I have been fortunate over the last few months to be able to discuss the present and future of the video search business with some of those executives best positioned to comment on this nascent industry.
The executives who were generous with their time are:
Tim Tuttle, founder and CEO of Truveo senior vice president, AOL.
Suranga Chandratillake, founder and CEO of Blinkx
Jeff Karnes, director of multimedia search at Yahoo!
Chase Norlin, founder & CEO of Pixsy.
Let me share with you their answers to some questions pertaining to video search and its likely future.
Heyman: How does video search differ from traditional web search?
Karnes: There is a big difference. Web pages have lots of data for text search. But there is not sufficient information for video search. The question is, "How do we get more meta-data so we can provide a more relevant user experience?"
Tuttle: There is a lack of published information about video. It is far more difficult than (text) search to find millions of videos. Standards are a mess. There are many different formats of embedded flash players. This will not be resolved soon.
Norlin: At Pixsy, we've realized that half our user activity is browsing, and therefore video search is more about discovery than it is necessarily about algorithmic search and relevance. Relevance is important, but not the Holy Grail.
Heyman: It sounds like there's a different paradigm at work in video search, in that in "enquiry search" users are often looking for purchase information while in video search, users are looking for entertainment. Is that correct?
Karnes: Video certainly has more of an entertainment usage for the mainstream audience. In web search, the users usually know what they are looking for. In video search, like Chase Norlin said, there is more an element of discovery. Yahoo is working with personalized viewing guides featuring channels that get the user started with the discovery process.
Norlin: In regard to "intent to purchase" vs. entertainment, obviously video search is geared towards the latter. But there are many other highly monetizable categories in video search that fall outside of entertainment, like automobiles, health, fitness, travel, et cetera. Ultimately, as I said, video search is really about discovery and keeping users engaged and informed, which of course is an effective tool for advertisers.
Heyman: How is video search likely to be monetized?
Norlin: Video search is currently being monetized through PPC text ads, graphical and in-text ads. Obviously, if the search engine is playing the video, then an opportunity exists for pre/mid/post roll ads. However, there are still a host of unresolved copyright issues associated with resyndication of video content and the area is still gray. The recent Perfect 10 vs. Google thumbnail ruling plays well for companies in the thumbnail category, like Pixsy, who exist to serve as massive thumbnail aggregators so that users can find what they're looking for and go directly to the source to watch it.
Tuttle: We are in an experimentation phase. As an industry, we are trying everything; pre-roll, post-roll, display alongside, overlay, hot spot. There is no favorite yet, but the 5-second spot is becoming an emerging standard.
Karnes: Monetization will evolve going forward. While ads seem most likely to be the prime means of monetization, it is likely there will be room for such models as direct-to-own, rental and subscription.
Chandratillake: Today, the most popular form of advertising is actually banners on the side of the video, which isn't video advertising at all! But, increasingly, we're seeing various forms of in-video advertising (pre, post, inter-roll and also banners that appear in the video stream itself). At Blinkx, because we're able to know what is going on in a specific video at a given point in time, we're working on a contextual video advertising platform. So you could be watching a video about luxury cars, and when the narrator mentions "Lexus," you'd see a small ad appear at the bottom of the screen that would take you to your local Lexus dealership's site. Later, when the narrator says "BMW," the ad will switch.
Heyman: Agencies and brands are posting videos on the web. How can they best optimize them for video search?
Tuttle: There are no technologies yet that do an effective job of automating analysis of the video stream and pulling out data. Broadcast video has closed-caption text data but most web video does not. So, therefore, in VSEO (Video Search Engine Optimization) you need to provide text info. Most effective is to provide RSS feeds.
Karnes: Invest in meta-data as you produce the content. Enter descriptions as you go. Embrace the social web by using tags and ratings.
Chandratillake: You can download our Video SEO whitepaper from the Blinkx site. Generally speaking, the big things are: a) think of what you do for image or text SEO and make sure you are doing the same for your video (you'd be amazed how many people don't!), and b) submit! It's important to ensure that the video search engines all know about your content.
Norlin: Video Search SEO is a natural outgrowth of the algorithmic video search business, very much like SEO was born from traditional algorithmic text search. Detailed metadata is key to getting indexed. Content providers should also take proactive steps to distributing their content by pushing out RSS feeds and contacting search companies to index their material.
Heyman: Web video has already had "Lonely Girl," but, as a medium, are we still waiting for the equivalent of TV's "I Love Lucy"?
Tuttle: I am not sure that a big hit or personality will catapult web video into the mainstream. There will certainly be web-only video success stories, but these will not necessarily be the only catalyst to trigger mainstream adoption. I think that new applications and websites might be a more likely trigger. That is my two cents.
Karnes: Technology is lowering the entry cost and is leveling the playing field when it comes to creating entertainment content. "The Blair Witch Project" innovatively used the internet and had great success but it couldn't be replicated. Creative is still an art form, even if distribution has become a science. The audience wins because there will be more and more content to choose from.
Chandratillake: There has been a lot of focus on user-generated content but ultimately that's a shorter term success; internet fame burns brightly but briefly. The content that has a lasting impact -- or stars that have that longevity -- will continue to come from professional sources. It's true that Hollywood occasionally overlooks talent, and that -- with such a low-cost barrier to entry and its viral qualities -- talented people can now use the internet as a way to build an audience anyway. But, ultimately, the Hollywood talent-machines have very deep pockets and they are experts: they do tend to find the best people and can give them a different vehicle for success. However, I do think what we will see is a lot of new forms and formats that fit the medium better, e.g. short-form content for watching on your PC, but also longer, relaxed-length form content for watching on your IPTV (no need to have everything conform to a 30-minute slot -- including ads -- once there's no such thing as the "schedule").
Norlin: In terms of web-based video, expect to see a universe of "prosumer" or "semi-pro" content creators to emerge. To date, content production has exclusively been the domain of major studios and production companies on the high end, and consumer content creators (e.g. YouTube) on the low end. I predict a rise of content creators in the mid-tier that create unique original content for web-based viewing. As content can now be syndicated to a wide range of devices and mediums, and as monetization of that content becomes more successful and sophisticated, there will be greater incentives for these mod-tier content creators. A mini-industry will emerge here.
Heyman: What's next for video search?
Tutttle: A chaotic arms race involving the video search companies.
Norlin: The next great wave in search is the trend toward personalization and automated content delivery. Ultimately, consumers will want to spend less time on the web and extract more value from it. The Pointcast model will be back in a big way. My Yahoo is an example of this.
Bob Heyman is chief search officer for Mediasmith -- a full service advertising media agency headquartered in San Francisco, California -- and is the co-author of Net Results.2 (New Riders) and the Auction App (McGraw-Hill). Read full bio.