From a pathophysiological point of view, type 2 diabetes is a well-characterised disease, since the glycaemic disorders result from three main mechanisms (the De Fronzo's triumvirate): a defect of beta-cell function, decreased disposal of glucose in peripheral tissues and overproduction of glucose by the liver. Each defect is subject to 24-h circadian variations and to inevitable worsening with time. As a consequence, therapeutic strategies should reflect whether patients retain sufficient insulin secretion or suffer from a more severe secretory defect that progresses from being responsive to oral diabetic agents to the insulin-requiring stage. Identifying the different pathophysiological stages is a prerequisite for successful therapeutic strategies. This assessment can be done by considering on the one hand the HbA(1c) and on the other the glycaemic profiles. For the latter, either discontinuous (self-monitoring of blood glucose) or continuous glucose monitoring can be used. However, many difficulties remain for bridging the divide between well-understood pathophysiological concepts and suboptimal glycaemic control achieved in clinical practice. The main drawback is the difficulty in providing therapies at recommended doses to stochastic phenomena such as either intestinal absorption of carbohydrates or fluctuations in both pharmacokinetics and pharmacodynamics of hypoglycaemic agents.