Highlights:• The diversity of oil palm smallholders management practice and their performances remain poorly known.• We aim to understand which farming systems optimize the economic performance of oil palm.• A first typology on agricultural practices revealed a gradient of intensification, a second on economic performances revealed an optimum.• Crossing the two typologies showed that intensive practices are not economically optimal.• The study recommends to adapt local recommendations using multicriteria approaches.Abstract:CONTEXT: Oil palm is widely cultivated in Indonesia and it is produced by various populations of smallholders. Among those, ex-plasma smallholders are characterized by the fact that they only start to manage their own plantations 5 to 10 years after those were initially planted and managed by a partner industrial company. Despite the large areas that are cultivated by ex-plasma smallholders, their diversity of practices and economic performances are poorly quantified.OBJECTIVE: The present study aims at investigating which smallholder farming systems optimize profitability at the plot scale.METHOD: We defined two typologies, one based on agricultural practices and the other based on economic performance, using multivariate statistical analysis. The typologies were crossed to understand the links between agricultural practices and economic performance at the plot level.RESULTS AND CONCLUSIONS: Our results showed that despite initial plot homogeneity during planting phase, ex-plasma smallholders can develop very diversified management practices. A gradient of intensification was observed from farmers implementing no fertilization or weeding management up to farmers following agroindustrial-like management. Economic performances at the plot level was found to be variable, with gross margins ranging from 1.1 to 26.9 M IDR.ha−1.yr−1 (eq. 66.4 to 1625.5€). We showed that the most intensive practices led to highest yields although these were not linked to the optimal economic plot performances. A trade-off based on input costs was revealed, as high fertilization practices did not seem economically optimal for smallholders.SIGNIFICANCE: The present study highlights the need to combine quantified agronomic assessments with economic analysis, to properly design recommendations to farmers.