For over 50 years patients with liver cirrhosis were considered to be at markedly increased risk of bleeding. This dogma was seemingly supported by abnormalities in standard laboratory tests (SLTs), such as the prothrombin time, that were interpreted as indicating a bleeding diathesis. However, publications from the last decade have revealed SLTs to be poor predictors of bleeding and it is now understood that stable patients with cirrhosis have a rebalanced haemostatic system and preserved thrombin generation. Viscoelastic tests (VETs), such as ROTEM® or TEG™ allow dynamic assessment of the entire coagulation process and provide a better illustration of the interactions between pro- and anticoagulants as well as platelets. Despite their documented success in reducing transfusion rates in liver transplantation more than 30 years ago, the adoption of VETs has been met with some resistance and has only recently gained significant momentum. Bleeding risk should be assessed in every patient undergoing invasive intervention and must consider markers of disease severity, underlying coagulation incompetence, anaemia and surgical factors. The recognition that bleeding in this patient cohort is predominantly linked to mechanistic factors such as portal hypertension, rather than primary coagulopathy, has led to a paradigm shift in their perioperative management. Cognizant of their detrimental effect, the use of large volumes of fresh frozen plasma to correct derangements in SLTs has given way to more refined haemostatic management with specific factor concentrates guided by VETs, coupled with measures to minimize portal venous pressure and meticulous surgical hemostasis.