Many companies were demonstrating products and services designed to improve business performance at IBC 2019. However, some have zeroed in on how to better leverage the movies and shows in existing libraries. In each case, artificial intelligence/machine learning (AI/ML) is playing a valuable part.
Conviva’s Content Pathway
Conviva is moving far beyond its roots in video quality measurement. The company is using AI to help analyze its vast pool of streaming video performance data to drive a new service called Content Insights. The service gives content owners and aggregators a unique view of how video is consumed in the home. For example, the Household Consumption Graph shows which service or content performs best through each screen active in the house.
Content Insights also allows users to see a customer’s video journey. Household-level Content Pathing enables a user to see common viewing routes through different shows and events. For example, a set of viewers often start watching The Sports News Show and move to ABC Basketball while another group starts at The Prank Show and moves to a DEFG Basketball tournament.
Armed with a clear picture of viewing habits and trajectories in the home, video service providers can improve marketing campaigns, content licensing activity, and pricing and packaging approaches.
Wicket Labs’ Attention Index
If you are looking to uncover hidden gems in your content library, Wicket labs Attention Index could help you find them. The company analyzes the length of time viewers watch a video. It subtracts the percentage of viewers that watch less than a quarter of the video from those that view more than three-quarters of it. The result is the attention index for that title.
For example, if 30% watch an hour show for less than 15 minutes and 40% watch for more than 45 minutes, the attention index is 10 (40%-30%.) The higher the index, the better the attention a title generates. A hidden gem is a title that doesn’t receive many views but has a high attention index.
Wicket Labs is using AI/ML to uncover other critical information. For example, it derives a customer happiness index (CHI) from many different data pools within a video business. The AI can also help recommend how best to approach a customer that is most likely to churn. Perhaps leading with a recommendation for a hidden gem could help save them.
Vionlabs’ discovery powered by emotions
Vionlabs contends that the vocabulary used in metadata is still remarkably limited. For example, the critical metadata phrases describing The Martian and Guardians of the Galaxy 2 are almost the same, resulting in recommendation engines frequently pairing them together. However, the two movies are very different.
The company generates a digital fingerprint for videos that provides far more details. It uses an in-house AI engine and algorithms to analyze audio, metadata, visual elements, and emotional structure. The souped-up content descriptions allow for many improved recommendations. For example, much better matches for The Martian include Passengers and Arrival.
Why it matters
AI/ML is beginning to find its way into real products and services for the video streaming industry.
It is having a significant impact on the area of content optimization by allowing content owners and aggregators to:
- See better how content is consumed
- Identify undervalued movies and shows
- Improve video recommendations.