Subject Line Optimization mechanics
Important: This feature is only available if the
Subject Line Optimization uses historical email open rates to measure the success of a subject line based on an Orale Eloqua instance's previously used subject lines. Those insights can then be leveraged to help Eloqua users optimize their new subject lines for better open rates.
Note: Subject Line Optimization provides a predicated open rate percentage as well as a Good or a Poor score.
In this topic, you'll learn about:
How Subject Line Optimization works
Subject Line Optimization results can be categorized by a Good or Poor prediction, as shown in the Eloqua email editors. The model also includes a predicted open rate. This predicted open rate provides a deeper insight into subject line performance.
Through machine learning, words used in previous subject lines within an Eloqua instance are associated with open rate performance. This is then used to predict the performance of future subject lines. Each Eloqua instance has its own unique data dictionary when making subject line predictions and whether its future open rate is poor or good. The Subject Line Optimization model calculates a cutoff for the good vs. poor rating, relative to how previous subject lines performed in a given Eloqua instance. This cutoff is consistently updated with more data and the model is continuously trained based on recent subject line performance. The model is automatically retrained once per week. For example, an open rate of 15% may be categorized as poor for one Eloqua instance but good for another. This all depends on each instance’s relative cutoff based on how their recent subject lines have performed.
Note: The order of words in your subject lines could make a significant impact on the open rate. If your predicted open rate is low, consider moving around the words for a better flow.
Subject Line Optimization is designed to be language agnostic. It matches exact words and phrases used in previous emails when available to create a prediction across languages. Subject Line Optimization supports all Latin based languages and partially supports Chinese, Japanese, and Korean. Any language used needs to have adequate data to improve the accuracy of the subject line prediction model.
Oracle Eloqua tracks and stores the following automatic engagement metrics separately from those generated by an actual email contact:
- Auto-opens and clicks by email scanning tools
- Apple email privacy auto-opens
Data requirements
There needs to be sufficient data for a prediction to be accurate. As the Subject Line Optimization predictions are based on previous subject lines and open rates, the more data, the more accurate the predictions will be.
Note: Subject Line Optimization is re-trained automatically on a weekly basis.
- For the model to be calculated, there needs to be 100 unique subject lines
- For the subject line to be included in the training model, there should be at least 100 emails delivered for that subject line
- The model uses the most recent 20k subject lines for training for an instance
- SLO uses unique opens (if an user opens the email twice, it only counts as 1 open)
Subject Line Optimization overview
Oracle Eloqua Advanced Intelligence Cloud Service