For decades, non-selective beta-blockers (NSBBs) have been the standard of care for the primary and secondary prevention of bleeding from oesophageal varices. However, several questions regarding the best clinical use of NSBBs remain unanswered and new data continue to emerge. Herein, we aim to delineate the therapeutic window of NSBBs in cirrhosis from a more individualised perspective. We address the current controversy of widening the therapeutic window and prescribing NSBBs to all patients with clinically significant portal hypertension. Although transient elastography is useful to rule-in clinically significant portal hypertension, we lack robust data supporting the use of NSBBs in patients without varices. While most data are based on propranolol, accumulating evidence suggests that carvedilol is superior and should be the first-line treatment until the decompensated stage. The clinical risk-to-benefit ratio appears to deteriorate in advanced decompensated stages and the risk of harm is high in patients with refractory ascites, low blood pressure and renal impairment, which clinically define closure of the therapeutic window. We also critically review non-invasive surrogates and biomarkers for predicting the haemodynamic response to NSBBs and confirm that the absence of reliable non-invasive methods is one of the main challenges facing the field.