Machine Learning to make and automate predications and forecasts ... initial results 99% accuracy in predicting reach and 83% in predicting individual viewing of specific TV show episodes ... next steps to improve predictive accuracy and operationalize for real-time use.
Developing calculations that determine best price options across rate cards to reach specific devices and people (linear only reaches some devices and people efficiently). This informs buyers what the best combinations of linear and addressable purchases are at varying price points. It then enables buyers to negotiate pricing with real-time updates on buy recommendations as pricing is negotiated. ... Conversely, it informs sellers on how to price their inventory to maximize yields.