Last week, Google entered the fashion world with their announcement of Boutiques.com, a fashion e-commerce site. Recognizing that the Web is not as suited to finding and buying soft goods like clothing and accessories as it is to hard goods, Google created a fashion search experience more like a brick-and-mortar store, complete with high-definition images of products, linking consumers to other sites where they can purchase the merchandise.
Boutique addresses the significant problems of volume and irrelevancy of search results plaguing other e-commerce fashion sites by limiting the millions of choices available to consumers through precisely identifying the user’s style. Users can take a personal fashion quiz to find his/her style or a user can identify with a particular style by choosing from a collection of hundreds of boutiques, created by anyone from celebrities to ordinary people. Within a boutique, a user can shop a “curated” selection of goods. In addition to the pieces added by the boutique’s curator, Google generates additional fashion choices inspired by the boutique’s content and by the pieces the user has clicked on. Finally, the user may simply search for an item using familiar fashion terminology and refine the results based on many relevant categories.
How does Google accomplish this intricate search? Google employed a multi-disciplinary team including both fashionistas and computer scientists, to infuse fashion sense into what Google does best: searching. The fashion experts identified factors important to a fashion decision such as color, silhouette, and length, and generated hundreds of words to describe these factors. Then, the engineers taught the computer to recognize these factors and words using a method called visual-search technology. Specifically, the engineers coached the algorithm to identify a particular color, for example, by categorizing thousands of photographs as either of that color or not. This visual method significantly improves upon key-word searches typical of other fashion websites, because the product website need not actually describe the color of a product in words for Boutique to recognize the item as that color and include it in search results. In this way, Google’s algorithm uniquely cuts through the clutter of clothing choices on the Web much like a shopper browses in person at a store, potentially changing the way consumers online shop.
The visual search technology at the heart of Boutique.com was developed by a company called Like.com, helmed by Munjal Shah. Google acquired this company and the intellectual property rights to their algorithm for a reported $100 million, as reported by the New York Times, and retained Shah to lead the Boutique project. This acquisition is part of a long line of acquisitions by Google, aimed at developing their intellectual property.