To remain competitive and manage their safety performance, many construction organisations have engaged in benchmarking themselves against lagging indicators provided by a statutory body. Aggregated metrics that are provided by statutory bodies are not useful for the purpose of operational benchmarking, as ‘best practice’ is unable to be identified. Access to safety statistics from leading construction organisations’ projects is seldom made available for the purposes of benchmarking. In addressing this void and to engender a process of operational benchmarking, a homogeneous dataset is used to examine 26,665 workplace injuries that arose during the delivery of 562 projects over a 10-year period by a leading international Australian construction organisation. The nature and the degree of severity of the injuries that arose are statistically analysed. The findings provide invaluable insights into issues contributing to workplace injuries during construction, which can be used as a basis for operational benchmarking and a platform for engaging in continuous improvement. Practitioner summary: Workplace injuries are a problem in construction. Recognising that safety is a key goal for construction organisations, we analyse the nature of workplace injuries that occurred in 562 projects. Acknowledging the challenges of using lagging indicators, an operational framework for engendering best practice in workplace safety is presented.