The characterization of production systems is important to determine limitations and potentialities that allow their management. This type of diagnosis has not been carried out in the study region; therefore, there is no information on cocoa-based agroforestry systems. Therefore, traditional cocoa farms in the municipality of Los Andes, department of Nariño (Colombia), were social and economically characterized. Basic information was reviewed and a semi-structured survey was applied to a stratified random sample of 60 cocoa farmers. Eighteen qualitative and 20 quantitative variables were analyzed simultaneously with multiple correspondence analysis (MCA) and principal components analysis (PCA). In the PCA, five components explained 58.7% of the variability. The variables with the highest contribution were farm area, income from cocoa yields, income from agricultural production other than cocoa, and family labor. In the MCA, 11 components explained 58.43% of the variability. The most important components were farmers' schooling, transport type, loans, marketing, technical assistance, gender, land tenure, production systems, roads, and domestic animals. Finally, topographic and climatic conditions, inadequate roads and marketing of cocoa and the low adoption of technologies limit the competitiveness of the cocoa farms, hence the need to create favorable conditions to enhance the potential of cocoa agroforests.