Across animals, there is remarkable diversity in behavior. Modern genomic approaches have made it possible to identify the molecular underpinnings of varied behavioral phenotypes. By examining species with plastic phenotypes we have begun to understand the dynamic and flexible nature of neural transcriptomes and identified gene modules associated with variation in social and reproductive behaviors in diverse species. Importantly, it is becoming increasingly clear that some candidate genes and gene networks are involved in complex social behaviors across even divergent species, yet few comparative transcriptomics studies have been conducted that examine a specific behavior across species. We discuss the implications of a range of important and insightful studies that have increased our understanding of the neurogenomics of behavioral plasticity. Despite its successes, behavioral genomics has been criticized for its lack of hypotheses and causative insights. We propose here a novel avenue to overcome some of these short-comings by complementing "forward genomics" studies (i.e., from phenotype to behaviorally relevant gene modules) with a "reverse genomics" approach (i.e., manipulating novel gene modules to examine effects on behavior, hormones, and the genome itself) to examine the functional causes and consequences of differential gene expression patterns. We discuss how several established approaches (such as pharmacological manipulations of a novel candidate pathway, fine scale mapping of novel candidate gene expression in the brain, or identifying direct targets of a novel transcription factor of interest) can be used in combination with the analysis of the accompanying neurogenomic responses to reveal unexpected biological processes. The integration of forward and reverse genomics will move the field beyond statistical associations and yield great insights into the neural and molecular control of social behavior and its evolution.