Retail customers are one of the early adopters of video analytics technology. Here we explore what retail video analytic features they want the most.
The developments in AI have helped video analytics offer several new features to customers across verticals. Perhaps nowhere has this been more visible than the retail sector, where video analytics has shown clear potential to provide value beyond security. The popularity of retail video analytics has prompted many companies to venture into this sector.
“Video analytics has provided one more tool to retailers who have always relied on gathering and analyzing data for doing things like optimizing product placement and maximizing customer engagement,” explains Amol Kulkarni, VP and Country Head of Dragonfruit AI. “In addition, loss prevention has become more efficient with the ability of video analytics solutions to generate alerts on various events, allowing retailers to optimally deploy their limited resources.”
While the abilities and accuracy of AI-enabled video analytics in retail continue to grow, understanding what the customers want will drive the developers in the right direction. This article explores the top video analytic features that retail industry customers worldwide want.
1. Occupancy metrics and movement patterns (footfall, max occupancy, heatmaps, dwell times, engagement times)
There is an increasing demand for retail video analytics that can improve customer experience and optimize store operations, driven mainly by the growing capabilities of deep learning and the intelligence being infused into these analytics.
Jammy DeSousa, Senior Product Manager for Security Products at Johnson Controls, gives two interesting use cases where such video analytic features have been used.
“A retailer in a shopping mall or city street can measure the success of a window display or promotion on display with cameras on the exterior of the store,” DeSousa explained. “Video analytics can correlate the number of shoppers who stop to view the windows vs. the number of people actually entering the store. Similarly, a drugstore or other retailer can track conversion rates by measuring how many people dwell or linger in front of the greeting card display vs. how many actually purchase a card.”
2. Loss prevention (loitering detection, shoplifting, vandalism detection)
Security may be listed second
but not any less important for video analytics customers in retail. Thanks to video analytics, surveillance cameras that were just recording footage for forensic purposes can report criminal incidents in real-time. Kulkarni explains that video analytics that can detect loitering detection, shoplifting, and vandalism are among the top features that customers are looking for at the moment.
“Retailers are continuing to use the standard security and life safety analytics such as object detection and object removed to aid in loss prevention efforts [to prevent] internal shrinkage and shoplifting,” DeSousa added.
3. PoS integration
Integrating video analytics with PoS transactions is yet another feature in which customers are interested. DeSousa said that his company sees use cases resulting from a combination of analytics and integration with Point-of-Sale systems to measure the critical retail metric of conversion rates or automate actions to improve the customer experience.
“With many retailers increasing their number of self-checkout stations, many stores maintain a ratio of dedicated self-checkout personnel to assist customers with the checkout process based on the number of checkout stations,” said DeSousa
Artificial intelligence plays a significant role in identifying the difference between store personnel and customers to help stores maintain the correct staffing levels. PoS integration can also ensure that what is being scanned by the customer is the correct product and that all products are being scanned.
4. Investigations & reporting
A major hurdle to using video surveillance data for investigations is having to search through hundreds (or more) hours of footage to find an incident that might have just been a minute long. Analytics has made this job easier with the visual search feature. Customers can now search for specific objects, like a car or a color, to filter the video to narrow down any search during an investigation process.
5. The COVID-19 features
Kai Bisgwa, VP of Sales & Business Development at Safety & Security Things, pointed out that throughout the past year, retailers have been primarily concerned with implementing solutions to solve pandemic-related use cases, like mask compliance, occupancy management guidelines, and social distancing.
“However, as we’ve shifted to a space of slowly realizing what life will entail going forward, grocers, convenience stores, and supermarkets are looking to optimize operations further and increase sales,” Bisgwa added. “Smart surveillance can assist by combining video analytic applications data with POS systems to allow the calculation of conversion rates down to specific store departments. Adding demographic attributes, which can be analyzed by analytics, an even more comprehensive analysis of shopper profiles and behavior can be obtained.”