4 MINS READ
Highlights
- Modern retail stores rely heavily on automation, video cameras, and sensors for managing inventory and logistics.
- These cameras and sensors generate heavy data streams that must be managed in real time.
- A 5G-enabled, edge computing platform can help deliver the high-performance, ultra-low latency solutions required for these efforts to be successful.
In this article
A vision of modern retail
Modern retail stores are designed to enhance customer experience.
The retail industry is transforming rapidly. Modern retailers rely heavily on automation for managing inventory, shelf design, customer service, and logistics. Add to it, video cameras and sensors that allow for unique store design and help enhance the customer experience. Technology is truly powering the retail transformation, setting modern stores apart from traditional brick-and-mortar ones. Here's how:
Buy online, pick up in store (BOPIS): After purchasing a product online, the customer drives to the store to pick it up. A set of video cameras near the store reads the license plate of the vehicle and triggers the fulfilment process, reducing customer wait time.
Customer profiling: Cameras identify the gender and age groups of customers, enabling visitor profiling and personalized experiences.
Efficient store layout: Cameras track the heat map (where customers spend the most time), dwell time (average visit duration), and conversion rate (percentage of visitors to a retail outlet who make a purchase)—helping optimize the store layout. They also aid in queue management, allowing stores to manage the number of available cashiers efficiently to prevent long lines.
Self-checkout: In a cashier-less, self-checkout (SCO) store, weight sensors determine if an item is picked up (or put back) on a shelf. Cameras then determine who picked up the item and add it to their tab. A modern retail store can have as many as 300 cameras per 1,000 square feet (or 1 camera per 3 square feet), analyzing shopper behavior in real time.
Shopping assistance: Some video cameras can detect body language and mood by scrutinizing shoppers' movements and facial expressions. Systems can then alert shopping assistants about customers who might need help.
Supply chain optimization: Retailers must maintain the right inventory in the right store and in the right quantity to meet customer demand. They also must predict how seasonal changes or other events might affect sales. Moving and storing inventory is costly, but the company may lose revenue if customers cannot find the item they want to buy. Cameras and other store sensors can monitor store inventory, evaluating and adjusting the recommended amount of inventory to have in stock.
Loss prevention: Retail shrinkage is the difference between inventory recorded in point-of-sale registers and actual inventory. Shrinkage is usually caused by theft, spoilage of perishables, product damage, human error, and supplier fraud. Cameras and store sensors can play a vital role by keeping a real-time tab on the actual inventory.
Smart mirrors: With these augmented reality devices, customers can try out clothes, makeup, and jewelry without wearing them
Drivers for technological change
Processing all the data captured by these video cameras and sensors in real time is imperative for modern retailers.
It is no easy feat sending all these video streams and sensor data to the cloud for real-time analysis. High bandwidth is required to move heavy data streams. So is low latency for quick data-processing and decision making, especially when robotics is involved.
This is where edge computing and edge-native applications become relevant for retail stores. They allow computing to occur closer to the source of data–right inside the store. Coupled with a private 5G communication network, retailers can deploy cost-effective and high performing 'edge-native' applications.
Key considerations for edge
Retailers need a blend of ‘latency-tolerant’ and ‘latency-sensitive’ applications to deliver a consistent customer experience.
While edge computing is essential for real-time analytics, it has a smaller footprint than cloud computing. For example, theoretically it is possible to use tens of thousands of high-capacity servers for computing on cloud, while edge computing is likely to be limited to a single, high-capacity server. This necessitates a rethink in terms of application deployment and a distributed computing approach.
‘Latency-tolerant’ parts of applications, such as machine learning (ML) training that require heavy computing, can be moved to the cloud. And ‘latency-sensitive’ parts of applications, such as a ML-trained model, can be run at the edge. The need to run some parts of an application at the edge and the rest of the same application on the cloud become important here, and a distributed computing architecture can help.
A 5G-enabled, edge computing platform can help deliver high performance, ultra-low latency solutions to address the modern needs of retailers. A pre-integrated, edge and cloud capability as part of the edge computing platform also helps in deploying distributed applications to leverage the distributed computing architecture
Future of retail
5G and edge computing have much to offer the retail industry.
Retailers can boost the scalability of their services with a combination of ultra-reliable, latency-sensitive edge applications and latency-tolerant cloud applications. They can reimagine their business to deliver a fully automated setup—from store management and inventory management to logistics and supply chain management—that delivers exceptional customer experience. They will also be in a better position to decide where, when, and how data should be processed. Time now for businesses to get an edge with these powerful technologies and accelerate their transformation to modern retail.
