How to Test That Deletion and Consent Withdrawal Actually Work

An opt-out often sets a flag in one system. The marketing tool connected last year never hears about it. Messages keep going out.

Deletion has the same weakness. The record leaves the main screen but survives in an export, a reporting copy or a shared sheet. Nobody notices until someone complains.

Create a test person with a phone number and email address you control. Give consent through the normal form. Confirm the record shows purpose, date and source.

Then withdraw consent the way a customer would. Rule 3 of the Rules, as published, says withdrawal should be comparably easy to giving consent. Your lawyer can say how that applies to you.

Send an erasure request for the test person through the channel your notice names. Check who received it, how identity was confirmed and what was deleted.

Some records may need to stay. The Rules as published require certain processing logs to be kept for at least one year. Tax and sector laws may require invoices to be kept. Your lawyer sets those periods.

Record the outcome in plain words. Note what was erased, what was kept and why.

A retention job that never runs looks the same as one that works. Look for its report. It should list what was removed and when.

Backups are harder. Removing one person from an encrypted backup is rarely practical. A common approach is to let old copies expire on a set schedule and keep a list of erasure requests to re-apply after any restore.

Write each failure down with an owner and a fix date. Then repeat the same test after the fix.

The PIB backgrounder says requests for access, correction or erasure must be addressed within ninety days. Testing well before a real request arrives leaves time to fix what breaks.