Monadic concept testing is an essential tool in phase gate processes to weed out risky product launches. As an optimization tool, however, it suffers from measurement error, poor discrimination, and limited ability to test multiple variations. By comparison, conjoint methods can test many variations but make assumptions that pose challenges (e.g. no interaction between concept elements). A newer technology, real time evolutionary optimization, can test large concept spaces with high discrimination while accounting for interactions. A case study. demonstrates how the technology works.