Jennifer Zaino
SemanticWeb.com Contributor

EveryZing, a universal search and video search optimization vendor, will soon turn on the first of its publisher sites to take advantage of integrating its MetaPlayer video player with the YuMe ad management platform and advertising network for the online video industry.

With that ad-serving technology embedded in its video player -- which takes the rich metadata EveryZing creates and makes it part of the user experience to enable searching within videos -- publishers increase their opportunities to maximize their inventory revenue through the addition of YuMe's advanced targeting capabilities, innovative video ad formats and ad sales network, EveryZing says.

"This is the logical next step in us providing not only a comprehensive solution for driving more consumption and engagement of rich media for our customers, but providing them the capability to monetize it," says EveryZing CEO Tom Wilde. "We wanted to pick one technology to do that ad serving but not give up the opportunity to connect other ad networks to our publishers. YuMe represented that nice mix and our publishers like that for driving more revenue through their business."

With YuMe inside the video player reaching out to its advertising networks, a visitor stopping by a site that features how-to videos on home repair who searched on the keyword "pipes" may wind up being served an ad from a hardware and home improvement chain, for example.

One important feature that YuMe provides is a rich rules set that lets advertisers and publishers figure out how to best serve a particular ad -- for example, a publisher may want to show a 15 second pre-roll ad at the beginning of a video and then mid-way through launch another one. Or perhaps an advertiser wants to show a particular ad to a user only once during the course of a day, even if that user combines five different videos at a particular publisher's site.

Wilde also is looking to the future and the next opportunities for YuMe and EveryZing to use the rich metadata the service produces using its speech to text capabilities and natural language processing to better target ads.

"Now ads are targeted with a fairly blunt instrument," he says, such as high-level demographic information. But more finite targeting should improve response rates. So, if you knew more about what is in the video you could show a more relevant ad -- not just, for example, a Home Depot ad when someone clicks on a how-to install a faucet video at a do-it-yourself site like BobVila.com, but a Home Depot ad specifically about a promotion in its plumbing department. "That's not possible today now -- contextual information has not been made part of the advertising experience for video."

It's been held back so far by a couple of things, including the fact that very few if any services can produce the quality of metadata that is required to finely target video ads. "We're unique in that way. Our patented ability to create metadata drives that opportunity," Wilde says.

But the other constraint has been the chicken and egg problem -- advertisers haven't been producing ads that are so finely targetable yet, because no one has been selling them a place to deliver them. "When that becomes available advertisers will increasingly tone their ads that way as they have for other [non-video] parts of the Internet."

In the meantime, Wilde sees that content providers are coming to a greater understanding of what the semantic web can mean to them, even if they're not calling it by that name -- and won't anytime soon. Rather, they identify the possibilities of the semantic web indirectly, only by means of how it can be put to work to solve their business problems: For example, one of EveryZing's customers, a big network, is interested in how to get its national news content onto local affiliate TV station sites and vice verse.

"They express it as a business problem, and we see it as a technical opportunity. The way we think about it is the ability to richly mark up all these content objects across their vast empire means we can help them create very intelligent rules that make sure the right piece of content is in front of a user at the right time," Wilde says.

"The semantic web is about rich markup so that objects can express their 'aboutness' better and be part of the user experience in a more intelligent way. So the semantic web brings that intelligence without putting too onerous a burden on the user."