Analytics for 360 media have traditionally been very qualitative. While this has been somewhat successful, we set out to do things the right way and focus on a quantitative approach to analyzing both 360 videos and photos.
Inspired by our ObjectExplorer system, which tracks visual gaze to specific 3D objects alongside controller engagement data - our solution to surfacing these insights is a new tool we’re calling Points of Interest (POI).
The POI tool allows customers to create hotspots within time series video or images that identify specific areas of value. These hotspots then provide a bounding box for our system to track visibility, gaze-at-ratio, and frame by frame metrics of user engagement.
The tool helps customers test content quality and user engagement with narratives, the quality and visibility of product placement, or even the level of engagement with ad units.
To use the tool, the user simply creates a new POI directly from our dashboard (above). Once the POI name and shape is selected, the user correctly sizes the POI around the target content, then selects in the timeline where this POI appears in an easy to use video player (below).
Once the POI is in place, the system, will automatically monitor inbound user gaze data for every customer to identify hits to our Points of Interest. This results in a quantitive measure - a solid comparable metric that can be benchmarked and measured over time, and across content. (below)
We are very excited to have brought Point of Interest to market, as it provides a quantitive measure of success for video and image content. With this foundation in place, we 'know the things that are valuable', much like a pageview or click, and can further our investment in machine learning, and computer vision research.