UC San Diego

Firms are increasingly organizing contests for crowdsourcing solutions to problems associated with sustainable growth. The effectiveness and efficiency of these contests relies on the decisions taken at the operational level by the firm. The current dissertation consists of three essays related to this theme.

In the first essay, I focus on the case where a firm wants to leverage the latent employee expertise across the organization, going beyond core R & D teams. Such intra-enterprise design contests offer a contrast with external contests where firms pose questions and engage an outside world of solvers not contractually employed by the firm. Internal contests, where solution-seeking firms internalize some or all of the participant opportunity costs, pose new opportunities and challenges, managing which forms the focus of this paper. While internal contests allow firms to pursue proprietary concepts, I learned from field studies that employee participation in internal contests requires carefully coordinated decision making. How can managers architect contests by setting participation cost-sharing policies, contest awards and specifying problems to ensure optimal outcomes from contests? An application-driven model is formulated and analyzed for managerial decisions and insights about when to resort to internal contests, how to structure and specify them, and when and how to cover the participant opportunity costs. I find that budget constraints combined with intellectual property considerations make internal contests increasingly attractive for solution-seeking firms. Interestingly, the contest reward in internal contests is curvilinear in the design problem uncertainty. Solutions of better average and extreme quality can be obtained with an internal contest under budget constraints. Firms with larger contestant pools can share more of their opportunity costs with employees, making internal contests more appropriate for larger firms. Also, a contest designer can under most cases exert less effort to specify the contest challenge problem in an internal contest relative to posing the problem to an external group. Curiously, it is optimal for open ended contest problems to be organized with smaller rewards in internal contest settings. We discuss how managers of solution-seeking firms can benefit with the careful design of innovation contests.

In the second essay I employ the perspective that innovation is a multi-stage process

that may be usefully split into an exploration phase followed by an execution phase. I analyze the open innovation setting

where a seeker sources innovations from outside solvers through a contest. The interaction between exploration and execution phases leads to insights into the effort decisions by agents competing in the contest. While the execution efforts follow the usual trends with costs and rewards, the implications on the optimal exploration efforts are more interesting. The model predicts that from the seeker's perspective, increasing the reward beyond a threshold can decrease the probability of finding the solution. I also use this exploration-execution model to study contest architecture along two dimensions (i) Reward structure and (ii) Information structure. Recommendations are provided for contest design based on the objective of the innovation seeker.

In the third essay, I focus on the individual and study the process that she employs while searching for a solution associated with creative problem solving. The objective is to start a research program focused on understanding behavioral patterns while performing tasks requiring creativity. The knowledge gained can then be used to create an optimized incentive and task design for increasing efficacy of innovation sourcing mechanisms. I use laboratory experiments to collect data on the process path adopted by participants while solving a unique insight problem on the computer. I observe that piecerate incentives lead to a better performance on the puzzle relative to an equivalent incentive based on competition. Analysis of the process path employed by participants shows that they spend less effort on the puzzle in the competition case and spend more effort when incentivized with an equivalent piecerate incentive. The performance on the puzzle task is observed to increase with an increase in the time spent on the puzzle. Several other process measures are defined and compared across treatments.