Googles new taps into its search expertise

Google's new taps into its search expertise

Computer nerds and fashionistas have little in common - or so you might think. But internet giant Google believes it has come up with a winning formula to help people shop for fashion online.

Its team of computer scientists has painstakingly adapted machine learning and computer vision tools to create a new way to browse and shop for apparel and accessories over the internet.

Indeed, by recommending items to shoppers and advising them on how to put outfits together - essentially identifying and analysing the browser's taste - it hopes its new site will ultimately become the first port of call for people shopping for clothes.

Google's move takes it into the space already occupied by the likes of other fashion aggregators such as ShopStyle and Polyvore - and, most recently, UK-based online fashion retailer Asos, whose new Fashion Finder feature refers customers to brands that aren't available on its website. They don't sell the items themselves, but instead direct shoppers to a store that does, in return for a commission.

Crucially, Google's new direction also taps into its core strength, which is of course searching, and this is where it gets exciting. After all, most search engines tend to throw up things you don't really want, with a search for "red shoes" more likely to pull up everything "red" and everything "shoes".

Google's techie team initially worked with celebrities, stylists, designers and fashion bloggers to try to understand and categorise "style and taste". They looked at the fashionistas' favourite items, colours, patterns, brands and silhouettes - as well as those they hated.

"Machine learning algorithms" were then applied to extract usable information from this trend data, in essence setting up special rules and formulas to put together looks that work.

Browsing boutiques
Shoppers can create their own boutiques or browse through those set up by celebrities, stylists, designers and fashion bloggers including Carey Mulligan, Mary-Kate and Ashley Olsen, Kelly Osbourne, and Nicole Richie and filled with items they'd like to wear.

On a more personal level, users can search the site by size, silhouette, patterns and colours. There's also help in putting a complete look together. A search for yellow pumps, for instance, would throw up matching outfit ideas, as well as images from street-style and styling sites to provide inspiration.

It will also solve the dilemma of finding items that work together to create a look, by throwing up ideas that tap into style rules suggested by people in the fashion industry - like "heavily patterned handbags don't tend to go with heavily patterned dresses" or "don't mix stripes with patterns."

The idea is to make it easier to sift through the vast amount of choice faced by shoppers, both on the high street and online.

As Munjal Shah, a Google product management director, explains: "If there are, say, 500,000 items in a store, that means there are literally billions of different combinations of outfits you can make with those items. How do you sort through all of this?

"With fashion, reviews and specs are less important; fashion shopping is about discovering something that fits your taste and feels right," Shah adds. "The web works well for buying cameras and other hard goods but for soft goods, such as clothing and accessories, it's not the same as shopping in a store."

Online sales growth
Well now, of course, it's come one step closer to the real thing. Google's move is also fuelled by huge growth in online apparel sales, which have consistently bucked the fluctuating performance of the high street. Forrester Research, for instance, expects internet sales of apparel and accessories in the US to reach around $25bn this year.

And not surprisingly it comes hot on the heels of increased activity in the sector by giants like Amazon and eBay.

Amazon is relaunching its clothing and footwear business to focus on high-end styles, and has improved the search capability of its men's and women's shoe stores to allow customers to browse for shoes based on how they look - as well as price, brand and colour.

And online auction site eBay is also focusing on the shopping experience with its new stand-alone eBay Fashion that allows consumers to browse by boutique, brand, category, similar items and trends. New "image similarity technology" enables surfers to find items with similar colours, patterns and shapes from among the 19m-plus average daily listings.

The new site is only available in the US, and only for women's fashion. But it looks set to raise the bar for other firms to not only keep customers interested but also enable them to find what they're looking for. And ultimately, if its referrals lead to a sale, the commission generated will be a big boost for Google's bottom line.