Elemental analysis of ancient ceramics is primarily used in provenance research, where defined compositional groups are attributed to particular raw materials sources or production locations. Requirements in data quality and analytical performance are high, as is the need for clear and reproducible methodologies, and the availability of information on the above to ensure interlaboratory comparability and long-term data validity. This paper outlines the measurement parameters of a dedicated calibration set-up for the analysis of ancient ceramics using wavelength-dispersive X-ray fluorescence spectrometry (WDXRF). The specimens are prepared as concentrated glass beads, allowing the measurement of 26 elements from a single sample, thus minimising sample size requirements. Certified and non-commercial standards are used to evaluate the performance of the method in terms of detection limits, precision, repeatability, and accuracy. The materials used cover a range of compositions in line with the matrix variability encountered in archaeological ceramics. The data confirm the high standard of the method and highlight specific limitations. An initial assessment of comparability with other set-ups used in ceramic analyses, primarily Neutron Activation Analysis, is given through a discussion of performance on commonly analysed materials. The advantages of the proposed method include excellent analytical performance, analysis of a large suite of elements including all major, minor and a good range of traces, relatively small sample sizes, and preparation of samples that can be stored and re-analysed.