Lack of knowledge about demand responses or about behavioural aspects of decision-making within procurement processes is a significant cost driver in modern supply chains. Very often, this lack of knowledge leads to a substantial increase in inventories and may endanger negotiated service levels. For instance, various studies reveal that decision-makers tend to anchor orders close to the average past demand although the target order size is significantly higher or lower. In order to improve this situation, feedback has to be systematically provided to the decision-makers. In combination with modern big data analytics and reporting instruments that enable exhaustive monitoring, effective indicators have to be applied in order to directly detect processes with significant potential for improvement. Hence, this paper proposes a new approach for measuring the intricacy in purchase order sizing that addresses self-awareness skills of decision-makers. By simultaneously analysing the amount and structure of occurring costs, processes with a significant and simple structured error pattern are identified. In order to identify these processes more reliably, a new approach that supplements former information-theoretic entropy measures by an additional cost value is proposed. By analysing costs and the structure of deviations from target values in a two-dimensional measure, a more comprehensive understanding of the considered order sizing process is pursued. In order to illustrate the application of the new approach and show limitations of one-dimensional measures, different scenarios that exemplify the new approach are presented.