This article aligns to the Silver User Certification, which focuses on independently launching and managing effective campaigns. Click here to learn more.
Send Time Optimization (STO) is the ability to send a campaign at the best time of day for each person in the audience. It is fully personalized on a per customer basis and based on the 24-hour calendar day.
NOTE: This is an advanced feature, please contact your Customer Success Manager to upgrade.
If a customer is scheduled to get multiple STO campaigns in a single day, it will decide how many emails should be sent and at what times of day they should be sent in order to optimize for site traffic over the day.
NOTE: You must publish this campaign before midnight the day of the send for the campaign to send as expected.
ADD STO TO A CAMPAIGN
Under the scheduling setting (When do you want to send your campaign?), you will have the option to make a campaign send at an optimized time for the following campaign types:
Simply select “Bluecore will decide the best send time for each customer”:
CONTROL GROUP AND TIME ZONE
We will automatically hold a portion of the campaign’s audience to serve as a control group for the send. Please contact Bluecore to configure the size of the holdout and time zone.
Navigate to the flag icon on the left-hand side of the navigation bar.
Go to Communicate.
Click the Send Time Optimization tab on the marketing settings page.
Reach out to your CSM for any reporting requests, we have out-of-the-box reporting available from our Product Support team for insights around delivers, open rates, click rates, revenue-per-email, and more.
STO TECHNICAL OVERVIEW
In order for Bluecore to determine the best send time per customer, we need several things:
For each customer, find an estimate of the value of each hour (probability to click on an email at each hour for that person). We would then have a value function (defined over 24 hours) for each customer that would indicate how valuable sending an email at a certain hour would be for that person. We can get this value function via a regression mechanism on historical data (past delivered times and clicks) and it will be personalized. We may also use other historical data (other online activities) to have a better sense of the customer’s personal online activity times or to group people together and find a common value function for people that “look” alike.
From that value function, we can find the best combination of x hours (for x going from 1 to the maximum number of STO campaigns we need to support) to maximize site traffic over the whole day. In order to find those, we also need a function that gives the probability of an email being sent (e.g. not being blocked by any frequency filters) for each time. The x-length tuple of times will be called strategy x.
Decide between different length strategies by looking at the customer state when they are eligible at some point in the day to receive an STO campaign. The customer state can be defined by many attributes (total number of campaigns the customer is eligible to receive at that point in time, how late/early it is in the day, how many emails in each strategy are left, etc…). We estimate the value of sending/not sending and make our decision based on that.