Conservation agriculture (CA) relies on three fundamental and inseparable pillars: no soil disturbance, diversified crop rotations, and permanent soil cover. Nevertheless, few studies have evaluated the interactive effect of these three fundamental pillars on a multicriteria basis. Here, we mobilize data from the French AGROSYST database, which gathers all farming practices and performances of the 3000 farms involved in the French DEPHY farmers' network. Linear mixed effect models were used to compare the performance (13 indicators) of CA (CAs, N=36) and pseudo-CA systems (pseudo-CAs, N=19, allowing one occasional superficial tillage) to conventional tillage (plowed, CTs, N=135) and superficial tillage (STs, N=90) based systems in similar production situations (climate, soil type, presence of livestock or irrigation, etc.). CAs required (compared to STs and CTs, respectively) more herbicides (+27 and +90%) but slightly less insecticides (-64 and-50%, non-significant), decreased time of traction/ha/year (-25 and-32%), fuel consumption (-21 and-39%), as well as mechanization costs (-20 and-26%), tended to slightly decrease profitability/ha (-7 and-19%, non-significant) due to slightly lower productivity (-19% and-25%) but resulted in better profitability per hour of field traction (+23% and +18%). Pseudo-CAs did not implement the three CA principles since crop rotation were as diverse and cover crop as frequent as in STs and CTs, and tillage occurred, albeit rarely. However, pseudo-CAs used less insecticides (-92 and-83% compared to STs and CTs, respectively), decreased fuel consumption (-25% compared to CTs), and resulted in similar productivity and economic profitability (per hectare and per hour of field traction). Further investigation is required to identify the diversity of responses across production situations, the determinants of multiperformance in given production situations, and to track down the rare innovative systems optimizing multiple performances and solving apparent trade-offs.