Beyond Related Experience : Upstream versus Downstream Experience in Innovation Contest Platforms with Interdependent Problem Domains

Innovation contest platforms are often organized around specific fields of interest and involve contests that span a variety of interdependent problem domains. While specialization in contests within a problem domain can improve a contestant's performance in a future contest in the same domain, prior research on individual learning also suggests that a contestant can benefit from her related experience—i.e., experience in contests whose problem domains share an interdependency with the focal problem domain. It is, however, unclear whether the benefits of related experience arise symmetrically from upstream related problem domains (i.e., upstream experience) and downstream related problem domains (i.e., downstream experience) or differ among them. Given that an innovation contest platform serves to provide an effective match between contest problem requirements and contestants' skills, it is important to understand how a contestant's prior experience on such a platform contributes to her problem-solving performance. To address the research question, we collect detailed archival data from TopCoder, a leading platform for software development contests, from its launch in 2001 to September 2013. Our dataset comprises detailed participation history of 821 contestants in 3,274 contests across eight interdependent problem domains involving 8,985 observations. We find that while a contestant's related experience on the innovation contest platform has a stronger positive association with her focal contest performance compared to unrelated experience on the platform, the benefits of related experience arise only from downstream experience. No significant performance benefits of upstream experience are observed. We also find that the performance benefits of downstream experience are more salient when the contest duration is shorter. Contrary to the notion of “hyper-specialization” that has been emphasized in previous research, our findings highlight the importance of targeted diversification of experience for participants on innovation platforms. They also shed light on the effective design of innovation contest platforms and resource allocation decisions in complex new product development projects involving interdependent problem domains