A London-based team of data scientists at Hoxton Analytics have developed a technology that categorizes and serves ads to people in stores based on the shoes they wear.
The technology relies on machine learning to identify a customer's gender with up to 80% accuracy, according to the company. The product sits in a small unit that houses a camera and a processor. It's typically installed low in a doorway or corridor to gather images of people’s footwear as they walk past. From these images and multiple layers of machine learning and artificial intelligence it can predict gender and age.
Hoxton Analytics' scientists believe by analyzing the style and size of people’s footwear as they walk past the sensors, the system can make real-time judgments on demographic and footfall. The shoes worn by a consumer makes them look like the type of person who might want to buy a blue dress with polka dots.
The idea is that the method in which the technology collects data to serve ads is less intrusive than taking a picture of someone's face, though people who venture out into public have their picture taken anonymously multiple times daily in retail shops, restaurants, banks, and airports.
Anonymity goes a long way when it comes to building relationships. Owen McCormack, CEO of Hoxton Analytics, told the Guardian that personally identifiable information isn’t being harvested, and that today a lot of shops do some pretty creepy and invasive things.
"If you know that everyone in Argos right now is a male, you’d be advertising PlayStations not hairdryers," McCormack told the Guardian. "For me that’s fine – you’re giving the retailers accurate information but you’re not saying this person is so and so who lives down the road and has an expensive house."