The TikTok Context Advantage 2026 Report - Explore How Context Drives Performance!Yesterday
Yesterday achieved massive outreach and viewership through video advertising using Mirrors!
Challenge
United International Pictures wanted to achieve significant viewership and drive box office success for the release of movie Yesterday in Philippines. However, it faced a unique challenge due to IP restrictions against the core storyline — The Beatles music and related content. Any piece of music, imagery, iconography or lyrics from The Beatles could not be used in the campaign materials, making reaching their fan base, and Yesterday’s target audience, a huge challenge.
Our Solution
Silverpush’s AI powered context detection platform, Mirrors identified in-video contexts like faces, logos, objects, and scenes, relevant to the movie; effectively serving ads in line with movie related content users were actively engaging with.
Music being the movie’s central theme, Mirrors detected music instruments, faces of band members, music scenarios, and logos of The Beatles and other contemporary bands — helping target relevant content without violating IP restrictions. Mirrors also strengthened the connection between music and the movie by reaching users at the moment when they were engaging with their musical interests.
Mirrors’ In-video Context Detection
Trained on millions of pieces of visual content, Mirrors uses computer vision to identify in-video contexts for highly relevant ad placement targeting.
Identified videos featuring objects such as Guitar, drums, bass, keyboards, amplifiers, and more.
Filtered videos featuring members of The Beatles band and their contemporaries including John Lennon, Paul McCartney, Bee Gees, The Byrds, Pink Floyd, and more.
And detected faces of Yesterday movie stars including Ed Sheeran, Lily James, Himesh Patel and more.
Detected scenes such as bands playing on stage, bands in music videos, rock concerts, and more.
Mirrors detected logos of The Beatles band and many contemporary bands.
Impact
Yesterday surpassed projected box office collections by 45%
Context Relevant Targeting
Mirrors effectively targeted context relevant content, which is not fully possible using standard content targeting.
Capitalized on Users' State of Mind
Served ads to the relevant audience in line with content they were engaging with, not based on past searches or affinities.
Brand Safety
Mirror's built-in brand safety mechanism filtered undesired content that can potentially evoke negative emotions like violence, nudity and more.
Impact in Numbers
Benchmark