A customer replies to a dispatch email with "stop emailing me". The reply lands in a shared inbox rather than at an unsubscribe endpoint, and the agent reading it picks from a menu built for tickets. The guidance turns on a purpose that menu does not model: the record is kept so that future sends can be checked against it [1].
Counting the rights makes the tension easier to size. The ICO lists four things a person can do, which are object, opt out or unsubscribe, withdraw consent, and ask you to delete their information [7]. Three of those four land on the suppression list, because the regulator says complying with an objection does not automatically mean deletion and that suppression is preferable in most cases [6][12]. The fourth arrives as a deletion request, and the guidance treats it as a separate question [13].
Which makes suppression a match-key problem. A list you check against needs whatever identifier the campaign would otherwise have used, which is why the ICO's own identity example is confirming an email address or a phone number in order to stop using those details [10]. Lead scores and the inferred interests that profiling produced do no work in that check, and profiling sits inside the scope of the objection in any case, alongside passing details to third parties [3].
Intake is where most tooling is thin. There is no required form of words, an objection can be verbal, and it can arrive at any part of the organisation, which is the ICO's stated reason for expecting firms to have a process that recognises one [4]. The channels that implies are the phone queue and a rep's inbox. Objecting also has to be free, and can happen before you have ever marketed to the person [9][11], and the deadline for telling people the right exists is at the latest your first communication, set out separately and in plain language [8].
One more column earns its place on the suppressed record: a reason. You cannot go back later to ask an objector whether they have changed their mind, since that contact is itself direct marketing for the purpose they objected to [5]. A re-engagement flow reading a blank send history as dormancy will mail somebody who told a salesperson, out loud, to stop.
So the test I would run through the contact schema this week: for each column, say whether it matches a future send or evidences that you honoured the request. Columns that do neither are what an erasure workflow should be allowed to take, and columns that do either are what has to survive it. The forcing function that follows is cheap. A colleague objects by the least convenient route you support, verbally, to someone in sales, and you measure two things afterwards: how many days pass before a campaign send is actually blocked, and whether the suppression row is still standing after the next deletion run. If the answer to the second is no, you have built the delete button the guidance warns about, and the person who asked you to stop will hear from you again.
