Scaling up Vignette, our prominence monitoring system on VOD
We've built a new version that allows us monitor more services on more territories
Over the past few months, we’ve been busy building a major upgrade for our Vignette system, which measures the prominence, or discoverability, of titles on various video-on-demand services. Unlike Vignette's first iteration developed in 2023, this new version, which we’re calling Vignette AI, leverages the power of a custom-trained machine learning model to scan the user interfaces of VOD services.
Trained on thousands of annotated screenshots of user interfaces, our model detects content tiles, row headers and other elements that appear on-screen. Our system then faithfully reproduces the complete architecture of scanned user interfaces, enabling us to track exactly where and when each piece of content appeared.
The platform-agnostic nature of Vignette AI has allowed us to drastically expand the number of VOD services and territories that we are able to monitor. While Vignette's first version only worked on three services (Netflix, Prime Video and Disney+), Vignette AI now incorporates local streaming services with ease — AVOD, BVOD and TVOD alike.
Our web-hosted analytics dashboard has also received an update, with new features including the ability to manually group titles for analysis and to inspect logos on promotional images.
This new iteration of the Vignette system is the most advanced solution for measuring prominence on VOD services, and positions Arvester as the partner of choice for all companies, rights holders and public bodies wishing to receive quantitative analyses of discoverability on those services.
The development of Vignette AI was made possible by a grant from the CNC’s Aides aux moyens techniques commission.