Create sophisticated Machine Learning Models with AMC Custom Models for Audiences

What launched?

AMC Custom Models for Audiences enables customers to leverage their proprietary modeling algorithms to create custom audiences in Amazon Marketing Cloud (AMC). Powered by AWS Clean Rooms, this solution allows customers to train custom models using their first-party data combined with Amazon signals, without moving sensitive information out of their AWS environment.

Why is it important?

While AMC provides lookalike audience modeling, sophisticated advertisers seek greater control over modeling algorithms and transparency in the training process. AMC Custom Models enables advertisers to build custom model-based audiences for specific use cases including:

• Value-seeking shopper identification
• Customer Lifetime Value (CLV) prediction
• New-to-brand prediction

Customers can enrich their first-party data with unique Amazon signals while maintaining intellectual property protection, as model algorithms are never shared with AMC.

Where is the feature available?

• North America: United States, Canada, Mexico
• South America: Brazil
• Europe: United Kingdom, Germany, Italy, France, Netherlands, Sweden
• Middle East: Turkey, United Arab Emirates, Saudi Arabia
• Asia Pacific: Australia, Japan, India

Who can use it?

• Technology Partners
• Enterprise advertisers

Where do I access it?

• Amazon Marketing Cloud (AMC)

API specific :

AMC Custom Models API details can be found here.