Online shoppers often fail to capture the object of their desires in words. And even when they can, the search results that come back often look nothing like what was intended.
EBay is on a mission to eliminate that disconnect and its latest effort comes in the form of a new image search feature. Revealed Wednesday, the update lets shoppers grab items in the marketplace and search for visually similar products by just dragging and dropping the photo into the search bar.
The company described it this way: “Let’s say you’re looking for a new chair, so you type ‘egg chair’ into the eBay search bar and you see a list of results for your search. If one chair in particular stands out, now you can search for more products that look like it — just drag and drop the image of the chair you like into the search bar to see visually similar results.”
Consider it the next step for the e-commerce company’s work in computer vision, a technology in which machines can read — and, more importantly, understand — the content of images. Known best for its online auction listings, eBay is no stranger to tech-forward approaches. It was quick to jump on Apple’s iPhone ecosystem, offering one of the very first offerings in Apple’s App Store in 2008. That long e-commerce history is a valuable asset now, with a library of 2.1 billion images giving the system plenty of data to inform its visual modeling.
“We use deep learning networks known as convolutional neural networks to process your images,” the company said. “Behind the scenes, when you submit your image in the search field, the neural network converts your image into a vector representation. Then, the vector representation of the image that you submit is compared against more than 1.1 billion live listings in eBay’s marketplace, using nearest neighbor search. Finally, we surface the best matched items, ranked by visual similarity.”
Indeed, it takes a lot of tech to deal with all those pics, which comes in different shapes, sizes and quality from a variety of sellers, ranging from individuals to established retailers. The system has to adjust for all the variables and zero out the noise, so it can find visually similar results for shoppers’ photo searches.
EBay is clearly banking on tech to drive business. This year, it unveiled augmented reality features that can visually assess shipping box sizes for sellers, as well as preference-based personalized storefronts for shoppers. But results have been mixed.
Despite showing growth in earnings and gross merchandise volume in the second quarter, the company still wound up disappointing investors. The issue was flagging customer engagement. After adjusting for new users stemming from its Giosis acquisition in Japan, active buyer growth clocked in at 3 percent, compared with 5 percent growth in quarter a year earlier. The concern is that the deceleration speaks to eBay’s inability to attract and retain new customers.
Its focus on artificial intelligence and visual shopping may be a gambit to reverse its fortune, but it’s also a natural next step for a platform bursting with product photos. The new feature builds on moves such as last year’s Find It On eBay update, which can locate products in its marketplace using visuals shared on social media, and Image Search, a tool that lets users hunt for products by capturing and uploading pics. Here, too, AI powers the drag-and-drop feature, processing the images behind the scenes.
The update goes live on Android and iOS next month for the U.S., the U.K., Germany and Australia. Beyond that, eBay teases that it’s not done tinkering with visual shopping yet, saying it will “continue to evolve it and add more ways to search with images soon.”
A picture may be worth a thousand words, but eBay is hoping it will be worth thousands of new users, at least. Certainly it can’t afford to see growth drag. Or drop.