How Einstein Engagement Frequency Works
Einstein Engagement Frequency (EEF) considers a prospect’s email engagement data from the past 90 days. The model then focuses on the past 28 days of engagement data to create a status: undersaturated, on target, or saturated.
Required Editions
| Available in: Account Engagement Advanced and Premium Editions with Salesforce Enterprise, Performance, or Unlimited Edition |
Data Requirements and Usage
- To build your model, we use marketing email send data only. Marketing email includes list email sends, and emails sent through Engagement Studio and Salesforce Engage. Data from emails marked operational isn’t used to build your Einstein Engagement Frequency model.
- EEF can generate a model and assign a status from 28 days of engagement data, but it works best after 90 days. More data fills gaps and helps the model identify trends.
- You have a customized version of the model that’s trained on your data alone. Data from one Salesforce customer doesn’t affect behavior for another Salesforce customer. While model training happens for each customer on their own data, the initial development of the model is validated with a representative set of pilot customers’ data.
- The model requires five or more variations of email sends within a 28-day period, based on at least 10 recipients in each group. It’s not sufficient to provide 90 days of data when you regularly send the same emails without variation to the same group of people only.
- The EEF model scans email sends for open, click, bounce, and unsubscribe rates for each recipient.
Example Tamari regularly sends tailored emails to her customers. She maintains a single list of
customers that’s segmented by their location and interests. She sends about four types of emails
based on customer preferences. She sends some combination of weekly product spotlight
newsletters, back-in-stock reminders, birthday coupons, and seasonal shopping guides. Because the
segmentation rule changes who receives each email type, she has enough data to build an EEF
model.
Example James uses a single Engagement Studio program to nurture his company’s prospects with
timely emails. Customers get one welcome email, and then they receive one of two emails depending
on how they engaged with the welcome email. Positive engagement leads to a third email. If they
don’t engage with the second email, they’re removed from the program. This program contains only
four email variants, which is insufficient for EEF to build a model. To meet the data
requirement, James can add a branch to the engagement program, create a second engagement
program, or schedule one-off email sends.
Did this article solve your issue?
Let us know so we can improve!
