Establish an authorized baseline
Record the validated relationships in the in-scope original sample. Then compare the transformed version against that baseline in a controlled environment. Without a baseline, describe structural consistency rather than claiming a retention percentage.
Check more than identifier format
A stable replacement ID is useful only if it maps consistently across permitted sources. Inspect split identities, collisions, shared accounts and system-generated actors. Removing an attachment may also remove the evidence for a decision even when every ID survives.
Report what changed
List lost links, excluded categories and altered timestamps with their effect on proposed tasks. Do not include re-identification keys in the release package. Preserving relationships supports utility; it is not evidence that re-identification is impossible.
Practical guidance, not legal advice or a buyer specification. Examples are illustrative. Reviewed October 20, 2026.