Complex calibration models developed on a first near-infrared (NIR) spectrometer cannot be used with spectra collected on a second NIR instrument, because of the differences between the instrumental responses of both instruments. The same problem occurs when the instrumental response of a single NIR instrument is subject to variations over time or when physical modifications of the samples such as particle size occur. To keep those calibration models, it is essential to correct those differences. This can be achieved by standardisation procedures, which are based on two main steps, namely estimation of the differences by measuring standardisation samples on both instruments, and correction of those differences. The different approaches available for each step are presented, and particular attention is paid to practical considerations.