Blinkx Reinvents Itself Again, Adapting to the Future of Video

I’ve rarely run into a company as adaptable, even Protean, as Blinkx. I’ve been following it since it spun off from Autonomy in 2004 with the mission of finding applications for Autonomy’s search and indexing technology. The first area that founder and CEO Suranga Chandratillake explored was desktop search for Macs and PCs, but within a year the company pivoted into video search. Since then it’s reinvented its approach to the video market every couple of years.

In its early incarnations Blinkx was a video search portal—it indexed video across the Web and helped visitors to find what they were looking for, monetizing the operation by showing its own pre-roll or post-roll ads. Later it de-emphasized its own site and became a provider of video search services to bigger content aggregators such as, Real Networks, and AOL.

But the company wasn’t entirely content to operate in the shadow of its big customers. In 2011 the company acquired its own network of publishing sites by buying Massachusetts-based Burst Media. And today, in response to the rise of what might be called “social content discovery,” it unveiled the beta version of its redesigned website,, with the aim of turning it into the company’s flagship product.

Blinkx's website, before its redesign (above) and after (below)

Chandratillake stepped aside as CEO earlier this year to focus on product development—his new title is chief strategy officer, with former chief operating officer Brian Mukherjee now holding the CEO reins. He says his team has redesigned the site “from the ground up” to encourage browsing as an alternative to traditional search. “Yes, there are times when you know exactly what you want, and we still offer that of course,” he says. But now the bulk of the site consists mainly of an endlessly scrolling wall of popular videos. “You can keep scrolling until you find something interesting. The whole concept of this never-ending stream is that if you are the typical ADD [Attention Deficit Disorder] user, you can flick around until you find what you are interested in.”

Another key change is social integration: if you connect Blinkx to your Facebook and Twitter accounts, the site will show you videos being shared by your Facebook friends and the people you follow on Twitter, or by people whose interests match yours. “You might find that Twitter is the most accurate view onto what you are interested in, while Facebook is more funny videos that your friends like. Either way, we have a pretty good chance of finding stuff you actually want to watch.”

Finally, the site has been redesigned to work equally well in desktop, tablet, or smartphone browsers. “When we last redesigned the site, 99.9 percent of people came in through desktops and laptops,” Chandratillake says. “But increasingly they’re using iPads and Android phones and even connected TV boxes. So we wanted to make sure that what we build was going to work on all these different screens, and do so seamlessly, without feeling like it was designed for one and rehashed for the others.”

In many ways, Blinkx’s redesign reflects the big, ongoing shift away from purely algorithmic search toward socially filtered or curated discovery experiences; call it the Pinterestization of everything. That’s why elements of Blinkx’s new setup resemble video discovery offerings from youngers startups like Remixation (maker of Showyou), Dijit Media (NextGuide), Matcha, BuddyTV, and Clicker.

But Chandratillake says Blinkx’s core expertise in understanding unstructured data—including the actual content of videos, not just metadata like titles and tags—still gives it a leg up over other discovery services that are based purely on social filtering. For Blinkx, he says, “those social influences are merely factors, more inputs into the algorithm, no more or less. They are valuable inputs that we didn’t have five or six years ago, but you need a level of filtration above that. Twitter is great, but there is too much noise.”

In other words, Chandratillake says you can think of the new Blinkx site as “algorithmically based concept of what you are interested in. It’s still your stream—everything that the people you follow have shared is still there—but the order is based on our algorithms.”

Blinkx plans to monetize the videos on its site in the same way it does on its other properties—by showing pre-roll and post-roll ads. But that will come later, after the “beta” label comes off the site, probably in October. Eventually, Blinkx may offer its new socially-filtered video discovery tools to its partners, starting with the Burst Media properties.

After Chandratillake gave me a tour of the new site, I asked him to explain the thinking behind the redesign of the site, and for his opinions about some of the big trends in online video. Here’s an edited version of our conversation.

Xconomy: To some extent, what’s going on with apps like Showyou and NextGuide and the new Blinkx is a shifting of some of the burden of discovering video back to the user. Using Blinkx now is more akin to channel surfing than traditional search.

Suranga Chandratillake: Yes and no. The burden of discovery isn’t being shifted back to users per se, it’s being shifted increasingly to the community at large. You are making the final choice; the difference now is that the options are driven heavily by your network of friends and connections. And of course search is still there. If you know you want to watch a specific video and hunt it down, Blinkx is still an excellent tool for that.

The social stuff seems like a really easy fix for how you figure out what to show someone, but actually it’s not. What you need is an algorithm or a filtration level above that, that helps people manage the chaos.

X: What signals are your algorithms paying attention to?

SC: Using the social graph data we get through Facebook and Twitter, we know who you interact with most, who appears to be high-value in your stream. Through that we can extend the graph and see videos that may be being tweeted by people who are similar to the people you follow. Thirdly, it’s about history and what you click.

X: How does Blinkx’s core search expertise still show through in all this?

SC: There are two pieces I think we do better. One is profiling. You’ve always been able to train Blinkx over time, but doing that well is really hard. Secondly, we have more data on any given video object because we analyze the video more deeply. Those two things are the key reasons we’ve been successful so far. All we are doing now is opening up to more powerful weighting sources, which are the social networks. The tough thing is using them without letting them over-weight what you are doing. Even if something doesn’t completely fit your social stream, we might still think it’s relevant to you.

X: Do you think traditional video search is becoming less relevant?

SC: The number of searches we handle has always been increasing. I don’t feel that it has become less relevant. But what we have noticed is that as people have become more confident about their consumption of online video, they have spent more time wandering around. The first version of Blinkx was just a search box, but as early as 2006 we had a video wall. For me, the best analogy is to the generic news sites like Yahoo News and Google News. They have a real balance between discovery and search. I can go to Google News to hunt down a press release about a company I care about, but I can also see the big stories of the day.

X: Thew new version of our site seems a lot more friendly to people who are in “lean-back” video consumption mode, rather than “lean-forward.”

SC: Weirdly enough people were already using us in a very lean-back way, but I don’t know that we’ve ever really helped them do that. This allows people to do that very easily if that’s what they want to do. They can start playing a video, then look for the next thing to watch and the next. And it works particularly well on mobile devices, which feel even more lean-back than the computer does.

X: Why did you give up the CEO role for the chief strategy officer role?

SC: I really enjoy building companies. I get as much excitement out of things like hiring my 300th employee or hitting profitability as I do out of the pride of shipping something for the first time. But there is a point at which you are doing so many of the incremental business things that you end up focusing just on that. There was a big part of my job that was not disruptive, that was just continuing or refining or growing. That is the part that I found harder to get excited about. At one point I thought that meant I would have to leave the company completely. I was quite stark in my thinking. But I talked to my board and they were very supportive and they all said, “Everyone goes through these phases. Make sure leaving is what you actually want.” And I thought hard and realized that I love what we’re doing, and that there will be a lot more change in online video. YouTube has been the one really big story in our space, and there will be at least two or three more before it’s all up. So I’m still just as excited about what we’re doing, but I wanted to focus differently.

X: Yes, speaking of really big stories in video, I’m surprised that the next one hasn’t started to unfold yet. It seems like we’ve been on the cusp of another big change for a year or two, something involving a major integration of online video and linear TV, and of mobile devices and traditional big screens. I would have expected Apple to jump into this in a big way, beyond Apple TV, but they haven’t. What do you think is going on?

SC: I have two totally different answers to that. The first is that from a technology perspective, there is a lot of stuff that needs to be hooked up that hasn’t been hooked up properly yet. For the new version of TV to really work, the first thing that is needed is for your TV to be much more intelligent than it is today. Sure, most TVs now ship with an Internet connection, but we know it’s not really there. Whether it’s an interface design question, or the cable companies strangling the economics of it, the point is it’s not realy working. Everybody is waiting for that to happen. Whether that happens in the form of a product like the iPod that delights the market so much that everybody follows behind, giving Apple the leverage they need to change the economics, or whether it’s something else, I don’t know.

The other answer, which is quite different, is that maybe [this transition] has already happened and we haven’t realized it. We spend a lot of time building an informal network of users that we profile, and we connect with lots of people’s children and nephews and nieces and siblings to try to get an unscientific picture of how people under 16 are using video. And one thing that fascinates me is that a massive number of them don’t watch regular TV anymore. They seem to spend all their time on YouTube, and they watch content that is non-professional, from my perspective. They are not watching anything that you and I might consider to be TV. I can’t tell whether that is an age thing and once they hit 20 or 30 they will become mainstream, or whether this is going to stick, and stuff that is relevant to us will stop being relevant to huge chunks of society.

X: On the first answer: What’s the holdup to making TVs smarter? We have the technology. Why can’t somebody like Apple just build a giant iPad and stick it on a wall?

SC: The second you stick it on the wall, people expect it to do what a TV does, and right now, because of all these commercial reasons, it can’t do what TV does. That means you will immediately disappoint the user. Apple can’t build this until the people who own and control the content give easier access to it. These things didn’t matter with the other devices: with the iPod you already had most of your music ripped on your hard drive, and with the iPad it was a whole new device so there were no expectations for it. I think with TV, we have so many expectations built in. It’s like electric cars. They’ve been around for a long time, but you can’t sell them until they can at least do everything that a gas-powered car does.

The good news for us is that there are lots of smaller shifts that will occur. I think that when we set out to build our new site, the two big ideas that Matt Scheybeler and I talked about a lot were discovery versus search, and online video as your companion to consumption of normal TV. We are both big fans of the idea of companion apps, and we don’t think that any of the current companion apps are that exciting. We ended up going away from that, because a lot of what companion apps need is not video—they need to be social and textual—and we are not experts in that area. But it’s a very interesting area, and maybe we will look at it later as a company. It’s a big disruption on the way. The reason TV is so exciting to advertisers is because it has these massive, engaged audiences. But if you can create a massive, runaway success in the companion area, it’s going to steal the commercial value of television. It’s not in our sweet spot, so we decided to do something we know well first, but it’s a great area.

Wade Roush is a freelance science and technology journalist and the producer and host of the podcast Soonish. Follow @soonishpodcast

Trending on Xconomy