NRF: ML and IoT in Retail Are Fueling Experimentation
Much has been made about a so-called “retail apocalypse” in the United States and elsewhere. The truth of the matter is more complex. Retail sales during the 2018 holiday season alone hit a six-year high, reaching $850 billion. In that time frame, 93 percent of U.S. adults visited a physical store while 70 percent went shopping in a mall. To be sure, thousands of shopping malls have closed in the past 15 years and traditional U.S. retail giants such as Sears, Montgomery Ward, Toys “R” Us, Woolworth’s, Macy’s and Kmart are either defunct or fighting to remain alive. And an increasing number of consumers are turning to online platforms to have goods — and meals — delivered to their homes.
Yet in-store retail sales outshine e-commerce by a wide margin. In 2017, in-store retail sales hit $3.0 trillion, according to Digital Commerce 360, which excluded items such as fuel, cars and restaurant and bar tabs. In that same year, e-commerce sales reached $453 billion, or 13 percent of the in-store sales figures. The same publication concluded that last year, e-commerce sales drove more than 41.8 percent of retail growth in 2018.
To remain relevant, the retail industry is steadily adopting technology like IoT that can help them enhance customer experience for users with redefined expectations. “More shoppers now tend to be digital natives,” said Charlene Marini, vice president of ISG Strategy at Arm, which unveiled a unified data management product for retail applications at the NRF retail exposition. “They want ease, efficiency and personalized experiences.”
But meeting customers’ evolving expectations can be challenging given how quickly they change and the considerable tech savviness required to successfully deploy machine learning and IoT in retail that often extend to both in-store and online activities. “On the data side, the fundamental challenge is that retailers tend to have multiple silos of data,” Marini said. While retail marketers with access to online customer data often have a significant amount of information about customers, they tend to lack holistic and detailed access to information on merchandising and store operations. “It is often a black hole,” Marini quipped.
See Brett Bonner, VP, research and development at The Kroger Co., give his Internet of Things World keynote on IoT enhancing the user perspective in retail.
But retailers need holistic customer data and insight into the entire customer journey to offer a personalized shopping experience. Providing such data requires not only integrating traditional data streams but often incorporating fresh IoT data streaming from various devices in retail environments.
Once companies have figured out how to integrate their data streams and harness IoT in retail, they must deal with the challenge of ensuring they have the analytics and machine learning capabilities required to act on the information they collect.
At NRF 2019, one of the largest retail events in the world, Arm showcased retail solutions that Marini said can help retailers address all three of these hurdles: unifying traditional data sources, blending it with fresh IoT data and using it to drive concrete actions. “None of them is solved by an average retailer today,” she said. Arm’s customer data platform leverages Treasure Data technology to harmonize and contextualize data sets, while also incorporating data streaming from IoT sources. To address the challenge of making real-time based decisions, Arm is partnering with Reflexis, an in-store operations specialist whose software is in use by 250 retailers internationally.