What to suppress before you send

The short answer

Eight checks, run at enrolment, against live lists rather than a spreadsheet somebody exported. Existing customers, open opportunities, active partners, opted out, recently contacted, hard bounced, competitors, and unservable regions. All of them before enrichment, because enrichment costs money and suppression is free.

Why this is the highest-value boring job in outbound

Suppression failures do not look like failures. Nothing breaks, no alert fires. What happens is that an existing customer gets a cold pitch for the product they already pay for, an open opportunity gets a generic first-touch email while sales is mid-negotiation, and somebody who opted out eight months ago receives another message.

Each of those is a small event and each is expensive. The customer one damages the relationship. The opportunity one damages the deal and, more lastingly, sales stops trusting marketing to touch their accounts. The opt-out one is a compliance issue.

The eight checks, in order

1. Existing customers. Checked at company level as well as contact level. A new person at a company you already serve is not a cold prospect, they are an expansion conversation and they belong to someone else.
2. Open opportunities. Anything with a live deal attached, at company level. This is the one that causes the most internal damage when it is missing.
3. Active partners and suppliers. Often missed entirely because they sit in a different object or a different system.
4. Opted out. Global, permanent, and checked across every sending tool rather than per tool. An opt-out in one platform must suppress in all of them.
5. Recently contacted. A recency window, usually 30 to 60 days, checked across all sequences at once. Per-sequence recency rules do nothing, because the problem is two sequences, not one.
6. Hard bounces. Suppressed globally on first hard bounce. Continuing to send to a dead address is the fastest way to damage a sending reputation.
7. Competitors. A short maintained list. Cheap to build, mildly embarrassing to skip.
8. Unservable regions. Anywhere you cannot legally market to, cannot support in the local language, or cannot service. There is no point paying to enrich a record you would have to decline.

Live lists, not exports

The failure mode is not forgetting suppression. It is doing it once, by hand, from an export. That works for the first campaign and decays immediately, because the export is a photograph of a moving thing.

Suppression has to be a set of live lists that every enrolment checks automatically. If a person has to remember to apply a filter, the system fails the first week that person is on holiday, and it fails silently.

How to check it is actually working

Two tests, both quick.

• The known-record test. Take one existing customer, one open opportunity and one opted-out contact. Try to enrol all three. All three should be blocked, and the system should say which rule blocked them.
• The overlap report. Before any large send, run the target list against every suppression list and look at the size of the overlap. If the overlap is zero, your suppression is not connected. It is never zero on a real list.

What suppression is not

It is not a substitute for targeting. Suppression removes people you must not contact. It does not make the remaining list good. A well-suppressed list of the wrong people is still the wrong people, which is why the fit definition comes before any of this.

Common questions

What should be suppressed from an outbound list?

Existing customers, open opportunities, active partners, anyone who opted out, anyone contacted recently by another sequence, hard bounces, competitors, and contacts in regions you cannot legally or practically serve.

Should suppression run before or after enrichment?

Before. Enrichment is priced per record, so every record you suppress first is money you do not spend. On a typical first build, between a fifth and a third of an imported list is removable before a single enrichment call.

How do you stop two sequences contacting the same person?

A recency rule checked at enrolment: no contact from any sequence within a set window, usually 30 to 60 days. It has to be checked across all sequences at once, not per sequence, or it does nothing.

Does suppression need to be rebuilt for every campaign?

No. Suppression should be a set of live lists that every sequence checks at enrolment, not a filter someone applies by hand each time. A manual step here fails the first week somebody is on holiday.

Where this connects