Multibeam sonar systems now routinely record seafloor backscatter data, which are processed into backscatter mosaics and angular responses, both of which can assist in identifying seafloor types and morphology. Those data products are obtained from the multibeam sonar raw data files through a sequence of data processing stages that follows a basic plan, but the implementation of which varies greatly between sonar systems and software. In this article, we provide a comprehensive review of this backscatter data processing chain, with a focus on the variability in the possible implementation of each processing stage. Our objective for undertaking this task is twofold: (1) to provide an overview of backscatter data processing for the consideration of the general user and (2) to provide suggestions to multibeam sonar manufacturers, software providers and the operators of these systems and software for eventually reducing the lack of control, uncertainty and variability associated with current data processing implementations and the resulting backscatter data products. One such suggestion is the adoption of a nomenclature for increasingly refined levels of processing, akin to the nomenclature adopted for satellite remote-sensing data deliverables.