Send time optimization delivers each subscriber their copy of a campaign at the hour they are individually most likely to open, based on their own history, rather than sending everyone at one time you picked.
It is the least glamorous AI feature in email and probably the best one, because it requires no editorial judgement, cannot damage your brand voice, and produces a small consistent lift on every campaign forever.
Why "the best time to send" advice is useless
Every study finds a different answer, and they are all right about their own dataset. Tuesday at 10am, Thursday at 2pm, Sunday evening — these are averages across millions of sends to audiences nothing like yours.
More importantly, an average is the wrong shape of answer. Your list is not a person with a preferred hour. It contains people who read email at 6am before work, people who read at 11pm, and people who clear their inbox on Saturday morning. A single send time serves the modal group and inconveniences everyone else.
There is also a mechanical reason timing matters at all: inbox position. Most people read from the top and stop, so arriving near the moment someone opens their inbox is worth more than arriving three hours earlier and being buried under fourteen newer messages.
How the prediction works
For each subscriber, the model looks at when they have historically opened and clicked, weights recent behaviour more heavily, and produces a predicted best hour. At send time the campaign is queued and each message is released at that person's hour, usually within a window you set — say, deliver everyone within 24 hours.
Subscribers with no history get your list's default, or a fallback based on people who look similar in whatever data you hold.
Because the model needs open history, it works better on established lists than on new ones, and its predictions improve as history accumulates.
What the lift actually is
Typically a few percentage points on open rate, with a smaller effect on clicks. Vendors sometimes claim more; treat those numbers with the same suspicion you would apply to any figure produced by the party selling the feature.
A few points is worth having. It costs a checkbox, it compounds across every campaign you will ever send, and it takes no time from anyone. Very few email improvements have that shape. But it is an optimisation, not a transformation — if your open rate is 12% because your list is unengaged, sending at a better hour will make it 13%, and the actual problem is elsewhere.
There is also a measurement problem worth naming. Open rate is distorted by Apple Mail Privacy Protection, which pre-loads tracking pixels at times Apple chooses. For those subscribers the recorded "open time" is not when they read anything, which means the model is learning from partly synthetic data and the measured lift is partly synthetic too. Judge it on clicks and conversions rather than opens.
When to turn it off
Genuinely time-sensitive sends. A flash sale ending at midnight, an event starting in two hours, breaking news. If the content has a deadline, everyone needs it now.
Sequenced sends. If an email is meant to arrive before another, spreading delivery across 24 hours can reorder them.
Very small lists. Under a few thousand, the per-person data is thin and the prediction is close to a guess.
New lists. No history means no signal. Send at a fixed time for a few months first, then switch on optimisation once there is something to learn from.
What matters more than timing
If you are looking for a bigger lever, timing is not it. In rough order of impact:
- Whether the email reaches the inbox at all. Inbox placement swamps every timing effect.
- Who receives it. Segmentation changes results by more than any scheduling decision.
- How often you send. Frequency drives both engagement and complaints.
- The subject line and preview text.
- Timing.
Send time optimization is worth turning on precisely because it is free and automatic — not because it is where the leverage is. Switch it on, then go and work on the first three.
It is included on every IGSendMail plan including the free one, alongside the rest of the AI features, and you can toggle it per campaign so the time-sensitive sends can opt out.
Frequently asked questions
How much does send time optimization improve open rates?
Typically a few percentage points, consistently, across campaigns. Larger claims usually come from vendors comparing against a poorly chosen baseline. The real appeal is that the gain is free and repeats on every send.
Does it work for new subscribers with no history?
Not directly — they get your list default or a prediction based on similar subscribers. Accuracy improves as each person accumulates open history, so the feature is more valuable on an established list.
Should I use it for time-sensitive campaigns?
No. If the content has a deadline within hours, spreading delivery across a day defeats the purpose. Turn it off per campaign for flash sales, event reminders and anything sequenced.
Is it better than sending at the same time every week?
Usually, though a consistent weekly slot has its own value because some readers come to expect it. If your audience is habitual and your open rates are healthy, a fixed time is a reasonable choice; for everyone else, per-subscriber timing wins on average.
