Low-cost and low-risk policy experimentation is an acknowledged virtue of simulation models of complex public issues (Ghaffarzadegan, Lyneis, & Richardson, 2011). But the virtue can become the vice. It is often too easy to ask, “What if this parameter value could be changed?” and get a quick quantitative answer. Model-based analysis that relies exclusively on parameter sensitivity testing may ignore how parameter changes in a computer model can be implemented by public organizations in the real world. Of course, testing a model’s sensitivity to variations in policy parameters is an important exploratory step in model-based policy analysis. Too often, however, there is no next step. A content analysis of three decades of articles published in the System Dynamics Review found that policy analysis has been limited to parameter sensitivity testing in nearly 75 percent of models of public issues (Wheat, 2010). This situation has developed despite admonitions from experienced modelers. Richardson and Pugh (1989) warned that “policies represented as parameter changes frequently tend not to be very effective in system dynamics analysis” (p. 332) and Sterman (2000) reminded modelers that “policy design is much more than changing the values of parameters” (p. 104).