Modern retailers must move faster, think smarter and put customer experience at the centre of everything they do. And in an industry that runs on data, this means bringing information closer to retail operations.
Speaking at an HPE and Microsoft retail executive event at Chefs Warehouse in Cape Town yesterday, Stephan Steyn, solution business manager at HPE, explained that if retailers want to make real-time decisions, they must analyse their data as close to home as possible.
A while back, a global fast food brand approached HPE with a cooking oil-related issue, Steyn shared. They need to closely monitor the temperature of its cooking oil to ensure food quality, safety and consistency. They also need to change this oil regularly. But the monitors used to track cooking oil conditions fed data into a central data centre, which meant there could be a 10-minute delay between a sensor detecting an issue and the issue being flagged.
“Edge computing changed this completely, turning delayed notifications into instant, real-time alerts and making sure that they don’t have any dissatisfied customers because their chips aren’t cooked properly.” This data is still fed into a central system, but now the purpose isn’t alerts, it’s for training and analysis, he said.
“If there are long queues in your store or if something has fallen off a shelf and spilt all over the floor or if a fridge has turned off and it wasn’t supposed to, retailers need to know about this immediately so that they can respond before customers get frustrated, someone slips or a bunch of products go off because they aren’t cooled properly.”
At the event, HPE highlighted leading computer vision and analytics tools for the retail industry, specifically solutions that leverage existing camera infrastructure, if available, to quickly identify and flag in-store incidents. This technology can also track customers as they move through the store, creating heat maps that can later be used to optimise store layout, identify any abnormal behaviour and ensure that staff are where they are most needed. HPE also showcased the value of HPE GreenLake for retail customers, delivering scalable infrastructure and enabling them to run their IT workloads in a cloud-like manner at the edge.
Being able to process data locally, in real-time, and close to the point of interaction is key in retail, especially in the era of artificial intelligence (AI). Retailers generate and collect vast amounts of data from multiple sources, but if this data is formatted differently, it’s difficult to derive value from it. “By processing data on the edge, you’re getting out of it what you need and making sure it’s in the correct format before uploading it to a central location,” said Steyn.
This is also far less data-intensive, explained Michael Langeveld, head of technology and business development for Emirates and Africa at HPE. Edge devices aren’t used to train AI models; they’re used to predict and classify new data based on trained models. “A few months ago, we worked with a store in Riyadh, Saudi Arabia, to optimise their parking,” shared Saleh Al-Nemer, chief technologist at HPE. “This particular store has a big parking lot and they wanted to understand how it was being used. They deployed a camera solution that uses AI analytics to track who was actually going into the store and who wasn’t. The store found that the parking lot was largely used by people who weren’t customers. So, they started offering customers free parking with their till slips and charging users who weren’t shopping at their store.”
Addressing the retailers in the room, Steyn said: “I believe that all of you have the data to deliver a differentiated customer experience. How are customers moving around the store? What products do they interact with? What's going wrong in the store, and how quickly is that fixed? How long does it take my truck to drive from point A to point B, and can I find a better time of day to send the truck out or a better time to accept deliveries in the store so this process is faster? You’ve got the data, but now you need the right tools and infrastructure to turn that data into meaningful and actionable insights.”
Share