Purpose: The resources needed to develop assessments of the twenty-first century skills, such as problem solving and collaboration, are huge, and require the introduction of cost-effective methods (Griffin and Care, 2015). The intention of the design template is to identify whether the presented approach to game development is viable for measuring collaborative problem solving (CPS) and if so, by using the template the same measures can be captured regardless of the specific task context. The paper aims to discuss these issues. Design/methodology/approach: This paper demonstrates a design template which utilises learning analytics and applies a measurement model to transform games into the twenty-first century assessments. The design template provides parameters for game design that provide sufficient data capture in which a measurement model can be applied. The learning analytics approach allows for capture of student behaviours in game play. By applying the measurement model, inferences can be made about demonstrated behaviours that are indicative of student ability. Findings: Following the task design template, it is evident that games can be designed taking into consideration the requirement for generalizability and measurement principles. In turn, tasks developed using a defined set of common characteristics and structure allow consistent measurement processes to be applied in an efficient and sufficient manner. In summary, this paper identifies a viable task design template in regards to design principles and scoring protocols for games generating measures of CPS. Originality/value: This approach combines various fields of research to present an approach that is a feasible, effective and efficient method for capturing data that are useful for understanding complex social and cognitive skills. The design template presents a method by which games can be designed in a way that assesses cognitive and social skills and provides a platform on which additional games can be readily created.