Have you ever seen an image of a celebrity wearing something you’d like to purchase for yourself? Or spotted a cool accessory in a stock photo that you wouldn’t mind owning? Google’s Similar Items system is designed to make that acquisition much easier. It uses machine vision technology to identify products within photos and then link to them on retailer websites.
While it’s certainly true that Google’s fancy image analytics algorithms are doing some of the legwork, Search Engine Land points out that Google also asks web developers to utilize metadata to help make the feature as fluid as it is. Without that, products may not be eligible for sale through the Similar Items system.
Google has reportedly been encouraging the use of the metadata tags since December, though that could simply be because it wants to seed the web with compatible sites and products before sending the feature live.
The feature doesn’t cover every kind of product out there just yet. As it stands, Google’s algorithm is able to pick up specific brands of sunglasses, handbags, and shoes, with the search giant promising that “other apparel and home and garden categories,” will be coming in the next few months.
Google claims that some of the most common searches alongside brand-name products are pricing and availability, so creating a shortcut to a shopping experience is likely to be a profitable move for the company.
