WPP, the global marketing and communications agency, faced a critical engineering challenge: marketing data fragmented across hundreds of global agencies made it nearly impossible to deploy AI tools efficiently and securely. To unlock generative AI’s potential, WPP partnered with Google Cloud to build a unified data backbone and platform engineering foundation.

WPP adopted a service-based project structure, centralizing Google Cloud Storage (GCS) and BigQuery into dedicated shared data projects while segregating compute and processing workloads. Granular IAM controls at the bucket and dataset level ensure teams only see authorized data. Raw data lands in dedicated GCS buckets, then Managed Service for Apache Spark executes custom Scala and Spark jobs to cleanse, normalize, and canonicalize information into standardized cohort definitions (SCDs) based on five keys: age, gender, geo, product, and interest. This uniform structure allows global-scale joins and aggregations without exposing sensitive underlying data. The type-safe Scala framework provides full source traceability for every data point, crucial for enterprise AI compliance.

For platform engineering, WPP developed reusable GitLab CI/CD templates that reduced cognitive load on teams and enforced corporate security standards. These templates manage Cloud Run for full-stack web apps and batch processing, deploy-only workflows, and Cloud Run functions for event-driven microservices. WPP embraced a “build once, deploy many” methodology using cross-project IAM and Artifact Registry to promote immutable container images from development directly to production, eliminating configuration drift and ensuring total parity. Progressive traffic migration allows gradual rollouts.

Security is integrated as a foundational enabler: Wiz scanning runs in the pre-push CI/CD phase, and Google Cloud Identity-Aware Proxy enforces zero-trust access across internal applications. Intelligent VPC template logic automatically resolves conflicts between legacy VPC connectors and modern Direct VPC access, preventing deployment failures. Operational health is monitored via strict metrics: request latency at p50, p95, p99 percentiles, 4xx and 5xx error rates, container startup times (to mitigate cold starts), and CPU/memory utilization.

The business impact was significant. WPP reduced creative and strategy time from four weeks to just three hours, achieved a 70% gain in production efficiency, a 33x increase in content volume, and a 2.8x increase in campaign return on investment. By standardizing serverless compute patterns and data workflows, WPP now securely deploys targeted marketing campaigns in days instead of months, operationalizing its agentic marketing system WPP Open.