A common and serious exposure.
WHAT THE PROBLEM IS
Production data copied into environments with weaker protection.
WHY IT IS SERIOUS
Those environments have more access, less monitoring, and copies proliferate.
WHAT LAWFUL BASIS COVERS IT
Rarely any, for using customer data to test software.
WHAT TO DO INSTEAD
Generate synthetic data, or anonymise properly.
WHAT PROPER ANONYMISATION REQUIRES
Replacing names Replacing contact details Replacing addresses Altering dates and identifiers where they identify Removing free-text fields, which contain anything
WHY FREE TEXT MATTERS
Notes and comments contain unpredictable personal data.
WHAT TO PRESERVE
Volume and distribution.
WHY
Development against small clean datasets hides every real problem.
WHAT TO NEUTRALISE ABSOLUTELY
Email addresses and phone numbers.
WHY
Test systems sending real messages to real customers is a frequent and damaging accident.
WHAT TO AUTOMATE
The anonymisation, as part of any copy.
WHY
A manual step is eventually skipped.
WHAT TO RESTRICT
Who may copy production data at all.
WHAT TO LOG
Every such copy.
WHAT TO NEVER DO
Point a development system at the production database.
WHAT TO AUDIT
Where copies of production data currently exist.
