Today the world began to talk to us (read as call about rumors they heard) about Riya 2.0. (I know I have still have one more episode of my story of how our launch led us forward ... but I often get ahead of myself).
Matt Marshall called and asked about where we are going next. He understood it well when he said," Riya Super-Sizes Plan ".
As he wrote, this new shift was driven by usage data from our users. Laurie Sullivan wrote about it here on Techweb.
Of course the true reporter/blogger in this was Dan Farber who coaxed it out of me even before I knew what to call it. His article is here . I had forgotten this in my first post. Apologies Dan.
Michael at Techcrunch writes about it here
For a very detail analysis of how we got to the decision see Episode 5 of my Riya launch story. For context you can also read Episode 4 3 2 1
Where are we going? Riya 2.0 will be a web-wide (public) visual search engine that uses face and image similarity to search the web. We are calling this new kind of public web search: Visual Search. Why? Because you will be able to search by clicking on / submitting a photo instead of having to type in text.
We are 100% focused on building the core technology right now and starting to crawl the web. When will it be available? Not for a while. Probably three or more months. Why are we talking about it now? Our crawlers are going out there and I want webmasters to let them in...;-). What will happen to Riya 1.0 and the personal search based upon face recognition? I will remain something you can do once you register for our search engine and create an account. You will even be able to do face similarity on your own photos without making them public.
One clarification to Matt's article: We don't have 14 Ph.Ds from Stanford. We have 14 members of our Research team many of whom are from Stanford.
Sorry, Munjal, for misunderstanding. So you when you said 14 Phds, you meant only some of them were for Stanford? I think I've got it now.
Thanks, Matt
Posted by: Matt Marshall | June 15, 2006 at 09:15 PM
The challenge with Photo Search or "Visual Search" is that most *image* searches today (something like 9 out of 10) are for, how should I say, adult images. It will take a while for internet users to think of submitting images as "keywords" and match those images.
I think probably the most relevant use of Visual Search would be for dating sites.
John
Posted by: John Lin | June 16, 2006 at 09:56 AM
What is backup plan #3?
It's never easy...good luck guys. At least you've got the VCs money hoss!
No greater joy in this world than to have the luxury of figuring out a market on VCs dime!!!
Posted by: Jack | June 16, 2006 at 12:20 PM
Revenue from referral fees will hardly support all the engineers you got in India.
Has anyone on your team analyzed the revenue potential from this new Idea? Or is this something you threw on the wall and hoped it sticks?
Riya is really neat and cool in term of technology, but it doesn’t really solve any problem. What pain point are you really trying to solve? And are people willing to pay for it?
Referral fee based services are great for sites like Fat Wallet or other coupon websites, but for photo search? I don’t think this one will fly.
But cool technology... I think Riya is a head of it’s time. I don’t see you guys in business 2 years from now. You’ll be either be acquired or closed shop.
I hope Riya proves me wrong.
Good luck.
Posted by: Jamesm | June 17, 2006 at 06:24 AM
Does your web spider have a name? I want to give you access to my website (and my thousands of pictures), but I need to know the name so I can unblock you in my robots.txt.
Posted by: Andrew Hitchcock | June 17, 2006 at 04:02 PM
Andrew - we'll get you a name. Jamesm me too, Affiliate plus it's brother CPC are $8B dollars+ btw.
Posted by: Munjal Shah | June 18, 2006 at 10:55 PM
PLEASE PLEASE PLEASE break the bot trend and only crawl sites that give you permission.
I know it adds gobs of overhead and slows the aggregation but we (webmasters) are tired of all the bandwidth sucking resources wasting bots that continually attack our sites.
Without my bot blocker turned on my bandwidth usage jumps from 2gb/day to 25gb.
-JayW
Posted by: JayW | July 24, 2006 at 03:06 AM
"We are calling this new kind of public web search: Visual Search. Why? Because you will be able to search by clicking on / submitting a photo instead of having to type in text."
It is interesting that you mention the term as if it is brand spanking new, while in fact, this is an old concept that has been tried before (including the term Visual Search)as early as 1999.
Check out this website http://www.evisionglobal.com
As "Jack" pointed out, monetizing visual search has been and will be the biggest challenge. The technology was ahead of its time in 1999, probably so even now.
In any case, I wish you the very best in your pursuit.
cheers
Prasad
Posted by: Prasad | August 22, 2006 at 03:27 PM