Roche’s yearly advanced analytics data challenge attracts 500 internal data scientists. The scope of challenges tackled has ranged from traditional prognostic risk models, through to furthering our understanding of Cancer Immunotherapy response with predictive modelling. The benefits of this initiative have, in addition to building a stronger analytics community, allowed us to rapidly prototype strategies. Further, sonsensus models, where a diverse set of models built by different teams using unique modelling strategies and domain knowledge vote on predictions has been shown to outperform individual models, providing new insights into the performance of our molecules.