How Macy’s Leans on Google Cloud to ‘Solve Our Search Problems’

By Glenn Taylor 2021.jan.18 sourcing journal

Join us on January 26 at 11 am ET for a webinar with Lenzing, TextileGenesis and sustainable brand ArmedAngels for a discussion on how they’re using Fibercoin technology to realize traceability’s benefits to sustainability, brand loyalty and risk mitigation.

Google might be known as the internet’s omnipresent search giant, but the Bay Area big tech firm has big ambitions in commerce, especially through Google Cloud. Major apparel and footwear retailers such as Macy’s, Kohl’s and DSW and other retail powerhouses like The Home Depot and Ulta have piloted the cloud service for over the past year, but it’s clear that after planting the seeds in 2020, Google is making a big play into retail-specific search, product recommendations and demand forecasting.

Carrie Tharp, vice president of retail and consumer at Google Cloud, shared some of the progress of the programs in an NRF Big Show session, illustrating that the online giant is leveraging its artificial intelligence and machine learning capabilities to eliminate one of the biggest friction points in online retail—finding the right product through search. “It’s really the concept of, how do you bring all that natural language understanding and understanding of query intent, or the matching up of product with the customer?” Sharp asked. “The customer might say one thing, but they think about it differently than a merchant does, or how the product data might be tagged. How do you sort through all of that for the consumer and get them to the right thing?” The Google Search for Retail feature helps retailers improve search results for their own websites and mobile apps using cloud AI and Google Search algorithms. While Tharp didn’t give a timeline on when the search feature would be generally available, Lori Mitchell-Keller, vice president of industry solutions at Google Cloud, noted that an additional feature that was in recent development, Recommendations AI, is now available. According to Mitchell-Keller, Recommendations AI, which is designed to adapt to real-time user behavior and account for variables like assortment, pricing and special offers to empower product discovery, has driven a 90 percent lift in click-through rates, a 50 percent lift in revenue and a 40 percent lift in conversions. Naveen Krishna, chief technology officer at Macy’s, said the department store company accelerated its partnership with Google Cloud at the start of the pandemic, and also advanced partnerships with other vendors such as Klarna, DoorDash and Verizon. The Google Cloud partnership further helped Macy’s as more shoppers bought online during the pandemic, particularly since the department store sought to improve its recommendations and search capabilities. “Instead of just building everything on our own, we decided to leverage Google’s mastery of search and product recommendations to solve our search problems,” Krishna said. “These new platforms and solutions are in the process of rolling out, our initial results are very promising and this is what I think we’ll be doing more of in the future.” The usage of Google’s cloud AI and search algorithms would be beneficial across the board for retailers beyond product recommendations. In a recent research report shared by Tharp, Google said that the top AI/ML use cases for investment by specialty retailers have the potential to drive between anywhere from $230 billion to $515 billion in value. In fact, 10 use cases account for more than 80 percent of the value and fall primarily within three parts of the value chain: merchandising and assortment; product lifecycle management; and logistics and fulfillment. The top use case of AI/ML by a substantial margin is inventory optimization, according to the Google report, potentially delivering as much as $88 billion in value by 2023. Second on the list at a potential $57 billion is design to value, which means retailers can assess material design choices and associated costs for a product launch, ultimately determining the most efficient way to produce merchandise. Coming in third through fifth on the list of top potential use cases are