Social network research has demonstrated increasing interest in individual agency. In this study, we adopt a dynamic perspective and a multilevel approach to investigate how individuals’ networking behaviors develop over time. We propose the concept of networking efficiency, capturing the successful conversion of the shared opportunities (i.e., referral) to business revenue. We argue that the efficiency of individual networking develops over time through a learning process, with individual engagement in the network (agency) and overall change in network composition (structure) accelerating or hindering this evolution. In a longitudinal empirical study of 5371 members across 81 networks in a large business association over 13 years, we find that while individuals share fewer business referral over time, the revenue generated from shared opportunities increases, thereby showing an increase in individual networking efficiency over time. Furthermore, we find that individuals’ active engagement in the network accelerates the learning process of efficient networking. At the network level, while new members joining and old members leaving both slow down the learning process, they reduce the efficiency of networking in different ways. By elucidating the dynamic of individual-level networking behaviors and their interaction with network-level renewal, our work contributes to the literature on network agency, network dynamics, and cross-level network effects.
