Compositional analyses of ancient glasses have the potential to unravel their provenance and identify the raw materials employed in their manufacture. During the last few years, the analysis of archaeological glasses has increasingly benefited from the introduction and further development of micro-destructive techniques with trace element capability, such as laser ablation-inductively coupled plasma-mass spectrometry. However, while a variety of analytical standards can be employed for quantitative analyses of modern glass, the lack of dedicated certified reference materials represents a major obstacle in the analysis of archaeological specimens. Heterogeneity and the uncertainty in the characterisation of commercially available multi-elemental standards can also negatively influence the calculation of the response factors required to obtain fully quantitative data. Finally, fractionation and matrix effects can strongly affect the measurement, resulting in the possibility of over- or under-estimating the concentration of the analytes. This paper describes a fully quantitative method used for the investigation and characterization of a large set of Roman and late Iron Age glasses used in the making of Iron Age British beads, the results of which are published elsewhere (Bertini et al., J. Archaeol. Sci., 2011, 38, 2750–2766). In this work, the impact of the aforementioned issues and the different methods of quantification used on the precision (determined as the bias between different replicates), accuracy and repeatability of the measurement are evaluated. Particular emphasis is placed on assessing the influence of laser sampling procedures, integration of the transient signal, data reduction strategies, quantification approaches and the possibility of adapting the method to suit a large range of glass compositions.