Google’s Boutiques places fashion in the cross-hairs of search engine technology
Created on Thursday, 14 April 2011 19:34 Published Date Hits: 291
Google launched, in what appears to be a break with its automated mass-indexing approach to the web, a new website aimed exclusively at fashion and the shopping online for apparel experience. While the launch of such a personalized website appears to be an aberration for Google who have traditionally looked at the globe as their playground and the world’s population as their targets it actually is not.
Google has always refined its search towards two guiding points: segmentation and personalization within the wider matrix of a global presence.
Boutiques (www.boutiques.com) takes the company’s philosophy regarding search into the next natural step. In creating boutiques Google has taken its search from global text-driven results which in search are an iffy proposition to visual, drilled-down ones. But that is not all it has done.
Branding as a means of focus
Google has, increasingly, being biased towards brands as focal points of online authority. Nike, sneakers, the company is always going to be thought of as a more authoritative source of Nike news than any retailer, no matter how successful, for instance. In Boutiques Google allows the manufacturers to register their presence ensuring that brands get a preferential status.
Brand new search engine mixes algorithm with the personal touch
To create its unique website which Google hopes it will become the focus of fashionistas everywhere, Google hired a designer to help categorize fashion items and code a search engine that can now see colour and it involved 100 fashion students to help tag items.
The result is that Google has used automation and personalization to create an algorithm which can now help you find a ‘look’ make automatic clothes choices suggestions for you and enables the fashion-conscious be erm, even more fashion conscious than usual.
Wetware and software
Google’s combination of wetware (i.e. flesh & blood designers) and software has taken search engine technology into an entirely new level. Just employing brains to do the hard work of breaking things down into rules of fashion and tagging items wasn’t enough. What Google needed to pull this off was to create a search engine that can actually see colour.
Theoretically that is not difficult to do. All you need is to look at a picture, analyse its RGB values and you suddenly have not just a good idea of the colour on a picture (which for humans can be subjective) but an exact match. Except that pictures are notoriously difficult to analyse accurately for colour. The model’s hair on a garment, the tone of her skin or a brightly coloured accessory are sufficient to alter RGB values and render a colour match useless.
To deal with this Google did what it does best: it pioneered new software. "So we created special computer vision technology just tailored for fashion that looks inside the photograph and really cuts out the item from its background," says Google’s Director of Product Management, Munjal Shah explaining this at an interview. "Nowhere on the Web has someone built an algorithm to do this," he continued.
Fashion is big
It’s kinds funny to talk about fashion when the global economy is still in a recession that Google would choose this moment to roll out a fashion search engine but the timing could not be better. Boutiques uses a unique combination of human expertise and a rules-based system to help users find the perfect look for them. But as always with rules-based systems the results are always less than perfect.
In true Google fashion which does not have a definitive beta-testing process they have brought out a site which is lightyears ahead of everyone else and they will now have to work really hard to refine it as it goes along.
With the global market picking up around 2012 Google’s timing could not have been more perfect.
Update: Following the success of Google+ - Google's social network, Boutiques is one of the victims of a rationalisation of Google's projects and its focus on socialising the web.
