Review your suppression list
Suppression is a list of people who must not be contacted, and every entry has an origin: an unsubscribe, a complaint, a bounce, a STOP on SMS, an opt-out on WhatsApp, or a do-not-contact decision. Reviewing it means checking the origins are right — not shortening the list. Most entries should stay, and one kind should effectively never be cleared.
What is on the list, and where it came from#
| Origin | Meaning | Clearable |
|---|---|---|
| Unsubscribe | The recipient asked to stop receiving outreach | Only on a fresh, recorded request from them |
| Complaint | The recipient marked a message as unwanted | No, in practice — this one costs you deliverability |
| Bounce | The address does not accept mail | Yes, if the address is corrected or verified again |
| STOP on SMS | The standard opt-out keyword | Only on the recipient's own re-opt-in |
| WhatsApp opt-out | Honoured and recorded like any suppression | Only on their own re-opt-in |
| Do-not-contact | A deliberate decision that this person is not to be contacted | Not casually, and never by the Assistant |
Compliance is checked in one place before any outreach, so an entry here takes effect everywhere rather than on whichever channel recorded it. That single point is also why the list is worth reading occasionally: everything downstream depends on it being right.
Suppression is not the same as blocking#
- Suppression
- A compliance state about outreach: this person must not be contacted. Checked before anything goes out.
- A block
- A memory row against the contact, tagged for a channel. Directives are read from the tag list rather than from the text, which is what makes a block hold across channels and into future conversations.
- The practical difference
- Suppression stops you contacting them. A block changes how Connect treats them, including when they contact you.
The review#
Group the list by origin rather than reading it alphabetically.
Result Bounces are the only large group with a legitimate reason to shrink. Everything else you are checking rather than pruning.
Work through the bounces: correct the address where you have a better one, and leave the rest.
Result A corrected address is a new identity on the same person, so the history stays intact rather than starting again.
Sample the do-not-contact entries and check each has a reason recorded.
Result A do-not-contact entry with no reason is the one that gets argued about in six months. Add the reason now while somebody remembers it.
Check that nobody on the list is also an active customer relationship.
Result That combination is usually a mistake in one direction or the other, and it is much cheaper to resolve before somebody notices the silence.
Leave unsubscribes and complaints alone.
Result Clearing those is how a sending reputation is damaged, and the damage is slow to appear and slow to undo.
Proving the list is doing its job#
- Attempt an outreach to a suppressed address from a draft. It should be refused immediately, with the suppression and its origin shown.
- Check the Decision Log for that refusal — a refusal is recorded as a decision, so the trail exists whether or not anybody was watching.
- Confirm the refusal happened before anything reached a provider. A suppression that only stops delivery is not a suppression.
- Check that a reply from a suppressed person still arrives normally. Suppression governs outreach; it does not make somebody invisible when they contact you.
That last distinction catches people out. Somebody who unsubscribed from outreach and then writes in with a question is a customer conversation, not a compliance problem — outreach and a customer reply are two paths with different gates and are never merged.
Questions#
Somebody asked to be added back. What do I do?
Record the request and clear the entry it applies to, keeping the record of both. For an unsubscribe or a STOP, their own re-opt-in is the thing that makes it legitimate — your recollection of a conversation is not.
Does suppression apply across channels?
The compliance check runs in one place before any outreach, so an entry is enforced regardless of which channel would have carried the message. Origins are channel-specific; the effect is not.
Will prospecting contact somebody who is suppressed?
No. Discovery runs through a compliance check before any first outreach, and identity resolution separately stops an existing customer being prospected. Both run before a message is written rather than after.