What Is Send Time Optimization and How Does It Decide When to Send?

Updated

Send time optimization (STO) predicts the moment each individual recipient is most likely to open or click an email, then schedules that person's send for that moment instead of releasing the whole list at one fixed time. It fits teams sending recurring or lifecycle email to a list with enough engagement history to learn from — typically HubSpot and Marketo users who want higher open and click rates without rewriting content. It works less well on brand-new contacts, very small lists, or shared inboxes where one address represents many people.

How it differs from a fixed send time

A fixed send time treats the list as one audience: everyone gets the email at 9 a.m. Tuesday. STO treats the list as many individuals and spreads delivery across the day, so each recipient's message lands near their own active window.

The practical effect is that a single campaign no longer arrives as one batch. Opens and clicks can rise because timing improves, not because the subject line or offer changed. Seventh Sense describes this as per-person send-time optimization and reports email performance improvement in a range of +7–100% across customers, with 500M+ personalized emails orchestrated and scheduled monthly. Treat that range as a vendor-reported outcome that varies by list, not a guaranteed lift.

What signals feed the prediction

The model needs evidence of when a person actually engages. The signals that matter most:

  • Engagement history — when this contact has opened or clicked past emails. This is the strongest input; without it the prediction is close to a guess.
  • Timezone — a 9 a.m. send is a different local hour for each recipient, so timezone corrects the raw timestamp into the person's own day.
  • Inbox provider signals — where the recipient's mail is hosted. Seventh Sense surfaces audience reach by inbox provider (Gmail, Google Workspace, Microsoft Office 365, Microsoft Consumer, AOL/Yahoo, Orange, and others) and breaks delivery, open, click, bounce, and unsubscribe rates down by provider. Provider context helps because filtering and engagement behavior differ across Gmail, Office 365, and consumer Microsoft, so the same send time does not behave identically everywhere.

A useful way to picture it: a contact who reliably opens between 7 and 8 p.m. local time, on a Gmail address, gets scheduled into that window — while a colleague on the same list who opens at 7 a.m. on Office 365 gets a different slot from the same campaign.

What changes in practice

  • Sends spread across the day rather than firing in one batch.
  • Each recipient's message is timed to their own predicted window.
  • Open and click rates can improve with no change to content, list, or offer.
  • Reporting has to be read per person and per provider, not just as one campaign-level number.

Limits and failure modes

STO is a prediction, and predictions degrade when the inputs are thin:

  • New contacts with no engagement history — the model has nothing personal to learn from, so early sends fall back to weaker signals.
  • Small lists — too few events per person to establish a reliable pattern.
  • Shared or role inboxes (info@, sales@) — one address stands in for several people with different habits, so a single "best time" is a compromise.
  • Lists with declining engagement — if people have stopped opening, timing alone will not revive them. Seventh Sense frames this as two kinds of people on every list: the ones engaging and the ones being worn out, and pairs STO with audience recycling and deliverability monitoring for that reason.

Expect results to vary by list. A large, long-lived, actively engaged list is where STO shows the clearest effect; a cold or brand-new list is where it shows the least.

How to measure it before rolling out

Do not switch the whole audience at once. Run a controlled comparison:

  1. Split the audience into a test group (STO on) and a control group (fixed send time), matched on engagement level and provider mix.
  2. Hold content constant — same campaign, same subject line, same offer — so timing is the only variable.
  3. Run for several sends, not one, since a single campaign is noisy.
  4. Compare open rate and click rate between the two groups, and check bounce and unsubscribe rates to confirm the spread-out delivery did not hurt deliverability.
  5. Check results by inbox provider as well as overall, because a lift on Gmail may not appear on Office 365.
  6. Expand to the full audience only if the test group wins on the metrics you care about.

If the test group does not beat the control, the list likely lacks the engagement history STO needs — fix data quality or engagement first.