A new feature from ThinkAnalytics recommends images selected directly for an individual viewer, based on their interests and profile.
Personalized Images takes the images from multiple catalogs on the UI, content carousels and program guides and applies advanced AI techniques and information science to automate the process of selecting the best and most appealing image for each viewer.
“With competition for viewers so intense, video service providers are constantly on the lookout for new ways to drive up consumption on their platforms. Personalized Images offers a really powerful way to help viewers find highly appealing content to watch. The results speak for themselves with our customers registering viewing uplifts of up to 15% when compared to standard poster images,” said Peter Docherty, ThinkAnalytics CTO.
Complementing ThinkAnalytics’ Content Discovery Engine, Personalized Images supports images and metadata from a range of sources, including those from customers and third parties. Results from early customer deployments have produced impressive and consistent uplifts of 10-15% compared to displaying a single standard image.
ThinkAnalytics’ says by dynamically selecting program images, Personalized Images helps to increased viewership, monetize content and increase the number of titles and hours watched per viewer.