Analyses of yield gaps i.e. the difference between observed and attainable crop yields in a given location, have raised expectations of significant potential progress in crop productivity in sub-Saharan African countries. However, an important question remains unanswered: Are those biophysically-determined attainable yields possible given the socio-economic context of farming systems in sub-Saharan Africa? In this study, we explored the potential increase in efficiency of crop production given the diversity of farming systems and livelihood strategies for the case study of smallholder farmers in Malawi. We implemented a non-parametric frontier efficiency method, Data Envelopment Analysis (DEA), which allows for the assessment of technical efficiency with respect to a production frontier. The frontier efficiency is based on best-performing farms in terms of input minimization and output maximization. Based on survey data of 102 households, we first built a typology of farming systems and distinguished two types, i.e. “maize-based smallholders under land pressure” (type 1) and “diversified crop-livestock producers” (type 2). By comparing results from farming system type-specific frontiers with those from an enveloping meta-frontier, we showed that the efficiency yield gap was overestimated by 13% in the case of the meta-frontier approach. Moreover, based on observed farming system-specific livelihood strategies, we defined different directions for reaching the efficiency frontier. For type 1 farming system, we assumed efficiency increase through reduction of both labor and inputs. For type 2 farming system, as income was mainly derived from agricultural activities, we assumed that efficiency increase could be achieved through increase in outputs, i.e. total calorie production from all cultivated crops. We quantified efficiency scores and identified their determinants to provide more specific recommendations on the levers for action to increase efficiency of crop produ