The global retail industry suffers from $100B in shrink—financial loss due to missing revenue, wasted inventory, stolen goods, or improper accounting — every year. Shrink is a huge problem for retailers today and can occur at many stages along the retail supply chain, from the distribution center to the checkout lane. The average shrink rate for retailers was 1.6% in 2020, and nearly 16% of retailers now have shrink rates above 3%, according to the 2021 National Retail Security Survey.
The best way to mitigate retail shrink is to improve security and increase visibility into product movement, but doing so at scale is challenging. Retailers can’t have employees re-checking inventory manually at the back door, in-store, at checkout, and outside throughout the entire day. Traditional security tactics simply can’t keep up as shopping methods diversify, which is why visual artificial intelligence (AI) technology (also known as computer vision) is gaining momentum.
AI provides a deep layer of intelligence and a constant watchful eye that can enable retail security executives to maintain inventory count accuracy, minimize shrink and improve real-time decision-making. Through this technology, retailers can boost onsite security and enhance inventory management in a truly scalable fashion.
Computer vision is particularly useful at the front of the store where staff, customers, products, and transaction events all come together. The technology elevates oversight at both self-service checkout stations and checkout lanes, allowing retail security professionals to keep closer tabs on product movements.
Common sources of retail shrink
Computer vision is valuable at the front of the store as it can secure two different types of checkout situations that contribute to retail shrink: malicious and unintentional loss.
Malicious loss typically emerges in the form of customers purposely neglecting to scan certain items or scanning cheaper products in place of more expensive ones. In some cases, store employees will also contribute to the problem by issuing unauthorized discounts. Of course, these actions throw off inventory counts.
But it’s really the unintentional side of the retail shrink equation that is a massive problem. Losses pile up due to shoppers accidentally forgetting to scan items. Some people will mistakenly leave smaller products in their baskets or miss products that they placed underneath their shopping carts. Checkout lane employees also inadvertently forget to scan items, especially smaller goods. These mistakes add up over time, increasing how much downstream work retailers have to do just to keep their records in order.
The challenge for retail security teams is that they often don’t discover issues until long after customers have left the store. Unscanned or improperly scanned inventory walks out the door, intentionally or unintentionally, creating problems later when employees attempt to reconcile sales numbers with inventory counts. In many cases, workers can only determine whether they have too much or too little of certain products without knowing why.
And when employees do have an opportunity to solve a problem at the moment, they still can’t make a perfectly informed decision. For example, a customer could improperly scan an item and place it in their shopping bag, triggering an alert from the machine. When the store clerk walks over, they have to take the customer at their word regarding what caused the error. If the kiosk doesn’t show the last item scanned or take a video of the transaction, there’s little the clerk can do to confirm the event and ensure that the inventory management system accurately reflects what just happened.
The solution to these problems is to increase real-time visibility at checkout, which is difficult to accomplish without computer vision technology. The right tool can prevent these common sources of retail shrink and mitigate inventory management discrepancies that would otherwise persist for far too long.
How AI helps
AI gives retail security leaders the ability to observe all checkout events as they unfold. For instance, AI can monitor live transactions, store video clips, identify un- or improperly scanned items, and flag when there’s an inconsistency between what the camera sees and what the self-service machine processes. The technology is sophisticated enough now that it knows what it’s seeing — high-quality solutions can distinguish a pair of shoes from sunglasses or an avocado from a banana, giving retailers more eyes on what occurs in their physical locations.
Whenever something unusual happens, AI can alert staff immediately, creating a real-time opportunity to prevent a future inventory problem. So, in the event that a shopper mistakenly fails to scan an item, an employee could walk over, check video footage to see what happened, and then re-scan any products that weren’t processed correctly.
Beyond checkout lanes and self-service stations, the technology can improve visibility into curbside pickups, buy-online, pickup-in-store counters, and other eCommerce-oriented channels that people are using more frequently today. In this way, computer vision is helpful for tracking product movements within newer-age shopping models that are only gaining momentum.
For those struggling to minimize shrinkage and efficiently manage inventory, computer vision technology is the answer. AI solutions, designed specifically for retail, perpetuate inventory accuracy, streamline operations and enable companies to deliver better overall shopping experiences. They increase visibility across the board, augmenting human workforces with the information they need to avoid inventory headaches and make better decisions in real time.