What it does
This KPI measures the effectiveness of branding in a video asset by quantifying how much attention the brand elements (name and/or logo) receive in the first scene (i.e. early).
Why it matters
The primacy effect in advertising is a well-established insight. It means people remember the first things they see or hear. So, if something important is processed in the beginning - such as the brand - it's more likely to be remembered. Brand cut-through in the first scene has an over-porportional impact on brand recall, correct assignment of the ad to the brand and with that ensures ROI an ad can only impact purchase if it is assigned to the correct brand. This is why Reach is sometimes referred to as contact X brand assignment.
How it works
Scenes are first identified and their key frames analyzed as for static assets. The highest resulting brand cut-through score in the first scene is then used to determine the effectiveness of early branding.
How to achieve great results
- Feature Branding in the First Scene: Ensure the brand name or logo is clearly present in the first scene to leverage the primacy effect.
- Maximize Contrast: Make the brand name or logo stand out by increasing contrast with the background. This helps draw attention to the branding.
- Utilize Hotspots: Place the brand name or logo close to existing attention hotspots in the scene to benefit from the viewer's natural gaze pattern. Leverage heatmovie to identify optimal locations.
- Leverage Center Bias: Position the logo centrally in the first scene to take advantage of viewers' tendency to focus on the center of the frame.
- Add Movement: Include movement near the logo or use animation if feasible. Motion attracts attention and can make the branding more memorable.
AI models used
The branding analysis uses a state-of-the-art optical character recognition model with high accuracy for text extraction. A fuzzy string matching algorithm matches recognized words against the brand name. Logo detection is performed by deep learning models tuned to recognize specific logos. Scenes are detected using a deep learning based scene detection model. Attention scores are based on predictive eyetracking models.
see also: