Portable X-ray fluorescence (pXRF) spectrometry is widely employed in the geoarchaeological analysis of archaeological deposits. However, the preparation methods and laboratory protocols used remain diverse, with established standards frequently disregarded. In this study, we examine the effects of grinding on the precision of mean analyte concentrations and evaluate the suitability of different factory calibrations and their associated measurement modes for geoarchaeological applications. Using a set of samples from the floor of a prehistoric smithy rich in microrefuse, we explore site-specific calibration standards, the relationship between the number of measurements and precision, and assess how precision affects the interpretation of spatial datasets. Based on our findings, we offer a series of recommendations for geoarchaeologists analysing large datasets with pXRF spectrometry.