WPP Systematizes Platform and Data Engineering for AI Marketing

WPP integrated its data and platform engineering systems with Google Cloud to build WPP Open, an agentic marketing system that predicts market changes and ad spend. This established a foundation for reliably collecting, cleansing, and providing marketing data scattered across hundreds of global agencies, supporting the use of generative AI.

WPP has introduced a service-based project structure in its current production environment.

• Google Cloud Storage (GCS) and BigQuery are centralized in shared data projects
• Computing and data processing workloads are separated into dedicated processing projects
• Data from multiple product suites is stored in unified infrastructure to serve as a single source of truth
• IAM is applied at the GCS bucket and BigQuery dataset level to implement granular access control

Source data from partners is stored in dedicated GCS buckets. Managed Service for Apache Spark then runs custom Spark jobs to cleanse, normalize, and transform the data into standardized cohort definitions.

By adopting these serverless computing patterns and standardizing data processing workflows, WPP can now deploy targeted marketing campaigns in days instead of months while maintaining security.