YouTube monetization: creating user-centric experiences using large scale data

Ramesh R. Sarukkai

Abstract: Over the last 4 years, YouTube has grown from a viral video sharing site to a platform that fuels a win-win ecosystem for video content creators, advertisers and users. A key driving force behind this successful transformation is building out products/platforms that focus and optimize for the user. In this talk, we will talk about user-centric efforts such as launch of user-controlled skippable ads (TrueView Instream), and Dynamic Ad Loads (machine learning system that balances realtime user impact with advertising policies). Leveraging very large amounts of real-time activity data is paramount to successfully building and deploying such user-centric models. We conclude the talk with challenges and opportunities in this important area of real-time user analysis and large data modeling.