Walmart captures 30,000 hours of videos at its stores every single minute, explained Prism Skylabs founder and CEO Steve Russell, but most of that footage goes straight to archive, and is being stored on hard drives without anyone ever looking at it again. The same is true for all that video that’s been recorded by most of the other 15 million security cameras in retail stores — which seems like a wasted opportunity. “In principle, within this video should be some really interesting stuff,” argued Russell.
Prism Skylabs has started to offer retailers ways to analyze all that footage more closely, and Russell showed off some possible applications during his presentation today at Gigaom Structure Data: Stores can, for example, measure which path consumers are taking within a store and then compare these results with what happens when they move a display case or a piece of furniture around.
Of course, uploading security camera data and crunching it for analytics touches on privacy issues. “It’s sort of creepy,” admitted Russell. That’s why his company has developed algorithms that can completely remove captured shoppers from the picture.
Also measurable is the attention that certain products are getting and the impact placement of these products, or even the positioning of a mannequin, is having on consumers. Retail chains used to make these decisions centrally, based on the assumption that there is a golden formula on how its stores should look like. But after analyzing data from cameras, Russell found out that this isn’t the case. “It turns out that all stores and geographies aren’t equal,” he said.
The same may be true for every sensor — and with camera phones becoming ubiquitous and technologies like Google’s Project Tango being just around the corner, retailers might soon have even more ways to capture and analyze video. “We are entering what I think is really a golden age of research,” said Russell.
Photo by Jakub Mosur