Einstein Send Time Optimization in Automation Studio

Reach contacts when they’re most likely to engage by using Einstein Send Time Optimization in a Send Email activity in Automation Studio. Einstein Send Time Optimization uses machine learning and 90 days of engagement data to determine the best time within the next 24 hours to send an email message to each contact.

Einstein gathers data for about 20 factors. Factors include engagement metrics, such as the number of messages sent to and opened by a contact and the day of the week that the message is opened. Then, by using machine learning, Einstein assigns a weight to each factor according to that factor’s influence on predicting future engagement behavior. Einstein analyzes this data and assigns a likelihood score to each of the 168 hours in a week for each contact. Send Time Optimization uses this information to recommend the send times that lead to increased engagement.

You can’t use send throttling with Einstein Send Time Optimization. If you select Einstein Send Time Optimization, send throttling is turned off automatically for this send.

This feature is available to customers with Marketing Cloud Engagement Einstein terms for Corporate, Enterprise, or Pro edition accounts with the Journey Builder add-on. To use this activity, activate Einstein Send Time Optimization for your business unit. Then use Einstein Send Time Optimization in Automation Studio in a Send Email Activity.

Note Before you use the Einstein Send Time Optimization for a Send Email activity for the first time, the information must be optimized. Optimization can take up to 72 hours. You perform this optimization only one time.

Take processing time into account when you choose your time frame. Processing time refers to the time required to calculate the best send time for each contact. It sometimes varies slightly based on factors, such as audience size and system load. This time is included in the sending duration that you set. It typically takes about 20 minutes per 1 million subscribers.

Before the Send Email activity begins, refresh and finalize the target and exclusion data extensions. To prevent an error, the target data extension must contain records. To prevent inconsistent results, avoid modifying data extensions during processing.

Example Cara wants to send emails to 2 million subscribers that arrive between 9 AM and 3 PM. For timely delivery, she adds up processing time and schedules the Send Email activity to start early at 8:20 AM. Because the start time is earlier than the intended delivery window, she extends the delivery duration to 7 hours instead of 6 hours.