MongoDB was chosen as the NoSQL database for Hudl, a sports video analysis platform, due to its speed, flexibility, and developer-friendliness. The dataset, which stores over 600 million video clips, was migrated with zero downtime and zero data loss using a two-fold approach of background data transfer and multiplexed DAOs. A schema design was created to optimize for frequent operations such as watching video clips and editing clip metadata, incorporating features like denormalization and shard key selection. The choice of TeamId as the shard key enabled targeted queries and improved horizontal scaling. After a successful transition, Hudl has seen significant performance improvements, with average clip load times around 18ms and 99th percentile times under 100ms, allowing for more focus on feature development and less on database scripting.