Before a show premieres, no one knows how it will perform. This fact poses a fundamental problem to content buyers, who will often have to bid on new shows without any indication of its future performance on their platform.
As a part of our professional services stream, we approach the problem from both a quantitative and predictive perspective by using data science. We use our existing massive data set and a neural network model to predict the demand for the pre-release shows.
We have achieved 86.7% correlation between prediction and the actual rank after a release on linear channels or OTT platforms. This high success rate allows the content buyer to make decisions on pre-release titles informed by data and advanced prediction techniques, thus improving their ROI for acquired content.