The benefits of good glycaemic control in both type I and type II diabetes mellitus are undoubtedly proven. Most national bodies have recommended glycaemic targets, with an HbA(1c) to achieve between 6.5 and 7.5%. However, it is well known that even in clinical trials, and routinely in clinical practice, the majority of patients fail to achieve optimal glycaemic control. The reasons for this failure are complex and multifactorial. Healthcare providers often delay the initiation and intensification of insulin unnecessarily. This stems from a fear of causing hypoglycaemia or weight gain in patients, from doubts about patients' self-care abilities and/or from inadequate resources to provide the necessary structured education to support patient self-management. Patients may be poorly adherent to treatment advice-particularly behavioural aspects such as self-monitoring, diet and exercise-although this may itself derive from inadequate access to effective diabetes education. There is, however, a limit to what can be achieved with existing exogenous insulin therapies due to their imperfect pharmacokinetic and pharmacodynamic profiles. Prominent among these imperfections is the problem of variability of effect from injection to injection with basal insulin formulations. Improvements in this area should benefit control and tolerability.