Case Study: a major US digital TV streaming company achieves 80% lower cloud costs and 30% lower CPA with Blend360 MLOps

Major US Digital TV Streaming Company cuts cloud costs by 80% with Blend360

The major US digital TV streaming company faced significant challenges in optimizing its real-time bidding for advertisement campaigns. Its existing system suffered from high cloud costs, slow model scoring that caused decision delays, and an inability to effectively run experiments. The company partnered with vendor Blend360 to develop an MLOps solution to overcome these hurdles.

Blend360 implemented a framework-agnostic MLOps system leveraging AWS services, Apache Spark, and MLflow. This automated, end-to-end CI/CD pipeline enabled the training of thousands of models daily and facilitated safe A/B testing. The solution provided real-time dashboards and alerts for data drift. As a result, Blend360 helped the client achieve an 80% reduction in cloud costs, a 30% reduction in cost per acquisition (CPA), and cut bid decisioning time to under one millisecond.

Major US Digital TV Streaming Company