Concentrators treating sulphide ores present significant business opportunities by way of improved concentrate grade and paymetal recovery. In the case of the Raglan operations, it was necessary to demonstrate the metallurgical performance and flowsheet opportunities for a future ore blend which was expected to have a different processing response. The future ore was campaigned at the Raglan concentrator, and sampling of the flotation circuit was performed. In addition to mass and value balances derived from this sampling, quantitative mineral measurement by Qem*SCAN was also conducted using polished sections on a size by size basis, and which required advanced sampling and sample preparation protocols. In this paper, a modification of the statistical benchmark surveying method is presented to meet the above objectives. This modification resides in the use of drill-core data for the estimation of expected grade of ore milled in a week-long surveying campaign. Use of Krige's lognormal models on the drill-core data on a weighted mean basis, and allowing for measured waste dilution, provides an accurate estimate of this ore grade. It is shown that this adaptation provides a mean ore grade estimate that is within 4% error of the survey ore grade estimate. (C) 2006 Published by Elsevier Ltd.