The Rolodex paradox : the effects of ties to venture capitalists on internet startup survival

Research on interorganizational relationships has established that organizations gain essential information and resources via their network ties. Less studied, however, are the costs and benefits of different types of network ties. This dissertation explores the effects of direct and indirect network ties on the survival of startups. Drawing on research on interorganizational relationships and organizational learning, I argue that direct and indirect ties have different effects on survival. Specifically, while network ties create access to resources, information, legitimacy, and reputation, they also come with maintenance and information processing costs. As the number of direct ties increases they become redundant and contribute diminishing benefits at the margin. At the same time, direct ties involve substantial costs, which will overwhelm the marginal benefits of additional direct ties eventually, leading to unfavorable outcomes. Therefore, the number of direct ties follows a U-shaped relationship with startup failure likelihood. Similarly, as the number of indirect ties increases, redundancy leads to diminishing marginal benefits. In this case, however, the costs of maintaining extra ties are negligible. As a result, having more indirect ties is expected to be associated with lower failure likelihood. Furthermore, agglomeration provides alternative sources of resources and information, and may dampen network effects. I focused on direct ties among U.S.-based business-to-business Internet startups and their venture capitalists (VCs) and two-step indirect ties between Internet startups from 1988 to 2001. I found that direct ties reduced the likelihood of startup failure, as did indirect ties. However, when indirect ties are accounted for, the effects of direct ties disappear, suggesting the mediating role of indirect ties. Having more direct ties benefited startups because they provided connections to more indirect ties. Ultimately, it was the number of indirect ties, and not the number of direct ties, that mattered. In addition, agglomeration introduced intense competition that led to higher startup failure likelihood but did not moderate network effects. These findings confirm that both network ties and agglomeration affect startup failure. More importantly, the results reveal a rolodex paradox – startups that are tied to well-connected others, regardless of the size of their own rolodexes, enjoy lower failure likelihood.