List Hygiene & Sunset

Every argument for cleaning a list usually starts with deliverability. Mine starts with the invoice. Your ESP bills on the size of your profile base, which means every dormant subscriber, every address that has been bouncing since spring, every mailbox sitting permanently full is a recurring line item. Holding bad emails is not neutral. You are paying rent on addresses that cannot buy anything. When a stakeholder measured on list growth pushes back on suppression, that is the argument that lands: this is not shrinkage, it is cost cutting, and the deliverability protection comes free with it.

The deliverability case is still the structural one. Every major provider uses engagement as a reputation input. Google says plainly that placement follows whether people want your mail, and holds bulk senders to a spam-rate target of 0.10% with a hard ceiling at 0.30%. Yahoo measures engagement server-side, from its own data, beyond the reach of any tracking pixel. Microsoft scores how bulk-like your mail is and routes accordingly. Sending to people who stopped caring does not just waste sends; it teaches the algorithms that nobody wants you, and the algorithms grade on a curve you share with every sender in your category.

Then there is the part of the list that is not merely indifferent but radioactive. Abandoned mailboxes follow a documented lifecycle: providers hard-bounce them for a year or more, then quietly reactivate some as spam traps. An address that bounced all last year and takes mail today is not a recovered customer. It is evidence collection. A sender who ignored twelve months of bounces, or who lost bounce history in an ESP migration, mails straight into that trap. This is why bounce processing is non-negotiable, why aged lists are never worth reviving, and why the hard line in this playbook is arithmetic rather than caution: a list dormant and unmanaged for twelve months or more gets suppressed, not re-permissioned.

Verification services have a place in this system, and it is narrower than their marketing. They catch typos, dead domains, missing MX records, role accounts, and disposables, which makes them a solid acquisition-time defense at the form. They structurally cannot catch pristine traps or recycled traps in their active phase, because both answer like real mailboxes. Verification is a front door check, never a retroactive cure for a list you should not be mailing.

I run this playbook at national retail scale. The bounce rules below match how I classify in practice, including the current reality at the big consumer mailbox providers where storage cuts have left dormant accounts permanently full: I treat those full-mailbox bounces as hard bounces after one or two hits, because a mailbox nobody empties is a mailbox nobody reads. Address verification runs as part of standing hygiene, not as a one-time rescue. And the cost-per-profile argument above is the one I have actually used to sell suppression to growth-measured stakeholders, because it reframes the conversation from list size to unit economics.

The sunset itself is a tiered wind-down, not a binary switch, and the tiers are how you resolve the real fight inside every marketing team: deliverability wants to cut unengaged subscribers early to protect the complaint ratio, revenue points out that late engagers surface for weeks and wants to evaluate longer. Both are right.

Reduce early, suppress late, and never keep blasting the middle at full frequency.