Even six months ago, it was difficult to find any meaningful video content on the Web. So why even bother worrying about how to organize it and make it searchable?
But seemingly overnight, the planets aligned. Broadband Internet access became a mainstream reality. Personal IT hardware, including more powerful PCs, the ubiquitous iPod and a range of devices-from hardware-based players to even cellphones-gained the ability to display Web-delivered video programming. Cheap technology for creating digital content, from Web cams to affordable HDTV cameras, began to proliferate as well. All around the Web, media sites cranked up their video production, and perhaps more importantly-with CNN and AOL as recent examples-moved video from behind subscription walls and out into the ad-supported wide open.
The result is that not only is video content on the rise, but advertisers are seriously looking at that content-ranging from everything from CNN and ESPN video clips to from-out-of-nowhere viral videos (think Burger King's Subservient Chicken) to emerging "video blogs"-as viable inventory against which they can place lucrative online video ads.
And just as quickly as video has become viable on the Web, video search has risen as a rapidly evolving new technology tool and advertising environment. Just look at this timeline as an indication of the explosion of video search:
Search giant Google began testing video search in January and in April upped the ante significantly by initiating a program that lets anyone upload-and eventually charge for-video that it will host and make searchable. Google also cut deals with media partners such as CNN and Discovery Channel to provide access to their video content. Later, in June, Google augmented its video-search product with an open-source video player that lets users watch video from within the Google environment.
Yahoo's search product took a similar evolution, from lab test at the beginning of the year to broader availability improved functionality through the spring and summer. Yahoo also forwarded Media RSS, a new format for adding metadata to video content. Yahoo believes it has the right DNA to tie the Web to Hollywood and do it right with video search. "Given the leadership of our company [including CEO Terry Semel, former Warner Bros. head], we have a huge understanding of content owners and their issues and concerns," said Bradley Horowitz, director-media search at Yahoo.
Also in June, Apple and video-search provider Blinkx provided what it claimed was the first podcast (audio) and video blog search engine, providing a record of the growing amount of user-created multimedia content.
Not to be outdone, AOL in the last days of June delivered its own beefed-up video-search product. The offering includes "featured videos," a 15,000-segment database of licensed video content from AOL and Time Warner content divisions and other partners, as well as 1.5 million "Video Results from the Web," featuring results catalogued by SingingFish, the video-search engine AOL acquired in 2003. AOL's play also includes an embedded video player that lets users view search results immediately, supporting multiple video formats. "Where we are today, first and foremost, is that we've gotten in sync with the market. Six months ago we were not in the position we are today in supporting video advertising," said Kevin Conroy, AOL.
Finally, a slew of grass-roots, user-created content media platforms have emerged over the past few months, including first-out-of-the-gate OurMedia.Org plus a set of plays backed by Web 1.0 tech-celebrities including Open Media Network and the still-incubating 24 Hour Laundry from Netscape's Marc Andreessen and BrightCove from Web-tools pioneer Jeremy Allaire.
It's quite amazing to watch video search emerge so rapidly. And although many of the early video-search offerings don't even offer advertising programs yet, the explosion of text-based search engine advertising-Google's annual revenues are closing in on $4 billion and its stock recently bounced above $300-means it won't take long at all for video-search-advertising models to begin to emerge.
The challenge is not that there isn't a business model for video search - but that there are so many possibilities.
Early video-search players can certainly see taking advantage of straightforward plays such as displaying text-based keyword ads against video-search queries. Yet at the same time, they don't want to limit themselves when the opportunity beyond that is so massive-basically creating an uber ad-insertion-engine-in-the-sky that could target, insert, deliver and play back (via their own in-page media players) finely-tuned video advertising on the fly and in response to highly-targeted user requests.
MAKES AND MODELS
"The ad-insertion model sounds quite reasonable to me," said Yahoo's Horowitz, adding that any number of revenue models-including pay-per-view, ad-supported and subscription-based-could emerge to support online video. "If you look at [Yahoo contextual ad unit] Overture, there are some natural analogs to what they do with targeted, contextual advertising that could be applied to the video world. It's a straightforward extension."
With video-search advertising still in its very early days, "advertisers and publishers are going forward hand-in-hand" with video-search engines such as Yahoo to figure out the right business model, Horowitz said. Indeed, the most profound change from the broadcast world is that online video advertising has the potential to emerge "not as a punishment [in exchange for receiving free content] but as something of value in tune with the intention of [the users'] request. That's the holy grail," Horowitz said.
One of the biggest open questions is whether the search engines and portals will be able to build their own video players and aggregate video content on their own sites. That would create incredibly valuable destinations for delivering video advertising to massive numbers of users. "We think publishers benefit more by leaning into a more aggregated approach," said AOL's Conroy. "But we aren't eliminating any options," including helping to enable a more distributed world where video resides on publisher sites. "There will be a continuum of approaches and relationships," Conroy said.
Indeed, a distributed approach would more closely mirror how text-based search advertising evolved. Search engines today don't "host" content; they help users find it out on the Web and extract advertising payment along the way. "For online video to really be a mass medium, that content needs to be everywhere, not just at portal sites," said Jupiter Research analyst Nate Elliot.
A portal-based, content aggregation approach may have worked in the early days of the Web, when users weren't exactly sure how to navigate this new environment, said Jim Spanfeller, president-CEO, Forbes.com, which today is selling video inventory as quickly as it can create it. But today, said Spanfeller, "people have become accustomed to going to specific places [such as Forbes.com] to find what they are looking for, including video."
Beyond the business model questions, there are also fundamental technology questions to be answered about video search. And how those questions are answered will have a huge impact on the evolution of video search as a business as well.
For instance, what does it mean to search video? Should it mean searching against keyword and metadata descriptions of video content? In that scenario, if I'm looking for video of Tom Cruise's appearance on "The Oprah Winfrey Show," publishers will tag their videos with text-based metadata tags using terms like "Tom Cruise," "Oprah," "Katie Holmes and "couch-jumping" in order to direct me to the content I'm looking for. With tag- and metadata-based environments such as Flickr (for photos), Del.icio.us (for Internet bookmarks) and Technorati (for blog search) strongly emerging, perhaps this approach to video search is not only sufficient, but creates a peer-to-peer environment where millions of users improve video search every time they use it.
"We've seen with Flickr [the photo-sharing site Yahoo recently acquired] that through a little bit of social engineering you can have millions of people tagging and annotating and adding value around photos," said Yahoo's Horowitz. "Flickr-like concepts can easily be applied to other domains such as audio and video and the power of one hundred million people can be brought to bear."
Yet at the same time, other video search providers-some more publicly than others, but in reality all are pursuing this vision on some level-want to enable consumers to literally search the contents of video streams.
"If you look at the video search engines that are emerging, they don't actually don't ever watch, listen or understand rich media itself," said Suranga Chandratillake, founder and chief technology officer of Blinkx, which pioneered an approach to cataloguing the content of video streams themselves.
To enable this approach to video search, Blinkx has banks of computers playing back video streams-typically in two-times real speed or more-and smart software analyzing the videos based on things such as image and audio patterns. Search giants Google and Yahoo are testing similar approaches as well. Using in-stream video analysis, not only could you find that Tom Cruise video we mentioned earlier, but the search engine could direct you to the exact spot in the clip where Mr. Cruise happily shouted, "I'm in love!"
And just imagine the targeted rich-media advertising possible against that video-search query.