
Store theft is a multi-billion dollar problem that has plagued retail since the first artisans opened their wares to the public. Since then, retailers have tried various methods to contain store thieves. installing security tags; creating shelf fixtures that make it difficult to work with multiple products; Use the “safer” case. Changed store layout to make it easier for staff to spot misbehavior.
Retailers are also investing in video systems to try to address this problem, hoping to detect and deter would-be thieves. Video deterrence is based on the idea that would-be thieves avoid surveillance areas because they don’t want to be caught on camera. Detecting thieves using video is more complex and often requires real-time monitoring of retail spaces in hopes of catching criminals in action. While relatively common in the early days (and still employed by a small number of retailers today), this approach is rarely adopted in most cases.
This can be explained by at least two interrelated factors. First, with hundreds or thousands of shoppers potentially in business, asking a human to watch dozens of camera feeds simultaneously for hours is an impractical proposition. While some competent video viewers can successfully look in the right place, at the right time, and with the right person to spot the thief, the reality is that in this kind of environment it reveals a significant minority. It is a difficult proposition to exclude many insignificant people in order to for most humans.
Second, building a return on investment (ROI) model for this approach has proven nearly impossible for most retailers. The labor cost of monitoring and responding to identified events far exceeds the value of the goods recovered. It’s usually cheaper not to catch a thief this way.
The rise of video analytics
However, the development of video analytics, which can be used to automatically identify and alert on predetermined events, has revived interest in using video to combat store theft. Some technology providers now offer systems that they claim can automatically identify theft, generating alerts for store staff. In his recent ECR Retail Loss online session, one of his antipodal retailers using such a system suggested that they were successful in significantly reducing unknown inventory losses.
Balance clarity, complexity, and accuracy
While analytics are certainly beginning to transform how video systems are used, a recent ECR Retail Loss study ensures clarity of purpose, consistency, and accuracy in identifying pre-determined events. It emphasizes the importance of recognizing the impact of environmental complexity on operational success. .
An example of video analytics that works well is a relatively simple binary environment with a well-defined purpose. For example, identifying when a burglar breaks into the back of a building. Under normal circumstances no one should be there. So in the presence of someone, analysts can easily spot the difference and generate an alert.
These factors should be considered when using video analytics to identify thieves in retail stores. For example, it is not always easy to determine whether an aisle event is a store theft. Uniformity of behavior is not a characteristic usually associated with shoppers. Additionally, as different types of selfie scans and his systems become more commonplace (such as mobile scans and shops where customers scan items and legally place them in their bags or pockets), honest shopping It becomes difficult to distinguish between customers and dishonest shoppers. Therefore, it becomes important to determine what constitutes suspicious behavior. Too wide and your staff can be overwhelmed with alerts, and too narrow and criminals can slip through the net.
Additionally, a busy retail store is a complex environment, with many customers interacting with each other, manipulating products, with or without carts and bags, and can be very dense. . This can be a harsh operating environment for video analytics to work consistently well.
Another important consideration is who will respond to the alert. Who receives alerts from these systems as retailers consider how to protect retail staff from violence and verbal abuse, and how much harm they can cause if the outcome turns violent. Some may wonder. Other studies have found that most retail violence is related to store theft incidents. Could it do more harm?
Finally, for retailers with very large stores with multiple aisles and potentially thousands of customers, is it realistic to think that this type of system is scalable? Can you ensure a level of accuracy so that your staff are not simply overwhelmed with potential events they have to deal with?
As with many types of interventions designed to address the problem of retail store loss, under the right circumstances, with well-defined processes in place, this kind of approach can reduce aisle theft. Incidents could be dealt with. In fact, it could be another useful tool in an arsenal designed to deal with problems that will forever plague the retail industry.

video watch is a monthly column authored by Professor Adrian Beck that shares insights on the positive use and impact of video technology in retail. It reflects the latest research and monthly discussions of the Video Working Group. ECR retail loss, the world’s leading think tank on retail losses. Research commissioned by ECR Retail Loss is supported by an independent research grant provided by Genetech and other leaders in retail loss prevention.