Problem
The publisher operated eight web properties through an external agency responsible for deployment. Every release depended on that agency's queue, so shipping a fix or a feature took days rather than hours. Meanwhile traffic was climbing and infrastructure costs were climbing faster — the platform was not architected to scale efficiently, and every traffic spike widened the gap between revenue and hosting spend.
Intervention
Delivery was brought fully in-house onto serverless AWS infrastructure. Caching and edge routing were restructured across all eight properties to reduce origin load and cut response latency. CI/CD was rebuilt from the ground up around a deployment cycle of under four minutes, replacing the agency-mediated release process entirely.
Result
Traffic and infrastructure cost moved in opposite directions — a rare outcome that came from treating caching, edge routing, and deployment pipeline as one connected system rather than three separate problems. Releases that previously took days now ship in minutes, and page loads across all eight properties are consistently sub-second.