Drug development is a risky business. Success or failure often depends on selecting one or two molecules for development from many choices offered by the engines of high-throughput discovery. A lead candidate needs to possess adequate bioactivity, appropriate physical-chemical properties to enable formulation development, the ability to cross crucial membranes, reasonable metabolic stability and appropriate safety and efficacy in humans. Predicting how a drug will behave in humans before clinical testing requires a battery of sophisticated in vitro tests that complement traditional in vivo animal safety assessments. This review discusses how to strategically identify which non-clinical studies should be performed to provide the required guidance and comfort to stakeholders involved in clinical drug testing.