The main cause of shrinkage in the UK is theft which accounts for over £5.5billion per year with over 1,000 daily incidents happening at the checkout! The top stolen items include the likes of delicious cheese, alcohol and sweets. A common way that sneaky shoppers are taking full advantage of less-than-intuitive self-checkout systems is by replacing expensive products with in-expensive products of similar weight!
In an effort to reduce retail shrinkage, the world’s biggest retail companies are now considering the implementation of new solutions that can produce both reactive real-time data and consistent analysis over time to prevent future theft.
Modern Store Security = Real-Time Shrinkage Data
Whilst self-checkouts improve a customer’s overall experience, they are a prime location for theft to take place as there is no member of staff overseeing the transaction, in most cases, there is one member of staff for 8 separate checkouts, that’s if there is one available. Due to this, there is an 30-70% increase in theft at these types of tills.
Implementing machine vision technology is like having hawk eyes in the back of your head. The technology is trained on both people and object identification to help prevent theft within retail stores. Using existing security cameras, we can retrofit the technology to begin processing footage at the edge using machine learning algorithms to detect suspicious behaviour, e.g. body language or items being placed in bags. However, because all images are processed at the edge it means no personal data is stored and less metadata is sent to the cloud, providing a quicker system that can send real-time alerts so no time is lost, which can be critical when dealing with theft.
The Science Behind The Solution
At Hark, we use specialist technology which is trained on object identification – using barcodes from the point of the product being picked up to the point of sale. The technology is trained on specific inventory and can also be trained on your own-brand products. Our unique features enable us to detect products that have been seen but not scanned or vice versa. If the products have not been scanned on a self-checkout till, this intuitive technology will automatically shut down the tills, whilst alerting staff managers or security.
If all barcodes were monitored, it would mean crime such as external theft or sweethearting for which 75% goes unnoticed, could be prevented as the systems would alarm staff when the incident is taking place. Sweethearting often involves actions like return fraud and unauthorised discounts which is more difficult to identify, whereas with object identification we can understand when an item is removed from a shelf and taken straight to the return desk and that the product was never bought in the first place.
How Does It Help? By Identifying Patterns!
Machine vision solutions collate all historic data (which can go back years) and when combined with AI and highly trained machine learning algorithms our tool is able to identify patterns and anomalies within shrinkage data.
That means store managers can understand which items are experiencing the highest level of theft in their store; this opens up an opportunity to take extra prevention measures such as using security tags on those items or moving them items to a more visible stand.
The analysis can be categorised on a seasonal and geographical level. Outlining which stores have the highest levels of shrinkage in general, which can be then investigated further.
Did you know: Christmas is renowned for the highest level of shrinkage, being reported at around £1 billion!
These insights enable managers to make more informed decisions and prevent future losses from occurring. The goal is to reduce the overall shrinkage rate in real-time, which will prevent those in-the-moment losses, along with preventing the overtime shrinkage by implementing new measures. All of which can be achieved with transparency over accurate retail shrinkage data.
Revolutionise loss prevention in your store, get in touch today!
