Three Essays on Crowdsourcing for Technical Problem Solving
Pollok, Patrick; Piller, Frank Thomas (Thesis advisor); Salge, Torsten Oliver (Thesis advisor)
Dissertation / PhD Thesis
Dissertation, RWTH Aachen University, 2017
Crowdsourcing has received considerable attention in management research and practice in recent years (Malhorta and Majchrzak, 2014). Today, many crowdsourcing initiatives are governed by specialized intermediaries who provide internet-enabled communication infrastructure and help clients gain access to the wisdom and creative potential of large and unknown populations of external contributors (Dahlander and Piezunka, 2014; Diener and Piller, 2013; Lopez-Vega et al., 2016; Surowiecki, 2004). While prior research has mainly focused on crowdsourcing of ideas and needs-related knowledge from ordinary users and customers, only little work has been done on crowdsourcing for technology-related solutions to innovation problems from professional solvers. The objective of this dissertation is to examine how intermediary-administrated crowdsourcing as a problem solving approach in innovation can be implemented, developed into an organizational capability, and effectively applied at the project level in knowledge seeking firms. I use a mix of qualitative and quantitative research methods to address these aspects in a series of three independent papers. The first paper investigates the implementation process of crowdsourcing in R&D-intensive firms. Following a design science research approach, I observe a set of crowdsourcing pilots in seeker firms and identify implementation barriers. I then propose and discuss an organizational intervention to overcome the observed barriers, and test the intervention in another set of pilot cases in the second research phase. I find that the active involvement of individuals with promoter characteristics helps overcome barriers and challenges in crowdsourcing pilot projects that may ultimately prevent the repeated use of the approach in firms. Results of the study suggest that seeker firms should actively search for and involve promoter roles in crowdsourcing pilot projects to ensure effective implementation of the method. The second paper examines the organizational antecedents of firm-level crowdsourcing capability. Although, in theory, the crowdsourcing mechanism should provide considerable advantages compared to other forms of innovation governance, most notably a quick and relatively low cost access to innovation-relevant knowledge, recent studies indicate that seekers differ considerably in their ability to reap the benefits from innovation crowdsourcing in the long run. Based on work on the microfoundations of organizational capabilities, I hypothesize how three lower-level organizational elements affect the development of crowdsourcing specific capabilities at the firm level. Drawing on survey data from a sample of 365 manufacturing firms and qualitative data from 26 managers of 14 crowdsourcing seeker firms, I find that organizational roles, investments in dedicated organizational structures, and deliberate efforts to articulate and codify crowdsourcing know-how (i.e. knowledge processes) are important antecedents of a seeker firm's crowdsourcing capability. More importantly, my results indicate that knowledge processes act as important mechanisms through which organizational roles and structures influence capability development. In the third paper I theorize about the factors that affect solvers' decision to participate in crowdsourcing contests and the actions seekers can take to exert influence on this decision. I argue that solvers' attention allocation to crowdsourcing is driven by the comprehensiveness of the problem definition, which again is contingent upon the stock of problem-related knowledge available to the seeker when crafting the problem statement. I use data of 635 crowdsourcing projects hosted by the crowdsourcing intermediary NineSigma to test the proposed theory. I find that search distance, proxied by the cosine between IPC-classified technical problems and seekers' patent portfolio vector, exhibits an inverted U-shaped relationship with contest participation. This result challenges the conventional view that crowdsourcing is a solution to distant search, and suggest that seeker firms can overshoot in terms of search distance when soliciting solutions from external solvers. I also find contingent effects. Significant two-way and a three-way interactions indicate that identity disclosure in crowdsourcing requests substantially improves contest participation, and that this effect is even stronger for high-status seeker firms. In line with research on status (Azoulay et al., 2014; Simcoe and Waguespack, 2011), my findings point to the important role of identity information in signaling status and indicate the presence of a Matthew Effect in innovation crowdsourcing.