M+E Daily

Gracenote launched a next-generation descriptive metadata solution that uses machine learning (ML) and offers “deeper, more descriptive” info about movie and TV show content, the Nielsen division said Jan. 3.

Gracenote Video Descriptors is the first of the company’s new Advanced Discovery products, and will enable pay TV providers, over-the-top (OTT) services and connected device manufacturers to “unlock deeper user engagement and loyalty in a hyper-competitive marketplace,” it said in a news release.

To generate the descriptive metadata, Gracenote “leverages a combination of expert human editors and advanced machine learning,” Simon Adams, its chief product officer, told the Media & Entertainment Services Alliance (MESA). He added: “Human editors with extensive knowledge of entertainment train smart algorithms to analyze content at scale and apply structured descriptors as well as establish correlations between related TV shows and movies.” 

By “delving deeply into the storylines, characters and settings, Video Descriptors better connect viewers to the full range of available content they might like based on moods, preferences, tastes and past viewing behavior,” the company explained in its announcement.

“While we cannot divulge specific customer names because their products and features have not yet launched, we can say that a number of major media companies have licensed Gracenote Video Descriptors,” Adams told MESA, adding: “Critical to these customers was the fact that the data solution was developed using a balance of expert human editors and advanced technology.”

TV providers are “dedicating increasingly large budgets to build sizable on-demand catalogs of both originally produced content and licensed TV shows and movies,” Gracenote pointed out. At the same time, these companies are “developing smarter search, discovery and recommendation capabilities, including voice and image-based search, to expose viewers to all available content, from big budget blockbusters to the most niche TV programs,” it noted.

“Major streaming providers average around 40,000 TV episodes and movies in their catalogs but put the onus on their viewers to sort through and find relevant programming,” Adams said in the announcement. He added: “By diving deeper into storylines and characters and assigning much more granular metadata to content, Gracenote Video Descriptors give TV providers powerful datasets enabling fresher discovery, recommendation and voice experiences that satisfy the full spectrum of viewer tastes or moods.”

Gracenote Video Descriptors “take entertainment metadata to the next level, enabling clearer understanding of content and subsequently, more personalized video picks,” according to the company. The new Gracenote dataset, it explained, “goes far beyond traditional genres to include descriptors” including characters, mood, scenario and theme. Gracenote Video Descriptors also feature structured keyword sets for individual TV shows and movies that describe content in progressively more granular terms, the company said.

Using HBO’s Game of Thrones as an example, Gracenote pinpoints themes such as “Greed” and “Betrayal,” describes scenarios including “Power Struggle” and “Manipulation,” assigns mood elements such as “Dark” and “Gripping” and classifies the program’s characters from “Royalty” to “Dragons.” Focusing on Scenarios, Gracenote data can be used to identify similar programs based on related descriptors such as “Sweet Revenge” or “Good vs Evil,” it said.

“By better understanding what Game of Thrones is about and relating those attributes to other programs, Gracenote Video Descriptors enable deeply nuanced picks and recommendations that surface the content that will best resonate with individual viewers,” it said.

Additional releases from Gracenote’s Advanced Discovery product suite will follow in 2019, it said.

Gracenote will demonstrate Video Descriptors alongside other innovative entertainment data and technology solutions at CES 2019 in Las Vegas, Jan. 8 – 11, it said.