| | AUGUST 20199CIOReviewof clothing and apparel to help cut down on a reason for a return. Bold Metrics is a company that provides a machine-learning fitment tool to retailers. The shopper answers four questions about their height, weight, age and either jean size for men or bra size for women. It uses that information in its machine learning to accurately predict the correct size for a brand. The retailer, SnapSuits, uses one of the products and has a return rate of only 13 percent--lower than the average for its category.It's also being used to make the supply chain more efficient. Machine learning can determine the type and amount of inventory needed for each warehouse so that the product mix is accurate and can be shipped at a lower cost to appropriate shoppers. Kohl's is using machine learning to determine how to fulfill an order from one or multiple stores or its fulfillment centers to lower the fulfillment cost per order.Enabling New Types of CommerceMachine learning is behind the next forms of commerce. In his 2016 shareholder letter, Jeff Bezos pointed out some of the ways that Amazon uses machine learning in its business, highlighting the AmazonGo and Alexa. Machine learning powers the computer vision that allows the AmazonGo store to be cashierless and powers the natural language processing behind Alexa that allows shoppers to use voice to shop.General Motors and IBM are partners behind GM's connected cars with the program, Marketplace. The inclusion of IBM Watson means that the program will learn the driver's routes and habits to provide recommendations and reminders. For example, it will know if the driver missed his usual weekly shop and recommend that he place an order. Another interesting use case is for click and collect orders. The program will allow retailers to see the driver's route to their stores and direct the driver to the correct collection point. Moving Forward with Machine LearningIn 2018, it is important for retailers to understand how to use machine learning in their operations. With numerous use cases available, a retailer will need to evaluate which ones are a priority to the business. Once those are established, retailers will need to know if they have enough data for machine learning to be effective. Without sufficient data, machine learning will not work when built in-house. A retailer can work with third party vendors to overcome this but that means the retailer's data commingles with other retailers' data. The machine learning program gets smarter, benefiting all on the platform, including competitors. It will be important for retailers to decide whether it's worth possibly helping competitors to leverage a better machine learning program.The final piece is to keep an eye on the bigger picture. Machine learning is enabling new ways for shoppers to shop, such as through voice assistants and cars. There are likely many more experiments underway. It will be necessary to for retailers to track new technologies that can enhance or disrupt their business. Machine learning programs can make predictions on the optimal amount of inventory to avoid out of stocks or too much inventory
< Page 8 | Page 10 >