The flotation process characterization is typically based on the mineral properties and related to the feed particle size. Laboratory testing allows for the evaluation of the batch flotation kinetics, while plant surveys are carried out for the plant evaluation, and sometimes the rougher flotation stage is also characterized by kinetics, considering either the full sampling of the circuit or the short-cut method. Comparisons of plant and batch results are useful for identifying the scale-up factors. The kinetic evaluation of cleaner stages is less common in plant surveys, and usually, only the overall cleaner and scavenger results are reported. This condition limits a more comprehensive understanding of these stages, which have significant differences from the rougher operation. In this study, the effect of main operating variables in cleaner and scavenger stages, such as finer particle size, higher mass recoveries, higher liberation, particles entrainment, froth recoveries, mineral grades, froth depth, gas rate, and others, was analysed by using an industrial simulation tool that was built from a wide industrial database. For this purpose, data from plant kinetic surveys was used to characterize the mineral feed entering the cleaner–scavenger stage, which allowed for calibrating the simulation tool and predicting the overall circuit performance. The metallurgical results of the cleaner–scavenger bank were compared with those when the bank was operating at a rougher stage (previous operation). The results allowed for evaluating the differences in metallurgical results of the cleaner–scavenger and rougher banks, mainly related to the differences in particle size and liberation as well as in the mass flowrate of collected particles, which affects bubble loading and consequently froth stability, that in turns impacts on froth recovery. The operating conditions and mineral characteristics of each stage also impacted the water recovery and gangue entrainment along the banks. The comparison of predicted recoveries and grades in rougher, cleaner, and scavenger stages showed a good agreement with plant data. These results validated the simulation tool, which is useful and flexible enough to characterize different stages, predict performance and explore new operating conditions. [ABSTRACT FROM AUTHOR]