Ooyala Debuts Content Discovery Technology, Drives More Views
On the eve of what promises to be an exciting NAB conference, online video platform Ooyala has announced a set of technologies called Ooyala Discovery that tackles the problem of getting viewers to watch more content -- and sit through more ads.
Currently, when viewers finish watching an online video, they hunt around for something else to watch or read, or perhaps leave the site. Major sites such as YouTube and Netflix have sophisticated personalized recommendation engines, but smaller sites don't. Ooyala aims to change that.
Using proprietary algorithms and anonymized viewing data gleaned from 200 million monthly users and 2 billion daily analytics events, Ooyala is able to recommend videos that viewers are more likely to want to watch.
According to Sean Knapp, Ooyala's co-founder and chief technology officer, the recommendations can appear in different forms, depending on what the site owner wants. The format getting the best results, he says, is one that offers picture-in-picture recommendations 10 seconds before the currently playing video ends. Viewers are offered a choice, and told that one of the videos will auto-play if they don't select anything. It lets viewers know "Here's what's coming up next."
The second-most successful model waits until the current video is done playing and then suggests options. People are currently trained to click away as a video is ending, Knapp notes, so the overlapping model combats that. Recommendations can be placed anywhere, including a site's homepage.
Ooyala's Discovery engine is essential as online video goes from the computer to the living room, says Knapp, and site owners look to create a ten-foot experience. Personalized recommendations allow the viewer to sit back and be entertained.
Video recommendations are from that site owner's library, not across Ooyala customer libraries. The engine looks at what viewers with similar viewing interests enjoyed when it creates the recommendations.
One important metric is which videos were watched to completion, not just frequently selected. The idea is to get viewers watching longer, leading to more opportunities for showing ads. The engine looks for content that's more engaging to viewers. Future improvements will help site owners earn more revenue from videos by showing them how doubled ad loads impact viewing times, or by suggesting the best place in long-form content to insert a paywall.
The technology has been in development for over a year, and is now in testing with select clients. Knapp says that they've seen a 4x improvement in video views by using the recommendations. Ooyala Discovery will be available to all Ooyala customers later this quarter.
"In the online video world, this is really unlocking the potential to finally reach the user in a one-to-one relationship. It delivers on the promise of online video to create a personalized consumer experience that ultimately will become the de facto across all screens, whether it's your PC, your smartphone, or your TV," says Knapp.
The following video interview with Knapp is used courtesy of Beet.TV:
Telstra becomes a major investor, as well as a customer and reseller, offering Ooyala services in Australia.
People are tuning in for longer online videos, says an Ooyala report, bringing an opportunity for more mid-roll video ads.
Video Index finds that tablet viewers are more engaged than their desktop counterparts.
By using Facebook data, publishers are able to compare audience segments or target ads.
Content owners can now rent or sell directly to Facebook's 750M monthly visitors.