As part of its new Watch Prompts dataset to help video service providers improve content discovery, Gracenote can offer analogies based on thematically similar content. An example offered is for ‘Barbie’, which could be described as “Legally Blonde meets The Lego Movie”.
This simple yet potentially powerful idea can harness either human editors, machine learning or a combination of both. Explaining what he calls the content comparison mashups, Trent Wheeler, Chief Product Officer at Gracenote says: “In instances where machine learning is in play, technology comes up with comparable content to provide a frame of reference for the target content.
“These assessments are based on content characteristics such as genre, mood and stylistic similarity. Machine generated pairs are subsequently reviewed by humans who decide which combinations make the most sense based on deep knowledge of entertainment content and editorial judgment. These results are used to train the algorithms to ensure progressively better relevance and accuracy over time.”
Gracenote Watch Prompts can also supply critical facts about programming and movies, like prominent award wins and praise from renowned TV and film critics. For example, the information page for the TV show ‘Succession’ could highlight its 13 Emmy Award wins, including two in the prestigious ‘Outstanding Drama Series’ category. There are also talent spotlights, including showcases for popular actors and creators.
In all cases, the new datasets are designed to be paired with viewer preference and consumption behaviour data – thus fuelling personalized viewing experiences. Gracenote Watch Prompts complement both basic programme metadata and Gracenote’s Video Descriptors.
“Streaming services are challenged to continually enhance the user experiences they offer to delight viewers, increase time spent and reduce churn,” Wheeler points out. “Watch Prompts leverages the expertise of our human editors along with scalability enabled by machine learning to deliver an entirely new dataset that will help our customers evolve and meet these challenges.”
As with any of Gracenote’s other datasets, the customer decides how to present the Watch Prompts data to users based on business rules and other proprietary considerations.
