The Cicerone Project conducted a grazed farmlet experiment on the Northern Tablelands of New South Wales, Australia, from July 2000 to December 2006, to address questions raised by local graziers concerning how they might improve the profitability and sustainability of their grazing enterprises. This unreplicated experiment examined three management systems at a whole-farmlet scale. The control farmlet (farmlet B) represented typical management for the region, with flexible rotational grazing and moderate inputs.A second farmlet (farmlet A) also used flexible rotational grazing but had a higher level of pasture renovation and soil fertility, while the third farmlet (farmlet C) had the same moderate inputs as farmlet B but employed intensive rotational grazing. The present paper provides an integrated overview of the results collated from component papers and discusses the inferences that can be drawn from what was a complex, agroecosystem experiment. The measurements recorded both early and late in the experiment were tabulated for each of the farmlets and compared with each other as relative proportions, allowing visual presentation on a common, indexed scale. Because of equivalent starting conditions, there was little difference between farmlets early in the experimental period (2000–01) across a wide array of measured parameters, including herbage mass, potential pasture growth rate, liveweight, wool production per head, stocking rate, gross margin and equity. Although the experiment experienced drier-than-average conditions, marked differences emerged among farmlets over time, due to the effects of treatments. During the latter half of the experimental period (2003–06), farmlet A showed numerous positive and a few negative consequences of the higher rate of pasture renovation and increased soil fertility compared with the other two farmlets. While intensive rotational grazing resulted in superior control of gastrointestinal nematodes and slightly finer wool, this system had few effects on pastures and no positive effects on sheep liveweights, wool production or stocking rate. Whereas farmlet A showed higher gross margins, it had a negative and lower short-term cash position than did farmlets B and C, due largely to the artificially high rate of pasture renovation undertaken on this farmlet during the experiment. Although farmlet B had the highest cash position at the end of the experiment, this came at a cost of the declining quality of its pastures. Modelling of the farmlet systems allowed the results to be considered over the longer timeframes needed to assess sustainability. Thus, returns on investment were compared over realistic amortisation periods and produced outcomes based on long-term climatic expectations which were compared with those that arose under the drier-than-average conditions experienced during the experimental period. The main factors responsible for lifting the productivity of farmlet A were the sowing of temperate species and increased soil fertility, which enhanced the amount of legume and increased pasture quality and potential pasture growth. The factor that affected farmlet C most was the low proportion of the farmlet grazed at any one time, with high stock density imposed during grazing, which decreased feed intake quality. The paper concludes that more profitable and sustainable outcomes are most likely to arise from grazing enterprises that are proactively managed towards optimal outcomes by maintaining sufficient desirable perennial grasses with adequate legume content, enhancing soil fertility and employing flexible rotational grazing.