Haemodynamic monitoring is a vital part of daily practice in anaesthesia and intensive care. Although there is evidence to suggest that goal-directed therapy may improve outcomes in the perioperative period, which haemodynamic targets we should aim at to optimise patient outcomes remain elusive and controversial. This review highlights the pitfalls in commonly used haemodynamic targets, including arterial blood pressure, central venous pressure, cardiac output, central venous oxygen saturation and dynamic haemodynamic indices. Evidence suggests that autoregulation in regional organ circulation may change either due to chronic hypertension or different disease processes such as traumatic brain injury, cerebrovascular ischaemia or haemorrhage; this will influence the preferred blood pressure target. Central venous pressure can be influenced by multiple pathophysiological factors and, unless central venous pressure is very low, it is rarely useful as a predictor for fluid responsiveness. Central venous oxygen saturation can be easily increased by a high arterial oxygen tension, making it useless as a surrogate marker of good cardiac output or systemic oxygen delivery in the presence of hyperoxaemia. Many dynamic haemodynamic indices have been reported to predict fluid responsiveness, but they all have their own limitations. There is also insufficient evidence to support that giving fluid until the patient is no longer fluid responsive can improve patient-centred outcomes. With the exception in the context of preventing contrast-induced nephropathy, large randomised controlled studies suggest that excessive fluid treatment may prolong duration of mechanical ventilation without preventing acute kidney injury in the critically ill.