Building the Momentum: 
Information Disclosure and Herding in Online Crowdfunding

Abstract

In online crowdfunding markets, backers face high uncertainty about the quality of a campaign. To mitigate such uncertainty, crowdfunding platforms often allow campaign creators to post communicative messages—i.e., campaign updates and creator comments—to dynamically disclose further information about the campaigns. In addition, previous funding transactions of ongoing campaigns are made publicly available, giving rise to herding among backers. In this research, we aim to understand how communicative messages and herding interactively shape the behavior of backers contributing to crowdfunding campaigns. Our results show that the frequency of communicative messages has a positive effect on backer contributions; however, it attenuate successors’ herding momentum towards predecessors, perhaps because the information disclosed in those messages lowers the informational value of previous funding transactions. To investigate the role of message contents, we extract topics addressed in update and comment messages using a Latent Dirichlet Allocation model. The results reveal that distinct messages have different impacts on backers’ contribution and herding behavior, and such discrepancies are found to be topic-specific. This study not only contributes to operations management literature on crowdfunding but offers implications for campaign creators and platform managers.

Keywords: crowdfunding; communications; herding, observational learning; topic modeling; instrument variables