This meeting was designed to fill a key gap in available spatial databases by bringing a group of people together to brainstorm on the difficult problems involved, and identify a way forward, in developing a dynamic global agricultural production systems classification that can be mapped, ground-truthed and refined through time. The work built on considerable efforts that have been made in the past, and will be based on “case study” systems classifications developed for a diverse range of countries, from which general lessons can be learnt for applying to the global scale. The outputs from this meeting will find immediate application in the work of FAO and the CGIAR centres, in targeting technology and policy interventions that are effective in promoting sustainable livelihoods of the poor in developing countries. The meeting was funded by FAO’s Pro-Poor Livestock Policy Initiative.