Experiment:
IoT and AI-based Digital Out Of Home (DOOH) audience measurement (with partner FastEdge.ai)
Business need
Digital out of home advertising has increased in popularity. From displays at bus stops to LED trucks to screens in local bars, digital out of home advertising is everywhere. The effectiveness of the advertising was usually measure through location ID technology provided by smart phone vendors. The DOOH ad provider used technology to detect the number of cell phones in the vicinity of the ad and — voila — the audience was measured. However, increased consumer privacy concerns have led many technology vendors to make location tracking and cross-app tracking an “opt in” or “opt out” function. As more and more users “opt out” the value of this approach to audience measurement decreased.
Solution overview
Low cost cameras would be directly connected to a Jetson Nano and attached to the DOOH advertising display. The Nano would run an AI Vision application and track the number of people in the vicinity of the advertisement, whether the people were looking at the camera (and thus the ad), their approximate age, gender, and ethnicity. The results would be stored and uploaded to AWS when feasible. Portals in AWS allow the advertiser to see the measurement.
Technology used: Jetson Nanos, assorted cameras, AWS – EC2, S3, network, Quicksight.
Results: Product is up, running and generating revenue at DOOH bar-based advertising in the US and an LED truck system in the UK.