Video services

AT&T Jumps On Board With Jinni

Jinni, the content search and recommendation specialist, has added a new high-profile company, AT&T, to its growing list of pay-TV customers.

Multichannel News reports that AT&T Inc. (NYSE: T) has now integrated the Jinni Media Ltd. solution into its U-verse TV service, extending Jinni's reach into another 5.5 million US pay-TV homes. The Jinni "Entertainment Genome" is powering content discovery features for the U-verse video-on-demand (VoD) library. According to Jinni CEO Yosi Glick, AT&T implemented the recommendation engine about three weeks ago.

AT&T is far from the only big name on Jinni's client roster. Jinni signed an agreement with Microsoft Corp. (Nasdaq: MSFT) in 2011 and then added Time Warner Cable Inc. (NYSE: TWC) and VUDU Inc. as customers in a roundup of wins announced in January 2013. In June 2013, Light Reading learned that Jinni is also the recommendation engine embedded in the latest version of Comcast Corp. (Nasdaq: CMCSA, CMCSK)'s next-gen X1 user interface. (See Jinni Lands Licensing Deals and Jinni Powers Comcast X2 Recommendations.)

Jinni's claim to differentiation is a technology that combines standard metadata with mood-based data tags and a machine-learning system designed to adapt to a user's specific tastes. The company offers web and mobile applications in addition to services for pay-TV and streaming video providers.

The world of standalone video search and recommendation solutions is quickly shrinking. Jinni and ThinkAnalytics Ltd. remain the two most popular offerings that haven't yet been snapped up by larger companies. TiVo Inc. (Nasdaq: TIVO) bought competitor Digitalsmiths Corp. in January, and Rovi Corp. purchased rival Veveo Inc. in February. (See TiVo to Acquire Digitalsmiths and Rovi Snaps Up Veveo.)

— Mari Silbey, special to Light Reading

mhhf1ve 5/9/2014 | 3:34:07 PM
Re: Mood ring TV Recommendation systems do have the "good enough" advantage -- in that no one is really holding the algorithms to some unattainable standard. Most people will look at the suggestions and think "that's okay" -- even if they're really not that good. It's amazing how much you can fool end users with branding -- just look at the various bing vs google taste test results.
Sarah Thomas 5/9/2014 | 1:46:59 PM
Re: Mood ring TV That's interesting about Netflix's prize, mhhf1ve. I am going to pay more attention to what's recommended to me now. Online, the program is sites like Amazon just recommend what you already bought, which would be completely unhelpful with TVs and movies. They can at least promote thier new releases you may not know about yet.
mhhf1ve 5/9/2014 | 3:26:11 AM
Re: Mood ring TV Recommendation systems have definitely gone through the trough of disillusionment. I'm not sure they can emerge again though. Netflix had a million dollar prize for a better recommendation system, but they ended up never using it because the content selection is the limiting factor in most commercial libraries. A good recommendation system will suggest titles that Netflix or any similar service provider can't supply consistently. Licensing agreements for Disney movies or Beatles music is always going to be tough to maintain as the royalties go up and up but users don't necessarily want to pay more and more. So recommendation systems end up suggesting what's available (and profitable for netflix based on licensing terms) rather than on what users would actually want to see. Maybe recommendation systems would work better for pirate sites where any and all titles are available....
Sarah Thomas 5/8/2014 | 5:13:41 PM
Re: Mood ring TV I highly reccomend Smart Start or Honey Nut Cheerios.
mendyk 5/8/2014 | 4:35:06 PM
Re: Mood ring TV I feel the same way about breakfast cereals. I'd like to know which cereals my friends would recommend, then I remember that I don't have many friends. And then I remember that I haven't had any breakfast cereal since probably 1977. So I just get a bagel.
Sarah Thomas 5/8/2014 | 2:41:46 PM
Re: Mood ring TV Movies are something I'd actually want reccomendations on. I never know what to watch, so I'd be curious to know what's popular. Rotten Tomato integration would be nice. I like to know what my friends are watching, but probably wouldn't accept the privacy and annoying factors of integration with Facebook or a similar site.
danielcawrey 5/8/2014 | 2:19:49 PM
Re: Mood ring TV I would try it... but since I know that services like Netflix don't do well on reccommendations for me I am not sure these would work well either. 

Maybe I am too picky. Maybe I would rather spend time searching for something perfect to watch rather than watching anything at all. Either way, I feel like some of these types of engines still need a lot more work. 
Sarah Thomas 5/8/2014 | 11:44:24 AM
Mood ring TV I'm interested to see how this Jinni works on my U-Verse, especially the mood-based reccomendations. I've been using Songza, which suggests different vibes based on the time of day and what you might be doing. I enjoy it. No clue if that's similar to Jinni....but interested to use it nonetheless.
Sign In