Most species pairs are isolated through the collective action of a suite of barriers. Recent work has shown that cryptic barriers such as conspecific sperm precedence can be quite strong, suggesting that they evolve quickly. However, because the strength of multiple barriers has been formally quantified in very few systems, the relative speed with which conspecific sperm precedence evolves remains unclear. Here, we measure the strength of both conspecific sperm precedence and cryptic non-competitive isolation between the hybridizing sister species, Chrysochus auratus and C. cobaltinus (Coleoptera: Chrysomelidae), and compare the strength of those barriers to the strength of other known reproductive barriers in this system. Overall, cryptic barriers in this system are weaker than other barriers, indicating that they have not evolved rapidly. Furthermore, their evolution has been asymmetric. Non-competitive barriers substantially reduce the production of hybrid offspring by C. auratus females but not by C. cobaltinus females. In multiply-mated C. cobaltinus females, heterospecific sperm outcompete conspecific sperm, as evidenced by the fact that heterospecific males sired disproportionately more offspring than predicted from the results for singly-mated females. In C. auratus females, neither sperm type has a competitive advantage. Such asymmetries explain why nearly all F1 hybrids in the field are from crosses between C. cobaltinus females and C. auratus males. We discuss these findings in terms of understanding the cost of mating 'mistakes' in the Chrysochus hybrid zone. In addition, our discovery that 95% confidence intervals for commonly-used isolation statistics can be very wide has important implications for speciation research. Specifically, to avoid biases in the interpretation of such isolation metrics, we suggest that studies should routinely include error estimates in their analyses of reproductive isolation.