Why You Must Simplify Your Edge to Accelerate Intelligent Retail

The edge is where about half of new retail applications and infrastructure will be deployed and where much of all new data will be processed.

Retail is poised to transform the in-store experience by using a variety of technologies to enhance the customer experience, reduce waits in checkout lines, deliver personalized notifications, and more. The key to accomplishing these things is to have an infrastructure in place that makes it easy to deploy and manage the edge applications that will power these innovations.

RTInsights recently sat down with Thomas Wallen, Edge Product Marketing Manager of Dell Technologies, to talk about intelligence edge applications, the challenges retailers will have deploying them, and what’s needed to overcome these challenges. Here is a summary of our conversation.  

RTInsights: What are some of the most exciting intelligent edge applications retailers are exploring?

Wallen: The intelligent store is a revolutionary concept in brick-and-mortar retail right now, and the impact of COVID has certainly accelerated this. Labor shortages and the move towards eCommerce purchasing preferences – such as online ordering and in-store pickup – have accelerated within the last two years and are driving retailers to find new ways to increase margins and increase loyalty among customers. The intelligence store is the response from brick-and-mortar stores to these increased and changed expectations.

Some of the specific use cases there that are really exciting are those that enable delivery and in-store pickup. So, micro-fulfillment centers are a very hot topic right now. There are a number of big-box retailers that are converting large parts of their existing retail square footage into warehouse units (similar to mini-Amazon fulfillment centers) that are robotically-enabled to pick out products, put them in boxes, or make them available for associates to easily pack. These centers may also move products into in-store pickup lockers or even prep them for automated drone delivery, which we’re already seeing in our cities.

Such efforts are exciting and are going to revolutionize what we would think of as brick-and-mortar. It’s starting to blur the lines of where you get your products from. Your local Best Buy will soon be where you get your bestbuy.com products.

Other things that are exciting right now are frictionless checkout or the reduced friction checkout world. What Amazon Go has been doing with fully autonomous retail is very exciting. This is starting to catch on across the industry, maybe not in a fully autonomous model, but as a way to increasingly make the checkout process more and more frictionless. You want to reduce touchpoints and queue times to make sure that customers can get in and out with their products without ever needing to interact with a store associate.

All of this is being empowered by what’s known as computer vision, which are very GPU-intensive workloads that must be run ‘at the edge’ inside retail stores. The question is, how can we take a video stream and allow AI to understand what is happening so that we can create actionable insights or determine whether somebody grabbed a Coca-Cola or a Diet Coke so we can charge them appropriately. These things are all happening right now. A lot of really big names that you are familiar with and engage with on a day-to-day basis are pursuing these use cases. Those would be the two areas that I would focus on the most, but when you talk about intelligence stores, there are dozens of use cases that retailers are considering.

RTInsights: What are some of the challenges they are having implementing innovative retail edge solutions?

Wallen: While there are many challenges at the edge, I would say first and foremost that the edge isn’t new. Retailers have been running edge applications for almost as long as technology has existed. The POS system, in a way, is an edge application. It is running in the physical retail stores. It’s communicating with the cloud, and there is a fundamental question of whether the data lives and how it’s managed.

When you start getting into inventory management systems, scheduling systems, pharmacy management systems, ordering, security, and many more. When you look closely, you’ll see that most retail stores are running 10 to 20 edge applications in the stores today that aren’t holistically designed. You have different application stacks and different infrastructure systems. I’ve been to stores where you see four or five or six different commercial PCs that are running mission-critical processes for the store. This is an inhibitor to innovation because it is a tax on IT and store operators.

This is shadow IT in its worst form because it’s not even inside of a data center; it’s inside of a physical store. How is IT ever going to know the stuff exists? When a company wants to move forward with an edge initiative, they have to do it within the context of what exists today. That ‘legacy spaghetti’ has overtaken the retail stores and needs to be addressed before a broader store transformation can occur.

As a result, we see a lot of customers take on their first ‘edge use case’ as application modernization and consolidation by containerizing and virtualizing all these edge applications onto a single IT-managed stack. Once they’ve freed up those cycles, they allow the business to innovate much more freely. This is the same data center virtualization story that we at Dell have talked about for years, but that IT modernization hasn’t happened at the edge yet. Much of the edge is still living in a very bare metal-focused runtime environment. So, the ability to virtualize those applications is not trivial. I would say that’s the primary challenge that we’re seeing retailers deal with to be innovative at the edge.

RTInsights: What’s needed to address these issues and overcome obstacles?

Wallen: The first thing they need to do is to establish the edge as part of their IT framework. The edge is where about half of new applications and infrastructure are going to be deployed and where much of all new data will be processed. Organizations are not going to have success until IT has edge as part of their overall management framework. They need modernized tools to be able to monitor, enforce policy, update, and allocate resources to these edge environments.

That is the first challenge that they’ll need to overcome, and that’s going to require IT to be closer to the business. IT is no longer just going to be a service provider to be told, “Here’s what we want to do, and your job is to do it.” Even if the compute is happening at the edge, there will have to be a lot of centralized enforcement policies so that the business can have operational control of the loss prevention environment or the inventory management environment. In contrast, the operational teams who are actually running the stores can be a lot more flexible about how they plan and deploy new services because IT is on board and can enable them to innovate much more quickly.

RTInsights: How does Dell Technologies help in this area?

Wallen: Our relationship with VMware is really, really critical here. There are many challenges that I have not addressed here that are solved by a number of VMware and Dell solutions overall.

One of our new launches, SONiC 4.0, helps retailers have a continuous network operating system from the core to the edge. SD-WAN is another solution that I’ve seen a lot of companies deploy to bring all their edge environment into a software-defined wide area network. So, all those 10,000 stores or other edge points inside of their distributed compute fabric, from a network perspective, are all virtualized as part of a single WAN or a number of WANs; however IT chooses to parse their network. That gives them a lot of ability to manage their networks. Overall, I would say Dell’s biggest differentiation here, and biggest value to customers is the breadth of our edge portfolio across relevant hardware and software categories, the rugged nature of our products, and the small form factor of our products.

Dell has also overcome many of the supply chain issues that plague our industry, and we have the ability to quickly deliver globally to thousands of locations on relatively short notice. Customers require a partner with our breadth of capabilities and offerings. And, of course, I’ll just emphasize our close relationship with VMware. VxRail is one of the most compelling offering that we see for edge deployments simply for its ease of use. You don’t need IT people on-site to add additional compute capabilities. It is as close to a plug-and-play platform as we see and can easily virtualize and manage your entire distributed compute fabric.

These are critical enables for customers who are investing in the edge in a big way. In this way, the edge will depend largely on a software-defined modality that central IT has full visibility over with minimal overhead costs.

RTInsights: Can you give some examples? 

Wallen: Lowe’s has been on an edge transformation recently, and they’re one of our premier customers at Dell Technologies World this year for what they are doing at the edge. There are a number of interesting use cases that they have started to deploy in-stores. A number of years ago. Lowe’s was behind their peers in terms of its technological offerings development and its edge infrastructure. So, Lowe’s embarked on a top-down IT overhaul of their edge. In 2019, they started with a full network modernization with SD-WAN.

Every one of their 1,800 stores were modernized on a software-defined network several years ago. That enabled them to step up their edge deployments. Today, they are rolling out a full overhaul of all of their infrastructure in their stores right now, replacing their existing three-tier architecture with VxRail. They’ve already done about 800 of their stores, and they plan to do the remaining 1,000 before the end of the year. They’re adding GPU capabilities in order to accommodate computer vision workloads, and all of this is happening seamlessly because of the nature of VxRail. They are ahead of schedule on all their deployments.

As a result, they have massively reduced their cost of management and reduced the time to deploy needed applications. They are a big transformational story that is led by their edge modernization. I can’t imagine what kind of use cases and transformation we’ll see in a year or two as they embrace these capabilities and we see more innovation in their stores become faster, and their time to market gets shorter.

RTInsights: Anything else we should talk about?

Wallen: So when we launched our first Dell Validated Design for Retail Edge at Dell Technologies World this year, initially, we were talking about the application modernization and the consolidation solution with VMware Edge Computer Stack. This solution allowed customers to easily virtualize and deploy all their edge applications, as we discussed earlier. The second Dell Validated Design for Retail Edge is a computer vision solution we’ve launched with Deep North now, delivering something called intelligent video analytics. This is a great way for customers who do not have an overly sophisticated computer vision or edge presence to start ‘digitizing’ their store footprint. This means a customer can take their existing cameras that they have in-store today, apply Deep North software with our validated design, and quickly get data and insights into how customers are passing through their stores, where customers are spending time in the stores, what products they are interacting with, and what products they ultimately purchase.

We like to say that we’re shining a light that is the black box inside the retail store. One way to think of this is in relation to eCommerce. When shopping online, the retailer knows what terms you use to search, what items to consider, what’s in your cart, and all the other information a user has shared online. Compared to a traditional retailer who knows nothing about the customer journey in their store unless they go through checkout (with that loyalty card).

With the Dell Validated Design for Retail Edge with Deep North, customers can chart and analyze the path flow inside their stores and begin to optimize around different segments or times of the day. Offering discounts and bundles that incentivize customers to spend more time in stores or interact with higher-margin displays. While the biggest retailers have been able to gather these insights in highly digitized ‘test stores,’ these metrics have been out of reach of the broader market until now.

This is the best place for retailers to start at the edge because they are getting access to data that they did not have before. And they can use that data to make more informed decisions and make more informed decisions about what capabilities to deploy next, such as endless aisles, augmented reality, personalized pricing and discounting, and even fully autonomous retail.

These changes things are coming quickly, but they require capabilities and skillsets that many retailers are still developing. Until they understand the data of what’s happening inside of their stores, it’s very difficult to go up to the next gear. That’s the most exciting part of the announcement for me, and the retailers I’ve spoken to share that excitement. Together, these solutions open a powerful capability to a lot of retailers and a great stepping stone for those that are maybe a little bit weary or uncertain about getting it to edge technology and the broader intelligent retail trend.

Learn more about Dell Retail Edge Solutions here.