I love shopping at a Nordstrom store, but hate shopping on the Nordstrom website. Unless I know exactly what I want, like my favorite shade of lipstick, I have a time trying to find what I might want.
Take a search for a blue dress. Nordstrom served up nearly 3,000 options and after narrowing it down to casual styles, I had a mere 1,000 to go through. Included in that search were jumpsuits (sorry, not a dress), plus sizes and maternity dresses.
I have been a longtime member of Nordstrom’s loyalty program, so it knows my size, my style/brand preferences and also knows I am of an age where I am not going to have a baby.
Nordstrom should take all that it knows about me as a shopper and serve up a first page selection of blue dresses that are most likely to catch my eye. Instead I have to wade through page-after-page of irrelevant selections. After five or so, I am out of there.
The blue dress I might want is relatively easy to find in the store, but nearly impossible on Nordstrom.com. It shouldn’t be so hard, which is why a new machine-learning commerce platform from SAP drew my attention.
Called SAP Upscale Commerce, it was launched late last year and was demonstrated at NRF’s Big Show. Afterwards I caught up with Lori Mitchell-Keller, co-president of SAP Industries, to talk through how it can solve the e-commerce shoppers’ needle-in-a-haystack problem when browsing online, but she also tells me it works to support sales in-store, too.
“Our application is surprisingly simple, but what it does is powerful,” Mitchell-Keller explains. “It uses machine learning to understand the customer and help build wardrobes around what the customer likes. It is like the ‘complete the look’ suggestions many websites offer [e.g. Nordstrom] but rather than having suggestions programmed by the buyers, it is based on the individual customer’s preferences.”
This application takes personalization to the next level to help the customer find what he or she is most likely to want when and where they are looking for it. Sales associates can tap the intelligence of the program too in order to deliver to the customer items that he or she might like.
“It is a deep learning system which automatically figures out which products to show to each visitor, while at the same time maximizing profit for the retailer,” Mitchell-Keller remarks. “This means that each visitor has a completely unique shopping experience tailored to them.”
Burberry and Ulta are early adopters of the new platform. Mitchell-Keller shares an example of how it works for Burberry.
“It knows that the customer doesn’t like peplum or double-breasted jackets, because she has tried on three of them and never bought, or only chooses pants to go with a jacket, never skirts,” Mitchell-Keller explains. “That kind of understanding of the customer is powerful and saves her time, especially considering the desire of customers to spend the least amount of time finding the right outfit.”
In addition to the customer-specific data, the program also trains itself to quickly spot products that are trending on social media, such as when a celebrity wears a particular dress or a duchess wears a trench coat, as Meghan did in her recent tour in Auckland, which Burberry promptly sold out of. “These hot products are automatically exposed more because they drive sales,” Mitchell-Keller says.
Of course, Mitchell-Keller stresses that the SAP Upscale Commerce application requires a base of experience with a shoppers’ history. The more the customer interacts with the system, either in-store or online or both, the better it will be at presenting the right options customized to his or her preferences.
Bridging the gap between online and in-store shopping
This application also works with the new way people like to shop, browsing in-store to touch, feel and try on selections, but then to make the actual purchase online.
“Shoppers might look around the store, make a mental note of what they like, then they may very well go online or call back the store to buy it. The system makes it very easy for the customer to shop how she wants, online or in-store,” Mitchell-Keller says.
This program would have been a great assist for her recently when she saw a Hugo Boss jacket in the store, then couldn’t find it online when she wanted to make the purchase. “I ended up searching for something else and only then did I find the jacket I wanted,” she shares, adding “There is a lot of improvement we can make across the board.”
Retailers are only beginning to put machine learning to work
Cosabella, the Italian lingerie brand, is one retailer Mitchell-Keller calls out as using machine learning to help customers find exactly what they are looking for quickly.
“Cosabella changes the website based upon the person who logs on. If I typically buy their lacy bras, the next time I log on, it takes me immediately to the lacy bra page instead of making me go through three or four pages to get to what I want,” she explains.
But neither Burberry or Cosabella are brands that customers are likely to build a whole wardrobe from. It is when brands like Nordstrom, Saks, Neiman Marcus, Macy’s or even Walmart and Amazon put this kind of machine learning to work that its real power will be realized by shoppers.
“Amazon knows a lot about me as a customer. It knows what I put in my cart and the wish lists I have there. But when I put something new in my cart, it suggests items that other customers might have bought together, but that’s not me,” she says.
“If Amazon was using machine learning, it would do a much better job of personalizing the experience for me. Rather it is approaching it from a ‘mass-customization’ point of view, but not getting smarter about me as an individual customer,” she says.
“There are so many applications of machine learning and pattern recognition that will help retailers deliver customer-specific experiences online and in-store,” Mitchell-Keller says.
“We are just on the cusp of helping retailers figure this all out. Not long ago when I talked machine learning to clients, it seemed like science fiction,” she continues. “But now they are latching onto the need to learn all about the customers whether online or in-store and adjust the interaction with the customers to take advantage of that.”