Purpose – After more than 30 years of research, literature on performance measurement systems (PMS) is characterized by diversity and fragmentation. Due to the multidisciplinarity of stakeholders and researchers involved the basis of literature is expanding, but not converging. The purpose of this paper is to dispersedly discuss failures of PMS in the abundance of literature written. Design/methodology/approach – Over 250 articles related to PMS have been analyzed in order to shortlist failures of PMS. Two criteria have been used: explicitly referenced being a failure; or mentioned as being essential for a successful PMS. Next steps were clustering, cross-checking with academics and professionals and re-allocation to appropriate levels. Findings – This paper identifies 36 failures and proposes an easy taxonomy for further referencing by attribution to three levels: metric, framework and management. Failures range from uncertainties in data gathering, lack of knowledge and dealing with complexity, toward the allocation of necessary resources. Research limitations/implications – Limitations are attributed to the abundance of research published on PMS. Consequently, incorporated papers are a subset representing the current state of the research domain. Furthermore, the completeness of the list can be discussed as well as the level of generalization of the proposed taxonomy. Originality/value – Both academics and professionals can benefit from this study as it creates an awareness of the risks involved when constructing, implementing and managing a PMS. Therefore, this original research ought to be seen as a catalyst for a learning curve, as it puts the research of PMS in a different perspective.