Traditionally, researchers and practitioners have been focused on deterministic costing models without acknowledge and handle conveniently cost uncertainty. This paper presents and discusses a methodology for measuring the risk within costing systems: Costing at Risk (CaR) which, considering a predefined confidence level, takes into account the worst expected outcome in terms of cost in a certain period. For the determination of CaR, a six steps methodology is proposed. CaR permits to identify the components which most influence the variability of cost objects allowing the design of more effective strategies of risk mitigation. CaR has been used to analyze the risk of the production cost of the Colombian Coffee which is produced across the country in different conditions. Indeed, the patterns of variability of the three components of production cost vary significantly when the main "cafeteras" regions are compared.