[Disclosure: I am a founder of Webshots and currently employed by CNET Networks]
Thomas Hawk’s recent musings on image search got my brain spinning. My thoughts (forgive the haphazard sequence):
Webshots has been terrified of Microsoft since we were a three person outfit in 1995 providing content and “photo sharing” with a windows screensaver. To us it seemed like at some point, Redmond would start including photos with the OS. It took many years for that to happen and it never happened in the doomsday scenario we imagined–an integration with Gates’ own Corbis. How about a library of Corbis images for MCE!
We are witnessing the democratization of photography and the resulting downward pressure on the “value” of images as they become commodities. The days of Life magazine are long over and the long tail has arrived. The flickr community is leading this charge within the segment of aesthetically sensitive afficionados with an open api strategy, great feedback mechanisms, and boatloads of pr.
Webshots has been an arena where this has been happening for 10 years and while we have enough users to roundly fill out the artsy bucket, we also know a little something about the mainstream having delivered literally BILLIONS of full screen photos to pc desktops. The most startling discovery: notions of aesthetics, popularity, and relevancy fall all over the map. I might have had an inkling back in 1998 when we were getting suggestions for new hand picked photos to distribute/share.
“I want to see pictures of tractors!” came in one day.
Amusing at the time, but now Webshots image search returns more than 26,000 photos of tractors contributed by members. How’s that for a long tail!
When our search engine was still small by comparison, we started to catch a glimpse of how a photo-sharing community could actually compete heads-up with the large portals. The key was metadata: we had long enforced an ontology as well as encouraged album keywords, captions, and titles. Webshots image search is weak in some areas: notably porn (we screen all of our photos), high-profile entertainment, and news. But these are the same categories that present some peril for the portals.
Our search index now has more than 170 million member photos available and we are working toward updating the index daily. Google has caught some heat for its own infrequency. Until recently, image search in a portal world has been all about the orphaned product manager. Neither Google nor Yahoo! has managed to devise a way to fully index the wealth of data that Webshots churns out every day and the most logical explanation for this and the general lack of innovation is the lack of incentive. Advertising is noticeably absent from portal based image search because there are unresolved copyright issues.
The tide is turning because Flickr has brought attention to public sharing and “tagging” is the metadata implementation du jour. Two questions that are ripe for discussion and debate:
Who are the players really poised to improve image search?
And what constitutes an improvement?
Webshots is a media property with broad licensing for the content it stores and as such has strong incentives to make search a core technology because it drives monetizable usage. RSS feeds of any search query are on the near term product road map which could be a big deal for the Web2.0 world because of the wealth of content that would be exposed and frequently updated.
The combination of Flickr and Yahoo! My Web2.0 starts to take a stab at context (another elusive element to relevancy) and that represents a potentially large leap forward. Webshots has an enormous social network but you still can’t search exclusively your photos or those of your favorite members.
There is an emerging piece to this puzzle that is critical: the trusted aggregator. Photo sharing in most services (Webshots is an exception) is publisher centric. Webshots derives great value from its broad browsing audience. There is huge potential (and we catch glimpses of this with the application of tagging) is creating systems and rewards that allow individuals to do for photos what bloggers have done for news. Aggregate and redistribute.
There is also the broader challenge of uniting disparate communities and services 😉