Einstein Send Time Optimization
Einstein Send Time Optimization (STO) is a Marketing Cloud Engagement feature that uses machine learning to determine the best time to send a marketing message to each individual contact.
Definition
Einstein Send Time Optimization (STO) is a Marketing Cloud Engagement feature that uses machine learning to determine the best time to send a marketing message to each individual contact. The model trains on historical engagement data (opens, clicks, conversions across send times) and predicts a per-recipient window when the message is most likely to be opened or acted on. When a journey or send is configured to use STO, the platform queues each message for delivery at the recommended time for that recipient rather than blasting the whole list at a single moment.
Send Time Optimization is one of the more measurable Einstein features in Marketing Cloud because the lift shows up directly in open rates and click rates. A campaign that sends at 9 AM in the marketer's time zone reaches some recipients at 6 AM local, some at midnight, and some at noon. STO reshapes that distribution so each contact gets the message at their personal best window. The lift is modest per send (typically 5 to 15 percent on open rate) but compounds across the program because every send benefits.
How Einstein Send Time Optimization picks a per-recipient send window
The training signal: open and click history
STO trains on the engagement events Marketing Cloud already tracks: opens, clicks, and conversions on past sends. For each contact, the model looks at the timestamps of those events and builds a per-recipient distribution of when engagement happens. A contact who opens emails at 7 AM gets a different prediction than a contact who opens at 9 PM. The model needs at least a few months of history to build stable per-recipient distributions; below that threshold it falls back to a default time-of-day model trained across the broader population.
The 24-hour send window
When a journey activity uses STO, the platform queues each contact's message for the recommended hour within a configurable 24-hour window starting from when the activity fires. A journey activity scheduled at 9 AM Monday sends to each contact at their best hour between 9 AM Monday and 9 AM Tuesday. The window can be narrowed (only weekday business hours, only daytime) when business constraints require it. Narrower windows reduce STO lift because the model has fewer hours to optimize across.
Where STO lives in Journey Builder
STO is a property of the Email or SMS activity inside a journey. Open the activity, switch the delivery option from Send Immediately to Use Send Time Optimization, configure the window, and save. The journey can mix STO and non-STO activities; transactional notifications often stay immediate while marketing nurture flows use STO. STO is also available outside Journey Builder for batch sends, configured at the send-time settings layer.
Limitations and what STO cannot fix
STO optimizes for engagement timing, not for content relevance. A message with weak content will perform poorly regardless of the send time. STO also cannot help recipients who have no historical engagement data; those contacts default to the population-level recommendation, which is usually mid-morning. For time-sensitive content (event reminders, flash sales with hard deadlines), STO can be the wrong tool because the optimal window for the recipient may be after the deadline. Disable STO on sends where timing constraints override personal-best.
Measuring STO lift
The honest way to measure STO is an A/B test holdout. Send half the audience with STO and half without (same content, same audience, same time). Compare open and click rates. Most teams see 5 to 15 percent lift on open rate and a smaller lift on click rate. Some industries see more, some less. The lift compounds over the lifetime of a contact because every send through the program benefits. Measuring once at launch is enough to confirm the feature is working; re-measuring quarterly is a waste of effort once the lift is established.
STO versus Einstein Engagement Scoring
STO and Einstein Engagement Scoring are sibling features that solve different problems. Engagement Scoring rates each contact's likelihood to open, click, or convert. STO uses the engagement history of each contact to pick the send time. Engagement Scoring informs audience selection (send to high-engagement contacts more often). STO informs timing (send each contact at their best hour). Most mature Marketing Cloud programs use both: Scoring to decide who to send to, STO to decide when.
Common configuration mistakes
Three patterns waste STO's potential. First, narrowing the window to a few hours so STO has no room to optimize. Second, enabling STO on sends with hard time deadlines (event reminders) where the personal-best window can land after the deadline. Third, expecting STO to fix poor content or audience selection problems; STO optimizes timing, not relevance. Avoid these and STO delivers the modest, durable lift the product description promises.
How to turn on Send Time Optimization for a journey activity
Enabling STO on a single activity is a few clicks. The setup work is in deciding which sends benefit and which should stay immediate.
- Confirm Einstein STO is enabled at the BU
Marketing Cloud Setup, Einstein, Send Time Optimization. The feature must be enabled at the business unit level before activities can use it.
- Open the email or SMS activity in Journey Builder
Inside the journey, click the message activity to open its configuration panel.
- Switch delivery to Send Time Optimization
Under the Delivery section, change the option from Send Immediately to Use Send Time Optimization. The 24-hour window controls appear below.
- Configure the optimization window
Pick a window aligned with business constraints (full 24 hours for most marketing sends, narrower for time-sensitive content). The wider the window, the more lift STO can produce.
- Activate the journey and measure with a holdout
Activate. Run an A/B holdout on the first major send to confirm lift. Once confirmed, leave STO on for similar activities without re-testing each one.
The default. Maximum room for STO to optimize. Best for evergreen marketing content with no deadline.
Constrains STO to a daytime range. Useful when brand voice requires daytime delivery or when SMS rules apply.
Excludes Saturday and Sunday. Sometimes required for B2B audiences or compliance reasons.
Contacts with insufficient engagement history fall back to a population-level recommendation, typically mid-morning local.
STO is configured per activity, so the same journey can mix STO marketing nurture with immediate transactional sends.
- Narrow optimization windows reduce STO lift. A 2-hour window gives the model almost nothing to work with. Default to the 24-hour window when business constraints allow.
- STO cannot rescue time-sensitive content. Event reminders and flash sales with hard deadlines should send immediately, not via STO.
- Contacts with thin engagement history fall back to the population default. STO has little impact on a new subscriber's first three sends.
- STO measures timing lift, not content lift. A campaign with weak content will underperform whether STO is on or off; do not blame STO for content problems.
- Per-recipient send time changes when the journey re-enters the activity. A contact who completes the journey twice will likely get the two messages at different hours.
Trust & references
Cross-checked against the following references.
- Einstein Send Time OptimizationSalesforce Help
- Journey BuilderSalesforce Help
Straight from the source - Salesforce's reference material on Einstein Send Time Optimization.
- Einstein Send Time OptimizationSalesforce Help
- Einstein Engagement ScoringSalesforce Help
About the Author
Dipojjal Chakrabarti is a B2C Solution Architect with 29 Salesforce certifications and over 13 years in the Salesforce ecosystem. He runs salesforcedictionary.com to help admins, developers, architects, and cert/interview candidates sharpen their fundamentals. More about Dipojjal.
Test your knowledge
Q1. What does Einstein Send Time Optimization do?
Q2. When is STO most useful?
Q3. What does STO need to make accurate predictions?
Discussion
Loading discussion…