Abstract The determination of the past ancient magnetic field intensity is typically very difficult and error prone. Fresh volcanic glass has been suggested to be an ideal material for obtaining direction and strength of the magnetic field. Domain state bias is generally absent, thermal alteration is normally negligible and, as demonstrated in this study, cooling-rate differences can be corrected for. Major issues remaining include alteration and the origin of the remanence: that is, whether it is a true thermo-, a thermochemical or, for altered samples, a chemical remanence. Hydration, devitrification and perlitization lead to incorrect estimates of the palaeointensity, which are very difficult to recognize as the palaeointensity analysis does not easily expose these biasing effects and points towards ‘reliable’ results. Particular care on sampling and/or volatile measurements are necessary to overcome these drawbacks.