Over the past twenty years, a major development in firms’ innovation strategies has been the emergence of crowdsourcing as a tool to stimulate new ideas. A growing literature has examined the process by which ideas are winnowed from such discussions for further develop¬ment. This talk focuses upon a surprisingly neglected factor: the substantive content of em¬ploy¬ee interventions. We demonstrate the utility of topic modeling as an approach to extracting thematic information about discussion content and serving as a basis for identifying significant collective concerns and the degree of consensus around them, the alignment of posts with both elite and consensus emphases, and evidence for deliberative learning during discussion threads. We use this approach to analyze an influential example of such discussions and demonstrate that posts selected to exemplify promising ideas for innovation emphasized particular themes, were aligned with cues from organizational elites as well as reflecting broader topical emphases, and were both generative of and influenced by relatively long ex¬changes. Finally, we consider feat¬ures of the discussion design and selection process that may have accounted for the results we report.