world. This distributed innovation model external to the firm has been the subject of a growing body of research defending the effectiveness of the approach ([6], [7], [12]). However, issues with intellectual property management -verification and transfer thereof [11]-, as Copyright is held by the author/owner(s). CrowdConf 2010, October 4, 2010, San Francisco, CA..
The Stock Market for Innovation (SMI) is an online innovation platform which replicates some features of a financial stock market. Company employees participating in the SMI can post their own ideas or speculate on ideas posted by peers. An idea is akin to equity owned by an individual contributor, upon which others can choose to invest using a virtual currency. Participants in the SMI collaborate by commenting upon an existing idea, both to suggest improvements to it as well as to challenge it. At any point in time, the spot value of an idea – together with the comments that support it– is proxied by the aggregate investment positions held on it relative to all other ideas in the market. The spot value of the idea represents the market’s belief in it. An idea ‘submitted’ to the SMI first goes through a process of ‘validation’ to ensure originality and clarity. A validated idea is then actively traded in the SMI. After the trading window expires, investors in the idea realize the gains or losses of their investment. An idea with a belief value higher than a specified threshold is ‘approved’. An idea approved by the market becomes candidate for implementation.
3. THEORY Over the course of the last decade, the literature on innovation management devoted much attention to new open models of
innovation that facilitate knowledge exchange among the different actors in the innovation process. One stream of research has set forth the basis for a theory of open innovation ([2], [3], [4]) referring to firms implementing an open trade of intellectual property. Another research stream has focused on open source processes ([5], [16], [17]), cumulative innovation [9], and the inclusion of online communities in the innovation process ([7], [18]). The SMI subject of this study lies in between these two research streams whereby online communities interact to generate, trade and cumulate ideas that lead to innovation for the benefit of the firm [13]. Laursen and Salter identified a strategic change in how firms across a variety of industries are searching for new ideas, “adopting open search strategies that involve the use of a wide range of external actors and sources to help them achieve and sustain innovation” [8]. In this same direction, Jeppesen and Lakhani showed the effectiveness of ‘broadcast search’ as a novel mechanism for innovation [6]. Through their study of InnoCentive, they identified that an important source of novelty in this competitive process came from marginal individuals. These marginal individuals have different knowledge than the main actors in the field of the problem, and hence provide truly novel solutions to otherwise unsolved problems. This has been referred to as ‘the advantage of marginality’ [6].
3.1 Definition: intra-corporate crowdsourcing Intra-Corporate Crowdsourcing (ICC) refers to the distributed organizational model used by the firm to extend problem-solving to a large and diverse pool of self-selected contributors beyond the formal internal boundaries of a multi-business firm: across business divisions, bridging geographic locations, leveling hierarchical structures. The Stock Market for Innovation (SMI) subject of the present study is an intra-corporate crowdsourcing initiative. The firm we studied uses the SMI to broadcast innovation tasks –traditionally allocated to a specific unit of the firm, e.g. innovation division, corporate headquarters– to all employees of the firm. These latter contribute their ideas on a voluntary basis, from across business units, geographic locations, and hierarchical positions.
3.2 The ICC marginality advantage Firms implementing online innovation platforms featuring search, interaction and collaboration tools, seek to facilitate access to individuals and sources of knowledge from across the four corners of the organization. This democratization of innovation [15] allows all individuals –equally throughout the corporation– to make use of their diverse perspectives and heuristics ([6], [10]) to potentially make an effective contribution to the innovation process. In this research, we hypothesized two types of intracorporate marginality as potential sources of innovative advantage: rank marginality – defined as being ranked lower in the corporate hierarchy – and site marginality – defined as being spatially distant from the corporate innovation epicenter. Lower ranked employees in the corporate hierarchy typically have lesser interaction opportunities with top-level management and hence have little say in traditional innovation processes. At the same time, lower ranked employees face customers and deal with operational problems on a daily basis. As a result, lower ranked employees may hold unique information that may prove valuable
in coming up with novel solutions to existing problem. The SMI provides all employees of the firm an open channel to contribute to the innovation process. While higher ranked employees with greater influence over the innovation direction of the firm may be likely to prefer traditional channels, lower ranked employees may find this alternate channel an effective mechanism in getting their ideas noticed. H1: Contributors ranked lower in the corporate hierarchy perform better at innovation. Similarly, sites that are physically distant from the company’s officially acknowledged centers for innovation, such as the corporate headquarters or innovation units –here referred to as the corporate epicenter– would traditionally have little say regarding innovation in the firm. At the same time, employees in distant locations face customers and deal with operational problems that may differ from those in other locations. As a result, employees in sites that are far may hold unique knowledge that may prove valuable in coming up with novel solutions to existing problems the firm faces. The SMI offers all employees, regardless of their site location, with equal opportunity for contributing to the innovation process. While employees in the epicenter of the company are already formally involved in the traditional innovation process, employees in distant sites may find the SMI a more effective mechanism to have their unique ideas heard. H2: Contributors spatially distant from the corporate epicenter perform better at innovation.
4. METHOD AND DATA This study was developed in collaboration with a company implementing a common online crowdsourcing platform for innovation across its multiple business units. It is a major information and communication services provider headquartered in Europe. The scope of the data and related analysis discussed in this paper encompasses permanent employees of the company in one country in continental Europe.
4.1 Data and Variables We used two sources of data in our study: (1) historical data on the performance of ideas submitted to the SMI over an eightmonth period, and (2) survey data of a sample of active SMI contributors. The variables are described in table 1. Table 1: Descriptive statistics of all variables in the study Variable performance age gender education tenure rank rank_high rank_medium rank_low site site_epicenter site_near site_far
N 370 370 370 370 370 370 370 370 370 370 370 370 370
mean 5.25 36.37 0.35 3.17 4.98 3.64 0.18 0.66 0.16 14.20 0.84 0.14 0.02
stdev 15.36 8.03 0.48 1.31 2.94 1.26 0.39 0.48 0.37 37.42 0.37 0.35 0.15
min 0 22.00 0.00 1.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
max 184.6 57.00 1.00 7.00 12.00 7.00 1.00 1.00 1.00 278.00 1.00 1.00 1.00
4.1.1 Dependent variable
5. RESULTS
As dependent variable we considered individual innovation performance. This variable refers to the cumulated portfolio of innovations achieved by each participant. As stated in section 2, ideas go through a three-stage process as they progress in the SMI. We attributed a growing weight to each stage, commensurate with the number of ideas that get past each stage in the process.
The results of the statistical regressions associated to the models described in section 4.2 are presented in Table 2.
performance = ideas submitted + 1.3 × ideas
validated
+ 3.5 × ideas approved
(1.1)
This variable has a negative binomial distribution which matches the type of regression used in the statistical analysis.
4.1.2 Independent variables Organizational hierarchy. We considered hierarchy in two ways: (i) as a continuous variable rank reflecting the company’s hierarchical structure, coded from lowest level (support technician) to highest level (board director) in the range 1 to 7 respectively, and (ii) as three dummies: rank_high for managers, executives, and directors; rank_medium for professionals and specialists; and rank_low for technicians and support technicians. Geographic location. We considered geographic location in two ways: (i) as a continuous variable site reflecting the contributor’s distance [in km] to the corporate headquarters or to the firm’s main innovation locations, and (ii) as three dummies: site_epicenter for corporate headquarters or main innovation sites; site_near for sites located within 100 km from the corporate epicenter; and site_far for sites located more than 100 km from the corporate epicenter.
Model 1, contrasting the effects on innovation of different ranks in the corporate hierarchy, shows that rank_low is statistically significant (p
