MCP

Campaign scheduling and pausing

Last verified: 2026-08-05

TL;DR

  • Watch two numbers daily — bounce rate and spam-complaint rate — and one as a proxy: a reply rate that falls while the copy stays the same usually means placement has slipped, not that the email got worse.
  • Published thresholds: pause and clean above roughly 2% bounces; reduce volume above roughly 0.05% complaints (2026 vendor guidance) [1]. The provider ceiling is 0.30% complaints, and that is a ceiling, not a target [2].
  • Recovery is measured in weeks. That asymmetry is the whole argument for pausing early.

Scheduling

A campaign's schedule is a promise about volume. Set the daily send so that every campaign drawing on a mailbox, added together, stays inside that mailbox's limit — Volume limits and ramp — and keep it steady, because a reputation built on 40 a day does not absorb a sudden 200-message Monday.

Two scheduling habits are worth building in. Launch new campaigns in a staggered way rather than starting three at once on the same senders, so that if something goes wrong you can tell which campaign caused it. And leave the follow-up gaps alone once a sequence is running: shortening them mid-flight compresses the sending load in a way that shows up as a volume spike, quite apart from what it does to the recipient.

The warning signs

Bounces. A spike above about 2% means the list, not the send. Pause, clean, and re-verify before resuming [1]. Bounce rate climbing slowly over a long campaign usually means list decay.

Complaints. Guidance is to reduce volume once complaints pass roughly 0.05%, well before the 0.30% at which mailbox providers start acting [1, 2]. Complaint rate is the most expensive number on this list: it moves reputation fastest and is the one most directly under the control of your targeting.

A quiet drop in replies. "A sudden drop in reply rate (with no change in messaging) usually means deliverability has degraded" (2026) [1]. Because most of a campaign's mail can land in spam without anything visibly failing, this proxy is often the first thing you notice. A related tell: out-of-office replies continuing while genuine replies stop — the automated responses come from mail that arrived, so their survival tells you the drop is engagement, not delivery.

Diagnosing which layer is at fault is Diagnosing underperformance's subject; this page is only about the decision to stop.

Pausing, and the cost of not

Recovery timelines from the same operating guidance: a minor dip takes one to two weeks at reduced volume; a moderate reputation drop three to six weeks with continuous warmup; a severe one six to twelve weeks, or starting over on a new domain (2026) [1]. Set against a pause that costs you a day of sends, the arithmetic is not close.

So pause on the first clear signal rather than waiting for confirmation. The instinct to let it run one more day to be sure is exactly backwards: the evidence you are waiting for is more damage. Resume at a lower volume than you stopped at, and climb back rather than restoring the previous setting in one step.

Stopping for the right reason

Not every pause is a deliverability pause. A campaign should also stop when the offer is being answered with the same objection over and over, when the list turns out to be mistargeted, or when someone asks you to stop — the last one being an obligation rather than a judgment call. See Stopping rules and suppression and Staying legal in a nutshell.

References

  1. Smartlead — Email deliverability guide (vendor guide, Jun 2026)
  2. Google — Email sender guidelines (support page, requirements in force since Feb 2024)