Abstract Data representing the glycaemic potency of foods and carbohydrates are being increasingly included in food composition databases for managing the worldwide problems of diabetes and obesity. This paper identifies a number of criteria for assessing the quality of food data intended to guide healthier food choices for postprandial glycaemic control. The criteria are then applied to various carbohydrate and glycaemic potency data that are now being added to food composition databases. Our analysis suggests that if communication is based on the glycaemic potency of whole foods, as consumed, it is more likely to be useful for consumers, nutritionists and food producers than if based on food carbohydrates alone. Basing glycaemic impact values on foods rather than on carbohydrates would allow them to be expressed per serving and per 100 g of food, consistent with other information presented in food labels, but would still allow carbohydrate-based food selections from food groupings of similar composition, as is required when using the glycaemic index. Furthermore, we recommend in vitro measurement of glycaemic potency, to overcome the expense, difficulty and imprecision of in vivo analysis, which makes it unsuitable in product development, quality assurance, and accumulation of food composition database values.