Understanding the wave before it hits: Converging digital and physical CX with edge computing
The potential for monetizing edge computing to improve customer experiences significantly increases with AR. This technology has long been ingrained within retail as a means of overlaying, for example, different furniture on pictures of customers’ living rooms to aid the buying process. More recently, it’s behind the personal stylist trend in which consumers pay clothing experts to digitally overlay wardrobe choices which they can “try on,” courtesy of AR.
Nonetheless, the most cogent means by which AR combines the digital and physical customer experience is via “wayfinding,” in which consumers pinpoint the precise location of a particular item within a retail store by following AR signage in a paradigm similar to Pokémon. Long mentioned a personal use case in which he was able to locate an item in a Home Depot store 5 minutes before it closed by finding it on the retailer’s site—before being digitally guided to it in the store. “It would’ve taken me 5 more minutes to find this thing stuck on the bottom of the shelf, but then I pulled it up in 5 seconds on the app, clicked it, and it was right there,” Long recalled.
The salient benefits of this confluence of digital and physical shopping include:
♦ Decreased costs: By empowering customers to find items in physical locations with digital means such as AR (although other approaches are equally useful), retailers can reduce labor and training costs. The latter is especially significant since traditionally, the only way to help customers in this respect was to devote considerable resources to training staff.
♦ Increased customer touchpoints: The combination of the digital/physical footprint enables the one to burgeon the other. “Now, you know that I just asked for where this product is, and if I’ve moved to where it is, you can assume that I’ve at least looked at it,” Long commented. “You can offer me similar products; you can do add-on sales. You can go ahead and start upselling me right there in my digital experience.”
♦ Customer satisfaction: Empowering customers with the means to find items autonomously increases convenience, customer satisfaction, and the overall shopping experience.
Although it may represent the most controversial aspect of edge computing use cases in retail, facial recognition “is being done way more than people want to admit,” Long revealed. Facial recognition technologies bring up obvious privacy issues and various regulations championing this dimension of the rapid dissemination and analytics of data. Facial recognition is widely used in retail for:
♦ Customer Convenience: This facet of facial recognition connects customers to their shopping profiles, which is particularly effective when most of that profile is compiled via ecommerce. The general premise is that once retailers are able to identify customers via facial recognition, they’re able to individualize their shopping experiences with smart price tags and other means.
♦ Sentiment analysis: Facial recognition technologies are deployed to understand customers’ affective states during their shopping trips. Retailers are attempting to segment customers and personalize their treatment by discerning which emotions customers’ faces reveal. “By grouping customers purely based on visual indicators and running statistics on their buying metrics, then it doesn’t matter what the person’s identity is,” Long explained.
♦ Customer segmentation: Facial recognition technologies are also employed to segment customers according to their facial features—which are themselves indicative of facets of gender, race, ethnicity, religious creed, and more—to better understand customer behavior and produce targeted marketing. Retailers use this technology to ascertain “what do people who look like this buy,” Long said.
The deployment of machine learning technologies is an integral aspect of these facial recognition use cases. Advanced machine learning has long been used across verticals for micro-segmenting customers for unparalleled understanding of user behavior. Once organizations effectively compartmentalize their customer base, they can exploit this understanding to predict customer needs and perfect how they’re treated. Retailers can leverage this insight via interactive price displays and smart price tags. According to Allee, aspects of machine learning and AI are leveraged in retail by “devices that sit at the edge level in distributed remote sites. They actually aggregate the data in real time on site so you don’t have these throughput issues that are holding back some aspects of AI adoption for real-time data and AI analytics.”
Use cases range from the fairly general (changing what’s displayed at gas station pumps) to the specific (delivering offers based on mutable information such as weather data). Other applications of machine learning in retail are based on a desire to make the digital world strikingly similar to the physical world. By analyzing metrics evinced between clicks such as “the distance someone moves a mouse or the velocity of it, we then can work out common behaviors such as frustration and confusion,” Harris mentioned. In this example, machine learning analytics shows what part of the digital experience retailers can improve for customers.