Online fashion retailer Zalando announces photo recognition to let users find similar products on their platform. Here is how it works.
The idea came up in the company’s hackathon that took place last year. To turn the idea into a reality as fast as possible, Zalando uses “findSimilar” by Cortexica. According to Coretexica’s CEO Iain McCready, Zalando is the first European retailer using this technology.
Zalando understands that online shopping no longer takes place exclusively at home (or in the office), but increasingly moves to where trends are directly visible: outside, where people meet.
Daniel Schneider, Head of Onsite Customer Journey at Zalando, explains the photo search scenario: “When [our customers] see a style they like on friends or on the high street, they can directly choose from corresponding products in our shop and buy them without a lengthy search.”
Since the feature is already live, VentureVillage tested the app in the office and here is how it works:
You open the Zalando iOS app (the Android update will come soon) and go to search. Below the barcode scanner, you will find “Photo Search”. Take a picture, pick a clothing category, find the 30 matching products.
Three pieces were tested: A blue-white striped long sleeve, a short cut light pink shirt, and a dress with zebras and zebra stripes.
For the blue-white striped shirt, which was easily recognized, similar options were shown. To the second photo, similar shirt cuts were suggestions, and besides other colors, shirts with prints were shown.
To the last and most difficult one, the dress with little zebras and zebra stripes on it: The app recognized it was a pattern, and suggested other patterned dresses. None, however, fitted a similar taste (or other animal prints).
Even though not all suggestions were the most fitting, the idea of the feature caught people’s attention.
The photo recognition feature is currently in beta testing and is only available in Germany.
“We want to understand if and how our customers use the feature. Are they looking for colors or patterns, or do they want to see items similar to a certain street style,” a spokesperson at Zalando, wonders. “How will they get inspired – by a pattern of a carpet or curtains? We don’t know yet.”
“left”]Tests will also determine the impact to local retailers. The idea of customers taking pictures of clothes and shopping for cheaper or similar ones is not a new idea.
Startups like HeyShops are working on a similar approach. The photo taken gets sent to all local retailers who joined the platform, will get reviewed by the staff and a specific offer can be sent to the customer.
Different to Zalando, the platform is smaller but the offers are not done by an algorithm but through stylists.
However, the collected data could improve user’s search and increase customized offers. It will be interesting to see how other large retailers (on- and offline) that are not listed on Zalando or similar platforms will respond to this technology.