Video analytics

Video analytics covers two related but distinct practices: analyzing audience behavior around published video content, and automatically analyzing the visual content of video frames using computer vision.

Audience-side video analytics tracks metrics like watch time and the audience retention curve to understand how viewers engage with content. Content-side video analytics instead processes the frames themselves to detect objects, faces, or activity, used in security, retail, and content moderation. The two are increasingly combined, for instance by detecting which scenes in a video correspond to retention drop-offs.

Publishers and platforms use audience-side video analytics to guide editing, ad placement, and content engagement strategy, while security, retail, and manufacturing use content-side video analytics for monitoring and automation. A common confusion is assuming "video analytics" always means frame-level computer vision; in media and marketing contexts it usually refers to audience measurement instead, so the term needs to be read in context, alongside how much engaged time the audience-side numbers actually represent.