To Watch All blinkx's Video Would Take 18 Million Hours
blinkx, founded in 2004 by Suranga Chandratillake, has cataloged - and thus made searchable - more than 18 million hours of online video, perhaps the most of anyone.
The blinkx video search engine is based on technology conceived and nurtured at Cambridge University over some dozen years. Besides making all that video available at blinkx.com, the company’s technology is used to power video search at Ask.com and others.
Chandratillake recently spoke with IBD about video trends.
IBD: How much video is on the Internet today?
Chandratillake: Nobody knows for sure, but the growth rate is certainly explosive. We search the Web for video, using speech recognition and visual analysis, and then index the data. At the end of 2006, we had indexed around 3 to 4 million hours of video. By April, that had grown to 7 million hours, and we’re now at 18 million. We believe we have the single largest index of all video on the Web, but even we don’t think we have all of it. It could only be half of what’s out there.
IBD: Does your video search differ from other such products?
Chandratillake: People have typically applied traditional text search when looking for video, but that is only reading words tagged around the video, like the title and summary. That’s like judging a book by its cover.
blinkx is different because we actually open the book. We watch the video and listen to it, using speech recognition software to analyze the words being spoken and visual analyses to understand what’s happening within the video frame by frame. We pull out a lot more information on what the video is really all about. So when a user comes along to blinkx, they can search for any topic they want and we can find very precisely and objectively videos that match that.
IBD: Is your technology used by other Web sites?
Chandratillake: Yes. Some of our big-name users are Ask.com, Real Network’s RealPlayer, Lycos and InfoSpace. A number of these very large Web sites have video search that’s powered by Blinkx. Those sites are able to monetize that service by placing advertising in and around the videos, and they share that revenue with blinkx.
IBD: What is the ideal ad platform for video?
Chandratillake: We have an agnostic viewpoint on that. We have an ad platform that’s called AdHoc that can match a video to ads that are about the same topic. When you can target people that way, the advertisers are happy to pay a higher rate because they know their ads are being played in a relevant context. We can work on multiple ad-delivery platforms.
IBD: Do you agree with eMarketer, which says online video ad revenue will grow from $775 million this year to $4.3 billion in 2011?
Chandratillake: That is one of the more conservative forecasts. Forrester Research recently put out numbers that are higher than that.
IBD: How many video Web sites are there?
Chandratillake: The last time we checked, we knew of at least 400 such sites. Some of them are huge, like YouTube, Crackle and Veoh, but many are very much smaller and it’s questionable if they can continue to exist. Many of them have a lot of overlap of content and users, and over time I wouldn’t be surprised if some of them merge. Advertisers will very happily pay high rates for professional content, but not for user-generated content.
IBD: So, is the market overcrowded across the board?
Chandratillake: In some areas, yes, but not overall. In professional media, it’s still pretty sparse. There are a number of big shows you can watch online, but the reality is you have better programming on the TV than online. But the big networks are starting to put more things online, and more is coming.
IBD: What influence has the flood of user-generated content had?
Chandratillake: The costs of producing and distributing media have been so high traditionally that only a very small number could take part in that process. And only a tiny minority has ever made it big in Hollywood. The Internet has changed that market entirely because you can now very cheaply produce and distribute your own content. Content creation had always been a stacked market controlled by a small number of companies or people. The Internet is changing the perception of what media is good and what’s not good.
IBD: Is the Internet, technologywise, in good enough shape to handle this explosion in video?
Chandratillake: It’s not currently capable of delivering the video that will be coming five years down the road, but the Internet is not a fixed thing. It’s always growing and expanding. It’s capable of sustaining the experience we’re talking about.
IBD: What about the emergence of high-definition video on the Web?
Chandratillake: I think that will happen in the next year or two.