Adobe is excited to announce Trait Exclusion in Algorithmic Models, an enhancement to Audience Manager lookalike models, powered by Adobe Sensei. This new capability provides unmatched transparency and control in lookalike modeling by making it possible to exclude selected traits from model analysis, ensuring lookalike output is valuable and focused on relevant audience traits.
Trait Exclusion enables use cases such as:
• Assuring that common catch-all traits, such as site visitors, don't bias a lookalike model, leading to flat results
• Removing unfamiliar or undesired traits from a data source to better understand the influential traits
• Excluding certain traits - such as demographic traits - from a third-party data source, thereby using third-party data in compliance with data privacy regulations
Customer impact:
Trait Exclusion in Algorithmic Models was rolled out on October 31, 2018 and is available to all Adobe Audience Manager customers. This enhancement is built into the standard workflow for Audience Manager lookalike modeling; there is no incremental fee or provisioning necessary to begin using this feature.
Additional technical documentation for Trait Exclusion in Algorithmic Models can be found in Audience Manager Product Documentation. A complete overview of Audience Manager Algorithmic Models is also documented.
For questions or concerns, contact amsupport@adobe.com.
