Same Message, Better Hour: AI Send Times Lift Push Clicks by 38%

Why we looked

Every CRM team asks the same question: when should this message go out? Most teams pick one hour for everyone, often the brand's busiest one.

Earlier industry research says that hour hardly matters. Studies of billions of marketing messages found that moving a whole campaign to another hour barely changes click rates. We wanted to know what happens when the hour is chosen for each player instead.

Smartico's Best Time model learns when each player is usually active (logins, deposits, bets) and predicts their best hours for every day of the week. We checked whether messages that land in those hours really do better.

How we tested it

A simple comparison can fool you. Players who are active at many hours are both more likely to get a message "in their best hour" and more likely to click anyway. So we compared each player with themselves:

• the same player's messages that happened to land in their best hour,
• against their messages sent at any other hour.

Then we ran the same test again with a twist: every player got another random player's predicted hours. If the lift came only from "evenings are good", this fake test would show it too.

The player's own hour wins

At the median brand, push clicks rose by 38% when the message landed in the player's own best hour. With someone else's best hour, the change was almost zero.

Email shows the same pattern on a smaller scale: clicks +11% with the player's own hour and about zero with a random one. So the gain does not come from the time of day in general. It comes from knowing this player.

This is not one big brand pulling up the average. Push clicks were higher in the best hour on every brand we checked, and email clicks on about 9 in 10.

There is no single best hour

Here is how the model sees one real player. On weekdays, this player is most active in the evening. On Friday and Saturday, the best hours move to around midday.

Now multiply this by every player on a brand. People spread across the whole day.

Even at a brand's busiest hour, fewer than half of its players are in their best hour. So a campaign sent at the "best hour for the brand" still reaches most players at the wrong time.

When the AI picks the hour

The tests above use messages that landed in the best hour by chance. Messages that the model scheduled itself got 35% more push clicks and 40% more email clicks than the same players' other messages. These are different campaigns, so this check is rougher, but it points the same way.

How this compares

• Campaign timing versus personal timing. Industry studies found that moving a whole campaign to another hour barely changes clicks. Our random-hour test agrees: a generic hour does nothing. The player's own hour is what moves clicks.
• Vendor claims. Send-time tools often quote 3–5% more email clicks. Our email click lift (+11%) is above that range, and push (+38%) is far above it.

What this means for operators

• Use best-time sending for push first, then email. These are the channels where the player's own hour gives the biggest lift.
• For SMS, a good time of day does most of the work. Avoid the night. Per-player timing adds only about 5% more clicks.
• Win the subscriber first. Timing only helps players who still receive push. Our push permission study shows when and how to ask without losing players for good.

FAQ

Data and method

How to cite

Smartico (2026). Same message, better hour: AI send times lift push clicks by 38%. https://www.smartico.ai/resources/labs/best-time-to-send-push-notifications

Smartico's Best Time model is one of the Smartico AI models. It chooses the send hour for email, push, SMS and messengers in any campaign step.