In recent years, atom probe tomography (APT) has been increasingly used to study minerals, and in particular the mineral zircon. Zircon (ZrSiO4) is ideally suited for geochronology by utilising the U-Th-Pb isotope systems, and trace element compositions are also widely used to constrain petrogenetic processes. However, while standard geoanalytical techniques provide information at micrometer scale lengths, the unique combination of chemical/isotopic sensitivity and spatial resolution of APT allows compositional and textural measurements at the nanoscale. This interlaboratory study aims to define the reproducibility of APT data across research facilities and assess the role of different aspects of the atom probe workflow on reproducibility. This is essential to allow correct evaluation of APT results and full utilization of this emerging technique within the geoscience community. In this study, nine samples from the same homogeneous, GJ-1/87 zircon reference grain were sent to nine APT institutes in Germany, the UK, USA, Canada and Australia. After preparing the sample out of a selectioned slab, each institute conducted three different rounds of APT analyses: using (i) unconstrained analysis parameters, (ii) pre-defined analysis parameters, and (iii) interpreting and quantifying a provided dataset. Data such as the measured elemental composition, acquisition parameters, or mass spectrum peak identifications, were recorded and analyzed. We observe a significant variation in the measured composition across this interlaboratory study as well as the number of trace elements identified. These differences are thought to directly result from the user's choice of atom probe data analysis parameters. The type of instrument does not seem to be a critical factor. Consequently, comparison of absolute trace element concentrations on zircon using APT between laboratories is only valid if the same workflow has been ensured.