AIM: Based on the results of 19 food microbiology proficiency testing (PT) schemes, this study aimed to assess the laboratory performances, to highlight the main sources of unsatisfactory analytical results and to suggest areas of improvement. METHODS AND RESULTS: The 2009-2015 results of REQUASUD and IPH PT, involving a total of 48 laboratories, were analysed. On average, the laboratories failed to detect or enumerate foodborne pathogens in 3.0% of the tests. Thanks to a close collaboration with the PT participants, the causes of outliers could be identified in 74% of the cases. The main causes of erroneous PT results were either pre-analytical (handling of the samples, timing of analysis), analytical (unsuitable methods, confusion of samples, errors in colony counting or confirmation) or postanalytical mistakes (calculation and encoding of results). CONCLUSIONS: PT schemes are a privileged observation post to highlight analytical problems, which would otherwise remain unnoticed. In this perspective, this comprehensive study of PT results provides insight into the sources of systematic errors encountered during the analyses. SIGNIFICANCE AND IMPACT OF THE STUDY: This study draws the attention of the laboratories to the main causes of analytical errors and suggests practical solutions to avoid them, in an educational purpose. The observations support the hypothesis that regular participation to PT, when followed by feed-back and appropriate corrective actions, can play a key role in quality improvement and provide more confidence in the laboratory testing results