Credit by Bernard Marr
Founded in 2008 by business school friends Robert Gentz and David Schneider, German-headquartered “etailer” Zalando is as much of a tech company as it is a retailer. Today, it’s Europe’s leading online fashion platform serving 17 European markets with 400,000 products from 2,000 brands. One reason the company can provide a personalized user experience to its 27 million customers is because of the way it uses artificial intelligence (AI) and machine learning just like the other 28 percent of retailers who used artificial intelligence in 2018. Here are only a few of the ways Zalando uses machine learning and artificial intelligence today.
Algorithmic Fashion Companion (AFC)—A Virtual Style Assistant
One of the ways Zalando uses technology to improve the user experience is through its Algorithmic Fashion Companion (AFC), a digital outfit recommendation tool that can generate outfit recommendations in real-time. The algorithm’s recommendations are based on products that the customer has put in their “wish list,” expressed interest in or purchased before. Human stylists provide adjustments to the algorithm to be sure its recommendations follow current fashion trends. AFC helps inspire Zalando customers as well as deliver outfit recommendations from what otherwise could be an overwhelming amount of choices. It’s not only a good user experience for the customer but drives business as well. Zalando reports that outfit recommendations drive 40 percent larger basket sizes. They also found that male customers are highly engaged with the recommendation engine. While AI helps the company scale its recommendations, it anticipates that it will always need human stylists. Zalando is working on how to create a similar recommendation engine for its beauty products as well.
Machine Learning to Understand Customers Better
At Zalando, there are teams, that include 120 researchers, focused on how machine learning, that can help the company better understand fashion, what makes a good outfit, the intent of the customer at the moment they are on the site as well as to discern what each customer likes. Ultimately, with the help of machine learning, Zalando hopes to be able to provide a personalized boutique experience for every single customer. When they are able to provide automated recommendations that are tailored to each individual’s in-the-moment goals, they will be closer to achieving their ultimate goal to become the continent’s fashion platform.
While machine learning certainly showed its abilities in fashion, there are limitations. Celebrities, "influencers" and other social influences are important when driving fashion trends, and it would be hard to imagine AI being able to fully respond to these human idiosyncrasies with automatic recommendations. Additionally, consumers aren't always shopping for themselves. Machine learning must analyze all the information at a given moment, and while it gets better over time with more data, it doesn't hit the mark 100 percent of the time.
Deep Learning for Visual Search
When customers see an article of clothing, accessory, or shoes on social media or worn by one of their favorite celebrities, they often want to try to find it and buy it through on online store. Visual search, where fashion items can be identified by an image and located online to purchase, is available through Zalando’s app and Facebook chatbot. The Zalando team continues to refine its deep learning algorithms for visual search by building on its FashionDNA system.
How Has AI Affected Human Zalando Employees?
While human fashion stylists might be safe for now, 250 marketing and communications professionals were cut as part of a strategy to drive a “personalized customer approach and AI-driven marketing solutions.” The realignment of the marketing team was aimed at being able to provide more personalized solutions for customers using AI. Even though this change was unexpected for the marketing and communications professionals, one first-hand account offers a perspective of the trajectory of change that might be experienced by humans when their positions have been eliminated by machines. Initially there’s uncertainty and confusion, but, in this case anyway, it led to a new start.