Most antidiabetes drugs tested in clinical trials reduce HbA1c by 0·5–1·5%, depending on study design, baseline HbA1c, and study population, with combinations resulting in quite similar glycaemic outcomes but offering a range of non-glycaemic effects and attributes. Consistent with the recent position statement of the American Diabetes Association and the European Association for the Study of Diabetes,25 the entire clinical picture should be considered so that treatment can be personalised based not only HbA1c reduction reported in many trials, but also on tolerability, safety, and frequency and ease of administration.