Social network analysis is an increasingly popular method for analyzing relational data in animal social systems. The abstract nature of network metrics primes them for use in cross-comparisons of social systems, but more work is needed to determine how well such measures can approximate meaningful biological properties. Finding biological correlates of network metrics, and extending the existing network methods to include the analysis of ongoing dynamics of social processes, will bring us closer to standardization of the terminology used to describe animal social and interaction networks. This will allow us to use the social networks in simultaneously studying the individuals embedded in a social setting, the consequences of their interactions, and the global properties of the social system. We discuss how certain facets of the existing network methods need be further developed to fulfill this potential and provide a multi-scale systems approach to the studies of animal sociality.