The optimization of metallurgical processes requires knowledge of the heterogeneity carried by constituents of interest. The common scenario, almost an industry standard, is to rely on the engineering firm who in turn relies on favorite manufacturers of sampling equipment to provide the appropriate sampling systems in the plant to perform accurate metallurgical accounting. This sequence is flawed from the beginning to the end because everything should be based on a careful understanding of the Theory of Sampling (TOS) by the company project team and their expert who invariably would suggest various approaches and different sampling systems.The necessary progressive and iterative approach should be that the company's project team (mining or other industries) would do its home work to have a good understanding of the geometallurgy, mineralogy and texture of the ore in each geological unit so consultants and engineers would have the best information to start with. Then a reputable sampling expert would recommend which sampling systems are needed, so the company would require engineering firms to follow pre-established, non-negotiable requirements, and finally select the sampling equipment existing on the market today that is most likely to lead to accurate and precise enough metallurgical accounting.It is the author's experience that when this necessary progressive and iterative approach is bypassed, sampling systems installed in the plant are likely to transgress the most elementary rules of sampling correctness. Invariably the argument is that the company cannot afford correct sampling systems for metallurgical accounting as suggested in the Theory of Sampling. There is nothing that is farther from the truth when later problems will develop and inevitable reconciliation difficulties with miners and geologists become daily routine with impossible answers; as a last resource the sampling expert may be called to perform a post-mortem analysis.The progressive and iterative approach suggested in this paper should be a due diligence requirement for any bank providing a loan for an important new mining project; anything less is most likely to generate unpleasant financial losses; historical, well-known examples are referred to.CITATION:Pitard, F F, 2015. A progressive, iterative approach leading to reliable metallurgical accounting sampling systems, in Proceedings MetPlant 2015, pp 78-91 (The Australasian Institute of Mining and Metallurgy: Melbourne).