Yahoo Takes Huge Step Towards Image Search Supremacy

Includes: AABA, GOOG
by: Thomas Hawk

Search Engine Journal reported Tuesday morning that Yahoo (YHOO) has added over 300 million indexed Flickr images into its image index. And so begins the single most important reason that Yahoo bought Flickr back in March of 2005 -- to use social search and rank to increase the relevancy of image search on Yahoo.

And with that, Yahoo just took a huge step in overtaking Google (NASDAQ:GOOG) in relevancy for the first time, even if in one small area.

I have been heavily blogging about this day for the past 2.5 years. Some of my past posts on this matter can be read here and here and here and here and well you get the idea.

The integration of Flickr images into Yahoo image search is only the first step though. It's being done cautiously. For instance: Do an image search on Yahoo Image Search for "San Francisco" and you will find that no Flickr images appear on the first page. But page forward and you will start seeing the images. Eventually you can expect much more forceful integration of Flickr's superior image search algorithm into all Yahoo Image Search main results.

Getting at what is inside of a photo is tremendously tricky. Google at present uses an image labeler game to try and get at this but this is not near as fun as Flickr and social photosharing -- or something like hot or not. But the fact of the matter is that social communities can be quite adept at determining what photos are better than others.

Where this sort of search technology will work best of course is at the long tail.

Take some of these examples for instance and you will see that with the Flickr integration, Yahoo mostly now produces superior results in image search than its competitors. It works especially well with long tail people search. All image searches were done with quotes.

Thomas Hawk: Yahoo Images, Google Image Search, Microsoft Images, Ask Images

Robert Scoble: Yahoo Images, Google Images, Microsoft Images, Ask Images

This is also technology that we've recently incorporated into Zooomr. Recently at Zooomr we developed our own algorithm using 28 different characteristics to determine rank and relevancy of photos. And even though today the Zooomr photo library is much smaller than Flickr, you can already see the success of letting social communities determine rank. Take this search on Zooomr for instance for sunflower, or dog, or cat, or beach. Now these are common terms of course. But as Zooomr's photo library keeps getting bigger and bigger, the image search associated with Zooomr will get better and better and reach further and further into the long tail. This has enormous implications for image search, but also for search for things like stock photography.

There is so much important work to do in the months ahead with image search, but the best image search going forward will be done around social networks. Yahoo was very smart to purchase Flickr back in 2005. It was the search team that bought Flickr, not the Yahoo photos team. Watch as this new integration evolves and gets better and better in the months ahead and watch as Yahoo leapfrogs its major competitors in image search.

Disclosure: I'm CEO of Zooomr.