Social video is an immense and complex place for advertisers. Ad campaigns typically run across hundreds of thousands of YouTube videos. Some videos are clearly suitable for advertisers – and some are definitely not. The hard work happens in between.
The problem of content safety and suitability is platform specific. Solving it requires deep thought about the YouTube content and advertising ecosystem. This has been OpenSlate’s primary focus for the past six years.
OpenSlate has access to data for hundreds of millions of ad-supported videos and ad delivery data for thousands of advertiser campaigns.
These data sets are used to develop unique insights and proprietary metrics about the nature and quality of YouTube content. YouTube has certified OpenSlate metrics for their methodology, accuracy and clarity. This is the data that powers our content modeling and auditing software.
Our engineering team has developed proprietary methods for ingesting, analyzing, storing and accessing this data. OpenSlate’s data environment serves as a research and development lab for our data science team to build the models that drive our clients’ decisions.
Our lab is currently utilizing supervised and unsupervised machine learning techniques, natural language processing and advanced statistical modelling to better understand the nature, quality and suitability of YouTube content.
Over the past six years, we’ve developed models to measure and predict content quality, brand safety, subject matter expertise and demographics, to name a few. Our models are extensible across languages and regions and are constantly evolving and improving. Social video never stops expanding; neither do we.