We develop a methodology for incorporating nutrition impacts in economy-wide analyses, providing entry points for where, when and how to act. It accounts for three channels of consumption, directly via primary commodities and indirectly via processed foods and food-related services, and produces indicators showing content by nutrient (currently calories, proteins, fats and carbohydrates), channel, source region and sector. The paper applies the framework in a CGE model (MAGNET) and uses FAO data to project nutritional outcomes resulting from the global food system over time. The analysis confirms that developing regions catch up with developed regions, with the USA at the high-end of nutrient consumption, whilst Southern Africa lags behind. In the USA the processed food channel dominates, whereas in Southern Africa the direct channel dominates. In the USA, and similar regions, fat taxes (thin subsidies) on unhealthy (healthy) processed foods, technologies reducing bad ingredients (e.g. trans fats, salt), improved food labelling, information and marketing campaigns, and/or targeted cash transfers may be worthwhile to investigate. In Southern Africa, and regions alike, technological advances increasing nutrient availability via primary agriculture and/or cash transfers enabling access may be more pertinent. The relative fixedness of sectoral origins shows that consumption habits change slowly and are visible only in the long term. For certain regions, including Southern Africa and USA, nutrient import dependency increases with substantial variations in regional sourcing. This implies that concerted action across the globe is crucial to reach diet, nutrition and health goals, and should include upcoming Asian economies, Africa (excl. Southern Africa) and the Middle East. Heterogeneity of results necessitates future ex-ante quantitative policy analyses on a more detailed and context-specific basis.