As the so-called “retail apocalypse” continues to claim victims across the brick and mortar (B&M) landscape, it’s clear that online retail is a force to be reckoned with. While physical retailers still have the upper hand in one area, customer experience (CX), technologies like computer vision (CV), e-commerce businesses are catching up and collecting valuable customer data to help them to cater to their customers’ individual needs.
Still, all isn’t lost for conventional retailers, especially with the value CV AI can provide for them too. Below are 4 CV technologies brick and mortar stores can invest in to gain critical data insights:
As the name suggests, face detection is a CV technique that can instantaneously detect and locate all the faces in an image or video. While you’ll need facial recognition to tell you exactly who it is, just knowing that a face is an image still offers valuable information.
For instance, for you to give customers the service experience they want, you’ll first need to ensure you have the right number of employees to cater to them. This is why most stores seek to optimize staff scheduling by identifying peak hours. Door sensors and in-store check-ins can offer you some insight here, especially when compared to the alternative of combing through hours of CCTV footage to make a manual count. None of these techniques, however, can compete with face detection.
As shown in this screenshot from our Face Detection Model demo, by integrating face detection with your camera system, computer vision can do the counting for you.
The technique can quickly and accurately tally the number of people in your store at any given time, giving you the data you need to determine the busiest days and times for your business.
Another thing that sets face detection apart is that this information can be delivered to you in real-time, alerting you to any unexpected surges or declines in customer traffic, so you can make informed staffing decisions, even on the fly.
While face detection can help you to determine your business’ peak hours, object tracking shows you how these customers are moving through your store.
Currently, online retailers can keep track of their customers as they navigate their site, and beyond. This ability allows them to collect crucial information they can then use to make valuable business decisions, like what products to recommend to who or what banner ads to show when.
Object tracking can offer physical retailers similar data. By being able to see how customers are actually navigating your physical store, you can discern the areas of your store that receive the most or least traffic and station your employees and promotional displays accordingly.
It can also help you analyze your customers’ behavior, seeing where your customers are really spending their time in your store and how long they dwell there. This can potentially indicate, for instance, whether your product or promotional displays are effectively grabbing attention, so you can decide if they are where they need to be, or if your planogram needs to be modified. It can also show you areas where the expertise of your staff may be required (e.g., if a customer is lingering in a particular aisle or section, they may be having trouble making a buying decision.) By monitoring this, you can see it happen in real time so you or an employee can intervene and offer assistance.
Just as this CV technique monitors where and how people are moving through a space, it can also be used to track objects, helping you ensure the safety of your employees and customers. For instance, you could potentially train your object tracking system to alert you if an object, like a motorized shopping trolley, is moving at a speed that violates your safety protocols, so you or a proxy can intervene before anyone, employee or customer, gets hurt.
Object Detection and Recognition
Both online and offline retailers need to manage their stock, but with object detection and recognition, physical retailers can now do this in real time. When combined with a camera system, these can be used to count the number of items left on a shelf or rack, offering some insight into how quickly that stock is moving. When compared to the remaining amount of that item in your inventory, for instance, an excess of that product on the shelf could indicate it’s time to retire it, while a dearth could mean you need to up the order amount.
Since computer vision mimics human vision, object recognition models can be taught to see and differentiate between objects much as we can. Just as you or an employee would notice when an item is out of place in your store, so too can a camera system that has been integrated with object recognition. The difference? The camera can do this continually, allowing you and your employees to focus your attention on tasks that really require human ingenuity, like catering to the individual needs of different customers, only acting where you’ve been alerted that an item is out of place.
For instance, both online and in-person, when a customer changes their mind about a purchase, they can simply remove it from their cart. In B&M stores, however, this can mean them leaving a can of soda in your breakfast cereal aisle or some other unconnected space. In addition to making your shelves untidy, that customer could also unwittingly hinder your stocktaking measures, as your employees will not know that an item that is missing from its rightful shelf has actually not been purchased. Should you train your object recognition model to recognize cans as being different from boxes, however, you can see where each can in the store is, and so find any misplaced cans quickly. Doing this helps you to keep the store environment more organized for both your customers and employees.
While the retail apocalypse will likely lead to the end of some physical retailers, the end of brick and mortar is far from a foregone conclusion. Physical retailers can offer customers a hands-on shopping experience they can’t get online. And while online retailers may still win in areas like convenience or price, technologies like computer vision can now help physical retailers to gather much of the same critical information as their online counterparts, equalling the playing field businesses and making their in-store experience worth the trip.