Einstein analyzes most of your commercial subject lines in order to detect and surface interesting insights and patterns in the language that you use.
When subject lines across multiple sends are identical or similar, Einstein automatically groups them. Then Einstein calculates a single, unweighted average performance rate to more accurately understand language performance. These grouped subject lines are the primary input data to all of the Einstein Copy Insights models, including those models used to generate language-related insights and performance predictions.
Often there’s a correlation between when Einstein detects a similar word or phrase and the AMPscript that’s used, but such a correlation isn’t guaranteed. Static text isn’t always recognized as unique across sends. AI-driven consolidation happens based on the language found, regardless of how the subject line was created, when it was sent, or who received it. Some examples:
- The subject line for sale email is the same for all recipients except for the product or discount value.
- A birthday message is the same for all recipients except for the recipient’s name.
AI cosolidates %%firstname%% in Content Builder, but more complex AMPscript strings aren’t always consolidated properly.
You can see language with AMPscript and ___ in ECI dashboards for up to 90 days after you activate ECI. ECI doesn’t collect fully rendered subject lines until the app is activated. To help decrease the amount of time it takes to create meaningful results, ECI pulls from a blend of rendered subject lines and subject lines as they appear in Marketing Cloud Engagement. Subject lines can include AMPscript.
A marketer for a global charity uses Journey Builder to manage donation follow-up. They refresh content quarterly, using AMPscript like %%=v(@customSubjectLine)=%% in their subject line. The marketer can dynamically use one of these three options based on the type of donation.
- %%firstname%%, your donation counts!
- Thanks for the donation, %%firstname%%.
- We love the love
When Einstein evaluates historical subject line performance each of the three options are considered independently, even though they can be part of the same send. For example, consider these engagement rates.
| Overall Send Engagement Rate | 33% |
| %%firstname%%, your donation counts! | 45% |
| Thanks for the donation, %%firstname%% | 22% |
| We love the love | 31% |
The unweighted average engagement rate for the send is 33%. But test subject lines are evaluated against the language and language-specific engagement rates (subject line groups) to determine potential effectiveness.
If the marketer sends a separate one-off email with an identical or similar subject line to one of the three options shown, the subject lines are grouped. Subject lines are represented one time in Einstein Copy Insights.