Schedule and send

This page covers the last step of a campaign and everything after it: choosing a send time, what happens if you need to stop a send that has already started, and how to read the report without over-reading the open rate.

Send now, or schedule#

Send now starts the queue immediately. Use it when the content is time-sensitive and you have already tested it.

Schedule holds the campaign until the time you set. Use it for everything else. Scheduling for even fifteen minutes ahead gives you a window to catch the mistake you are going to notice thirty seconds after you commit.

Sending is not instant either way. A large campaign is delivered as a queue over a period rather than all at once, so the last recipient receives the message some time after the first. That is normal and it is better for your reputation than a sudden burst.

Time zones

A scheduled time is only meaningful alongside a time zone. Before you schedule, confirm which time zone the scheduler is interpreting your time in, and remember that it may not be the one you are physically sitting in.

Two things that catch people out:

• Your audience is not in one place. A campaign scheduled for 9am goes out at one moment in absolute time. For a list spread across several countries, that moment is breakfast for some readers and the middle of the night for others. If time of day matters to you, segment by location and schedule separate sends.
• Daylight saving shifts. A recurring send scheduled around a clock change can land an hour off. Check the first send after a change.

Pausing and resuming a send#

A campaign that is partway through sending can be paused. Pausing stops the queue where it is: recipients already sent to have their copy, recipients not yet reached do not, and nothing is undone.

Pause when you spot a real problem, a broken link in the main call to action, a wrong offer, the wrong list. Do not pause for a typo in the third paragraph. A paused campaign that goes out four hours later in two uneven halves is more visible to your readers than the typo was.

Resuming continues from where the queue stopped. Contacts already sent to are not sent to again. If you edited the content while paused, the remaining recipients get the edited version, which means your list received two different emails. That is sometimes the right call and it is always worth being deliberate about.

The campaign report#

Once a campaign has sent, its report shows what happened. The figures you will use most:

Delivered. Messages accepted by the receiving mail server. This is the denominator for every rate below it, not the number you sent, because bounced messages were never delivered to anybody.

Opens. Recipients recorded as having opened. Read the caveat in the next section before you build anything on this number.

Clicks. Recipients who clicked a link. This is the most trustworthy engagement number in the report, because a click requires a deliberate action. Track it as unique clickers rather than total clicks when you are comparing campaigns, since one enthusiastic reader clicking six times is not six people.

Bounces. Messages that could not be delivered. A hard bounce means the address does not exist, and hard-bounced addresses are suppressed automatically so you do not keep sending to them. A soft bounce is a temporary condition such as a full mailbox. A rising bounce rate is the earliest sign your list quality is slipping. See Bounces and complaints.

Unsubscribes. People who opted out. A small unsubscribe rate on every send is healthy and means your list is self-cleaning. A spike after one campaign is feedback about that campaign, usually about frequency or relevance.

Complaints. People who marked the message as spam. This is the number to watch most closely. Complaints are weighted heavily by mailbox providers, they damage your sending reputation for every future campaign, and one complaint costs you far more than one unsubscribe. If complaints rise, look at how people are getting onto your list and how easy the unsubscribe link is to find.

Open tracking is approximate

Opens are measured with a small tracking image loaded when the message is displayed. That mechanism is unreliable in both directions, and it has become more unreliable over time.

Opens are undercounted when a mail client blocks images by default and the reader never loads them. Those people read your email and are never recorded.

Opens are overcounted when a mail client or a security scanner fetches remote images automatically, before any human sees the message. Some providers preload every image in every message as a privacy measure, which registers as an open for recipients who never opened anything.

So treat your open rate as a rough indicator that moves over time, not a count of humans. Compare a campaign to your own recent sends to the same audience, where the measurement bias is at least consistent, rather than to a published benchmark from another platform or year. Where a decision matters, base it on clicks or on what happened after the click.

An A/B test on subject lines is measured on opens by necessity. That is one more reason to require a real difference and a decent sample size rather than acting on a one point gap. See A/B testing.

If something goes wrong#

The campaign is stuck or sending slowly. Large sends are queued deliberately. Check the report's delivered count over a few minutes. If it is climbing, it is working.

Delivered is much lower than the recipient count. Look at the bounce figures. A high hard bounce rate on a list you have not sent to in a long time means the addresses have gone stale. Verify the list before the next send rather than sending to it again.

Nobody received it and there are no bounces. Check that the campaign actually completed rather than remaining scheduled or paused, and check the segment resolved to the contacts you expected.