We discuss two methods that can be used to estimate the impact of internal and external variability on nonlinear systems, and demonstrate their utility by comparing two experimentally implemented oscillatory genetic networks with different designs. The methods allow for rapid estimations of intrinsic and extrinsic noise and should prove useful in the analysis of natural genetic networks and when constructing synthetic gene regulatory systems.