Pseudonymisation

The technique lowers the chance that someone handling the data can tell who it is about. Analysts see only the codes, while the key sits with a small team. Because the key lets the process be reversed, the data is still personal data and privacy law still applies. That separates it from anonymisation, which cannot be undone. Methods include replacing names with random tokens, hashing identifiers with a secret value, and encrypting fields. Weak methods can be broken, for instance when a code is simply a customer number or when other fields give the person away. The GDPR names it as a useful safeguard, and it supports security, research and testing. Analysts can still link records over time through the codes, with far less exposure of identities.

A bank's analytics team studies spending patterns. Before the data reaches them, the customer name and account number are replaced by random tokens. The table linking tokens to customers stays with a separate security team. When a fraud pattern is found, that team can reverse the token for the few accounts that need contact.