There are several approaches to considering classifications on user data. One of the most prevalent classifications is PII, which refers to personal information fields (often columns in tables/views). This enables Atlan to disguise data, i.e. Atlan can hash/nullify/redact PII-tagged column data.
Use RTBF (Right To Be Forgotten) as a classification if you need to be GDPR-compliant.
If there are schemas/tables with data on users/customers that need to be cleaned up on a regular basis, this classification might serve as a label for them as well as power any cleanup tasks you would want to do at source.
You can also consider CIA classifications.