Crowdsourcing "blockbuster" ideas: A dynamic structural model of ideation
Yan Hua Huang, Param Vir Singh, Kannan Srinivasan · International Conference on Information Systems · 2011
Abstract Crowdsourcing initiatives are becoming an increasing popular tool for new idea generation for firms. Although such initiatives are widely adopted in many different industries, the number of ideas generated often decline over time and the implementation rates (percentage of posted ideas that are implemented by the firm) are very low. Critics of crowdsourcing often attribute these observations to users’ restrictive view about firms’ products leading to contribution of mainly niche ideas, and limited knowledge about firms’ cost structure leading to contribution of mostly infeasible ideas. To investigate these criticisms in detail and to devise policies for firms to alleviate these concerns, we propose a structural model to capture users idea contribution dynamics. We estimate the model using a rich dataset obtained from Ideastorm.com which is a crowdsourcing website affiliated with Dell. Using the peer voting score we are able to infer out the true potential ideas, whereas a firm's costs of implementation are indirectly imputed from the idea implementation data. We find that users tend to significantly underestimate firm's costs of implementation of their ideas while surprisingly also slightly underestimate the potential of their ideas in the beginning of the posting history. Because the underestimation of cost is much larger than the underestimation of potential, users tend to post very few viable ideas. However, users learn about their abilities to come up with potentially ideas and the cost structure of the firm through a Bayesian fashion from peer voting on their ideas and firm's response to all posted ideas. We find that the users learn very quickly about the firm's cost structure but the learning regarding their abilities to come up with blockbuster ideas is quite slow. As a result of the learning process, contributors of low potential ideas eventually drop out, and high potential idea contributors remain active. Over time, the average potential of generated ideas increases, while the number of ideas created decreases, and the firm can reduce the cost of screening ideas without losing high potential ideas. Through a policy simulation, we show that the firm can significantly increase the learning rate of the users regarding their abilities to contribute ideas by running short term promotions where the users whose ideas are implemented are given very high incentives.