Ask the expert: Computer vision for c-stores
To make better decisions, you need accurate and timely information. It doesn’t get much simpler than that. This applies to individual and organizational decisions the same, but sourcing accurate and timely information is not always so simple.
Today, advanced technology plays a large role in collecting and organizing pertinent information, allowing businesses and consumers to make informed decisions through real-time data. This results in greater c-store and grocer business efficiency and improved customer experiences.
Insights born from advanced data analytics are transforming how businesses operate and identify trends and provide metrics to assess KPIs with greater accuracy. This is not new technology, but until recently, even the best systems were marked by a gap in time between the present moment and reported data. This lag was particularly relevant for food retailers, where inventory turnover is high. Retailers who lack real-time visibility of what’s happening in their stores can face several issues, including empty shelves, misplaced products, and inaccurate inventory reporting on websites and mobile apps. These issues all impact customer experience, operational efficiencies, and ultimately, profitability.
Retail, in real-time
For retail, a subcategory of artificial intelligence (AI) known as computer vision will revolutionize the industry. Computer vision utilizes computers to analyze and interpret the world in the same way that humans do. The capacity of this technology has been relatively limited for a time, but recent advancements in AI technology have greatly accelerated its capabilities.
Computer vision can provide retailers with highly accurate insights and reports about what’s happening in their stores, on their shelves, in real time.
Every shelf can be monitored 24/7, and computer vision can actively interpret what it “sees” to provide live updates and notifications to store employees, integrated business systems, and customer-facing technology like websites or mobile apps. It can discern when shelves need to be restocked, displays need to be tidied, if price tags are missing from a product and much more.
The power of AI
In Canada, food and beverage sales reached $8.5b in January 2023 alone, with hundreds of thousands of different products on shelves. In a single store, the logistics of monitoring every single shelf and instantly processing visual data seems like an immense, impossible undertaking, but computer vision makes it possible.
Google Cloud’s new shelf checking AI technology utilizes the company’s enormous database and sophisticated algorithms in conjunction with high- performance, high-fidelity network cameras to quickly recognize billions of products and help store personnel ensure shelves are stocked.
Modern network cameras from top-quality manufacturers record, process, and transmit high- resolution video to cloud-based AI analytics software where products can be analyzed and identified based on visual and text features.
Products can be recognized from a variety of images taken at different angles from different vantage points. The flexibility of visual input is a key strength in computer vision, as retailers can supply images from a wealth of sources to their
AI-backed systems, from ceiling-mounted cameras to a store associate’s mobile device. No matter the source, computer vision can spot and identify unique products accurately.
The use of AI analytics for surveillance systems means that retailers could have a system that can differentiate between 10 or 11 boxes of a given cereal brand on a shelf and identify if packages are not neatly aligned.
Computer vision can determine if the bar codes on the shelves match the products above them and detect product stock placed in incorrect locations by customers. Computer vision functions as a retailer’s superpower, allowing them to see and understand more of what’s happening in their spaces with new levels of granularity.
Data that can make a difference
Of course, all the data in the world is meaningless unless it can be acted upon or put to good use. The real value of this technology comes through its ability to share key insights with people and systems capable of working with the information it gathers. As an example, notification systems could be designed, which push text messages to onsite or offsite personnel, reported in customizable online dashboards, pushed to inventory management platforms, and used to trigger responses by other technological systems at a retail location, like security or operations systems.
Computer vision can do more than identify anomalous situations, and store managers can program the technology to recognize specific conditions and push notifications or trigger other actions to address them. Managers want their staff notified when a number of units of a certain product remain on a shelf, or if all remaining units are on the top shelf making them difficult to reach. The capabilities of the technology effectively eliminate guesswork and manual checking, enabling retailers to optimize their sales strategies and gain a greater understanding of their inventory flow.
Beyond monitoring product on shelves, computer vision can provide further insights and actionable information through its advanced capabilities. Some of these include:
- Tracking visitor numbers and visitation trends to determine the busiest times of the day/week/month
- Identifying areas lacking traffic with heat maps of customer movement
- Monitoring individuals who loiter near high-value items for extended periods
- Monitoring queue times and having customized alerts when queues become long which notify employees
- Monitoring customer purchasing behaviour by analyzing how long they spend at a given section of shelving picking a product
These can help streamline a retail space’s operation and ultimately support pursuits of profitability.
Computer vision and its growth into the future
Computer vision is poised to change the face of retail, and as more data becomes available, its potential utility will only increase. The timing of this advancement falls perfectly as food retailers continue to navigate a world faced with supply chain issues, heightened customer expectations, and worker shortages. Computer vision could be the superpower that will aid retailers in overcoming these hurdles by providing actionable insights that help optimize inventory management, customer experience, and employee utilization. This all serves to help retailers improve their bottom line and grow.
James Stark is segment development manager, retail with Axis Communications. He brings more than 30 years working with the retail industry and specializes in loss prevention, safety, e-commerce fraud, and supply chain security.
