Upload data for the Target personalization algorithms

Offline data, such as CRM information or customer churn propensity scores, can be incredibly valuable when building personalization models in Adobe Target Automated Personalization (AP) activities.

There are several ways to input data in Automated Personalization (AP) and Auto-Target personalization algorithms. In addition to the methods in Methods to get Data into Target, Experience Cloud shared audiences (Adobe Analytics, Audience Management){target=_blank} and in-activity reporting audiences are also used in our algorithms.

For information about the data automatically collected and used by Automated Personalization and Auto-Target personalization algorithms, see Automated Personalization Data Collection.

Best Practices

The following list presents best practices for uploading data for Target’s personalization algorithms:

• The more high-quality data that is available to Target’s personalization algorithms, the better the quality the resulting models in your AP and Auto Target activities.
• Limit using multiple profile scripts or attributes that serve the same purpose.
• Don’t pass a unique Id, such as a session ID if not needed.
• Review what data Target automatically collects ( Data Collection for Target’s Personalization Algorithms) so that you don’t send duplicate information. For example, Target uses IP addresses to determine visitors’ ZIP codes. It is not necessary to pass this information as a separate variable.
• Do not pass multiple values in the same attribute/variable. If multiple variables are concatenated, Target’s personalization algorithms treat each string as a unique value, reducing the value of the information for personalization.
• Use a memorable and meaningful naming convention to make your Personalization Insights Reports more understandable.

Business.Adobe.com resources

Multi Armed Bandit Testing Enterprise Governance Mobile Optimization Behavioral Targeting Recommendation Engine Multivariate Testing