Analytical techniques based on gamma activation analysis (GAA) are rapidly gaining popularity in the mining industry for analysis of Au, Cu, and Ag. Since being introduced commercially by Chrysos Corporation in 2018 under the brand name PhotonAssay™ (CPA), 130 companies have reported Au exploration results to major stock markets using data collected with this technique, and 12 companies have used such results to support a publicly reported mineral resource estimate (MRE). While it has been established that, under best-practice operating conditions, results from CPA are accurate and precise, test results may not prove accuracy for every style of mineralisation and any individual CPA instrument. Therefore, many operators carry out their own test work, comparing CPA assays with results obtained by conventional techniques (eg fire assay). The authors’ experience with comparative studies around the world is that they are often not designed for rigorous scientific testing, and hence do not meet the objective of drawing robust conclusions on the quality of CPA data. Additionally, the issue of materiality, key to every mining and exploration project, is rarely discussed, and studies do not define a data quality objective (ie when is something fit for purpose). In this paper, we add detail to previously published recommended testing programmes, setting out the sample selection process, and detail all critical steps during sample preparation, including the collection of true duplicates for precision benchmarking. Based on several real-world examples, we then provide a step-by-step process to assess the validity of the testing data, accuracy and precision, and whether any observed differences are both statistically significant and material to the project.