Skip to main content

Getting Started

Portrait Engine allows organizations with multiple customer data sources to create unified and customizable views of their customers. Portrait Engine is a cloud-native system for identifying persons and households from permissioned data sources as client-specific identity graphs. A portrait is a unique identity, usually a person or household that contains attributes. Portrait Engine will match, deduplicate, and merge your permissioned data to establish a unified view of your customers across data sources and silos.

Portrait Engine's identity resolution process includes:

  • Data Sources: Portrait Engine works with datasets from disparate cloud sources, validates them, and combines them for analysis. Data can be processed in place, consumed in batch, or from streaming sources allowing for real-time processing of data. Data sources can include Personal Identifiable Information, licensed first-party or third-party data.

  • Rules: You specify the rules to perform customer matching and construct a First Party identity graph using traits from your data in addition to links from LiveRamp’s graph. Business rules can supplement Portrait Engine’s Machine Learning algorithms. Different applications can use their rulesets to create different views of the customer across business units or use cases.

  • Algorithmic Training and Experimentation: Using the Machine Learning Workbench in Portrait Engine, a user can train an algorithm on the features present within datasets and run experiments and tests against truth sets or synthetic data. Through pairwise matching, identity candidates are specified.

  • ID Generation:  By applying formulas, functions, scripts, or Machine Learning techniques individual identities are intuited and are assigned a Portrait ID. This client-specific ID reflects the custom consolidation that has occurred in the process at a specified resolution level. These IDs will be stable across generation builds as the source data evolves to maintain a stable client identifier.

  • Portraits: The Portrait has attributes passed through from the original data sources and can add calculated attributes such as purchase behavior, lifetime customer value, and more using your data. Other LiveRamp IDs such as ConsumerLinks can be retrieved or created with the proper permissions and data control.

  • Metrics and Monitoring: Results can be analyzed using the resolution metrics produced in the Resolution Dashboard.

The first-party graphs created in Portrait Engine are used for analysis in cloud-based analytics or visualization tools or fed into other downstream systems such as personalization engines, customer data platforms, data management platforms, and other systems. The customer has full ownership of their data and flexibility in how they represent their results.