Retailers Are Rethinking Retail Shrink With Computer Vision and Advanced Analytics

Automated Security Tools That Stand Guard

Ananda Chakravarty, who is vice president for IDC Retail Insights and responsible for the retail merchandising and marketing analytics strategies practice, explains that these technologies run on pattern recognition. The machines are trained on extensive video archives to spot suspicious behavior.

“The AI tools are trained to look for the signs of a person shoplifting in a store with a pretty high rate of accuracy,” Chakravarty says.

Over time, the data is trained on hundreds of hours of recorded incidents. “Any good system today will have some of that information already predefined for the training purposes, so automated tools can take over from a security guard that used to be watching a monitor,” he says.

But computer vision’s capabilities now go far beyond simply spotting theft in action. “It’s also tracking items in conjunction with what is being presented at the counter for payment,” he says. “It knows if that scanned item matches the tag or if it has been swapped for another item at half the price.”

Add radio-frequency identification technology into the mix and suddenly, the tools are even more powerful.

“RFID tags, combined with the computer vision and actual store surveillance, allows for a much more holistic view of what’s happening inside the store, strengthening the ability to detect theft or malicious intent,” he says.

Real-Time Analytics Spots Potential Loss Before It's Too Late

Real-time analytics capabilities allow retail organizations to spot potential loss events as they happen. The systems can also trigger direct intervention by flagging a situation in real time and notifying a human in the loop or activating the alarm system.

Meanwhile, edge computing is enabling faster processing at scale by keeping calculations local, removing the need to relay data back to a remote server.

“Ultimately, edge computing allows these computer systems to operate without having to rely on a data center sitting hundreds or thousands of miles away,” Chakravarty says. “That makes for quicker and easier processing of the data.”

The stronger a retailer’s IT and network infrastructure, the more seamlessly it can deploy advanced technologies.

Retailers today want “more surveillance across the board, especially at the point of checkout, and computer vision can help with tracking items all the way through,” he says.

Chakravarty points out that IT leaders must ensure the technology investment is weighed against cost and customer experience.

“If you have thousands of stores in a chain, it’s not so easy to justify the cost, and there has to be a determination made as to whether that 1% or 2% loss for the year from that store is worth the investment” he says.

EXPLORE: The retail technology solutions and services from CDW. 

He cautions that deployments that go too far can create diminishing returns. “The more intense you get with the amount of technology you want to use, it just becomes less cost-effective over time,” Chakravarty says. “Good enough is good enough, and that’s really what a lot of retailers think in terms of how much they are willing to invest.”

The key to a successful computer vision deployment is making sure the tech operates across a large, distributed footprint while delivering actionable intelligence to both store associates and corporate teams.

A Layered Security Approach, Especially at Check Out 

Most retailers are pursuing layered approaches, combining RFID, computer vision and existing security infrastructure.

“We’ve seen most of the advancements being made bundled into the existing point-of-sale system, or existing tools that they already have for internal security,” he says.

Ultimately, retailers must balance loss prevention, operational costs and the customer experience. This means protecting revenue (via real time analytics and surveillance) without overwhelming staff or alienating shoppers.

“Retailers want these systems to be autonomous and work on their own,” Chakravarty says. “The biggest focus is really more deterrence than it is actually catching the thieves in the act.”

Click the banner below to learn about the comprehensive IT solutions modernizing the retail experience.