Analysis of large scenario databases has become a critical tool for identifying climate mitigation strategies, as shown in the latest IPCC report. However, key elements of these strategies differ significantly among scenarios. Possible reasons include differences in climate target, models used, and assumptions on behavioral, technological and socio-economic developments. For policymaking, it is important to know which of these factors are the main cause of the spread, but quantification of this is still missing. Here, we aim to identify consensus in climate policy scenarios by analyzing how much of the variance in scenario outcomes can be explained by these factors, using Sobol decomposition. Some results, eg concerning future use of fossil and renewable resources, are mainly determined by climate target, while others, such as the composition of different renewables in the electricity mix and key outcomes of end-use sectors are more model dependent. Scenario aspects beyond model and climate outcome determine only a limited part of the variation, eg, in nuclear power and hydrogen use, which suggests the need of more cross-model scenario variation. The outcomes put mitigation strategies in a new perspective by identifying which findings on the energy transition are robust and reveal key areas for future scenario development, model improvement and research.