Although sensitivity and uncertainty analysis and the quantification of error in model outputs have been regarded as an essential part of rigorous model design and application for some time, the developers of larger process-based agricultural and ecosystem models have been slow to adopt these approaches. Part of the reluctance might have been due to the relatively slow execution time of such models combined with the large number of model runs needed in these analyses. In recent years model execution time has improved through a combination of faster processors and parallel and cloud computing techniques. There have also been advances in the development of techniques used for these analyses that have increased their efficiency. Over those same years the demand for knowledge regarding the certainty of model estimates and how to obtain greater certainty has only increased as agricultural and ecosystem models are applied in increasingly contentious domains, such as informing policy. Given these changes we would argue that simulation modellers should use sensitivity and uncertainty analysis as a standard part of their applications ‘tool box’ for agricultural and ecosystem modelling.