When you pseudonymize data, you substitute one or more values with non-identifiable keys. However, contrary to anonymization you store the mapping between keys and original values away from the data themselves. Pseudonymization makes it possible to reverse the process, if you hold both the mapping table and the pseudonymized data set.