Veoh Targets Video Ads Based On Past Viewing Patterns
If behavioral targeting is the great hope for display advertising on the Web, can it work for videos as well? Web video startup Veoh thinks it can and is bringing its behavioral targeting advertising program out of beta today. The ads are targeted at one of nine groups, including viewers interested in action videos, cars, pop culture, sci fi, anime, and family fare.
Veoh groups viewers into these interest groups based on their past viewing, searching, browsing, tagging, and commenting activities on the site. The ad-targeting technology uses some of the same underlying algorithms as its recommendation engine, and were both developed by chief scientist Ted Dunning. He previously built the recommendation engine at MusicMatch (later bought by Yahoo) and credit-card fraud detection algorithms at ID Analytics. The company claims that during the beta, ads that were behaviorally targeted performed twice as good as ads that were not.
Everyone’s trying to figure out how to make ads work on Web videoâ€â€from YouTube to VideoEgg. A big issue is the quality of the video inventory out there. Many advertisers don’t want to risk associating their brands and products with user-generated video. That includes a large portion of the 100 million videos a month watched on Veoh.
Also, for behavioral targeting to really work, it needs to be done at Internet scale. Veoh not only needs to prove that it can provide better response rates to its video ads, but that it has a large enough inventory of advertiser-safe videos to matter. To do that, it would have to somehow monitor video-watching behavior beyond its own site (which it could do via partnership agreements) and become more of an overall video ad network. It would then have to make sure it doesn’t get tangled up in some of the privacy issues that behavioral targeting for display ads are running into.
Veoh raised $30 million in June (bringing its total capital raised to $70)
Crunch Network: CrunchBoard because it’s time for you to find a new Job2.0
Original post by Erick Schonfeld