TechCrunch is reporting that Google is the late stages of
acquiring visual search engine Like.com. The company also owns several other
properties that utilize its image search technology, including Style By
Jacquie and the fashion profiling site Covet.com. Originally developed for
facial recognition purposes under the moniker Riya, the search technology has
also been applied to user generated photos, not just product/celeb images.
If this is indeed to be, there are certainly merchants out
there wondering what will happen to these existing sites and the
traffic/revenue that originates from them. Like.com is a significant AdWords
advertiser and has become a large referrer of traffic to some merchants in the
shoes, apparel and accessory categories. Like also acts as a publisher of
Shopping.com, Shopzilla, and PriceGrabber offer content, so merchants not
sending a direct feed into Like.com may still receiving benefit from their
properties.
It seems likely that Google's true target here is not any of
these sites but the underlying technology, which generally does a very good job
of returning similar product images by color, shape, pattern, etc. It doesn’t
seem likely that Google would maintain these sites long term in parallel to
Google Product Search. For one, it’s a conflict of interest since AdWords is a
main acquisition channel for Like.com, but also Google generally thinks way
bigger than that.
Integration of Like.com’s visual system into Google's
existing search experience, both for product search and otherwise, could result
in a very attractive option for users, especially in instances where words are
not nearly as representative of the query subject as an image (this is exactly
why Like.com focuses on clothes, shoes and accessories and not laptops or
TVs). Giving users not only more ways to search but options that allow
clearer communication of what is sought is definitely a win for Google. In
addition, expanding the already overhauled image search to build a stronger
response to Bing’s “Visual Search,” which is frankly more of a visual browse
experience, is probably seen internally as a plus.
If this sort of integration is the long term goal, the
product search integration is the logical first step since that is the
technology’s primary use right now. Such a feature could mean an increase in
stickiness for the already popular Google Product Search. However, there isn’t a
lot of certainty outside of Like.com as to how many of their users choose
visual search over the trusty old text box. Similarly, there is little clarity
as to what the conversion rate of users who choose visual search over text
based search looks like. When it comes to clothes and shoes, different brands
and models can look very similar, so using an image to find that exact item
seems a lot less effective than a text based search. Really only when users
don’t know exactly what they are looking for and want to be guided to
aesthetically similar items does an image search seem extremely valuable. That
sounds like a less qualified shopper. Then again, as long as Google Product
Search traffic remains free, does it really matter how qualified the user is?
There is also the possibility that this is intended to
augment the already live Google Goggles app on the android platform, which
allows for searches based on user photos. Though media products seem to already
work fine in Goggles, other types of products have not yet been a focus. Both
the technology and the existing product catalog could act as an accelerator to
expanding the scope of the Goggles application, acting, at least in the short
term, like a RedLaser for clothes, shoes and accessories.
Speculation aside, it seems safe to assume that Google wants
to use this software to expand what they are already good at: helping users
find things, and either serving ads during that experience, or driving Google
usage/loyalty so they can continue to serve even more ads during other points
in the Google experience.
