Amazon is testing out a shopping site for consumers who don't know specifically what they want but are willing to take some automated recommendations to help them find it.
The new service, Scout, asks shoppers to like or dislike a product (thumbs up or thumbs down) and responds by showing other products based on their choices. For example, if you're looking for a coffee table and you give a thumbs up to a rectangular table with a reddish wooden top and a thumbs down to a circular table, you may be shown a weathered gray table with a shape similar to the one you liked.
Amazon is using machine learning technology to address one of the perpetual criticisms of the site — that it's a great place to buy things but a lousy place to browse. While Amazon is by far the largest U.S. e-commerce company, it's left the door open for e-retailers like Stitch Fix and Bonobos to provide a more personalized experience and given Instagram and Pinterest more room to use their vast amounts of data in turning their social networks into fledgling commerce sites.
Scout is currently available for home furniture, kitchen and dining products, women's shoes, home decor, patio furniture, lighting and bedding. The site says that more categories are coming soon for things like clothing and handbags.
"This is a new way to shop, allowing customers to browse millions of items and quickly refine the selection based solely on visual attributes," an Amazon spokesperson said in an emailed statement. "Amazon uses imagery from across its robust selection to extract thousands of visual attributes for showing customers a variety of items so they can select their preferences as they go."
The news sent shares of online furniture sales site Wayfair down 4 percent to $136.93 in midday trading.
Amazon hasn't started promoting the service yet, though some shoppers may find themselves stumbling upon it in various categories. Under select listings, they may see a gray box that says "SCOUT | Style explorer," which takes them to the site. Or if they're on a general Amazon page looking for light fixtures, the link for Scout is on the bar at the top, in between "Shop by Room" and "Shop by Style."
When you arrive at the page, a blue box says, "start liking now."
It's the latest step in Amazon's effort to move consumers away from a world of traditional search while collecting more data on their habits and preferences. The company's Echo home assistant, powered by Alexa, lets people shop by voice, looking for discounts or buying things again from a prior order.
With Scout, Amazon is trying to "free people from the need to use words to describe what they're looking for when shopping in highly visual categories," a spokesperson said.
Customers can also filter by prime-eligibility, customer rating and price.
Here's the complete statement Amazon sent to CNBC:
This is a new way to shop, allowing customers to browse millions of items and quickly refine the selection based solely on visual attributes. It is perfect for shoppers who face two common dilemmas: "I don't know what I want, but I'll know it when I see it" and "I know what I want, but I don't know what it's called." Customers simply like an image (by clicking or tapping the thumbs up) or dislike an image (by clicking or tapping the thumbs down) to get tailored recommendations in real time and find the products that best suit their needs. Amazon uses imagery from across its robust selection to extract thousands of visual attributes for showing customers a variety of items so they can select their preferences as they go. This innovative shopping experience is powered by Machine Learning. The result is a beautiful and inspirational image feed, which gives customers the ability to explore a wide range of products in a playful and personalized manner with just a few clicks. It is currently being tested on Amazon.com and in the Amazon App for home furnishings and women's shoes. To learn more, please visit www.amazon.com/scout.
Correction: A previous version of this story mistakenly said that users can't filter by price.