NBCUniversal is using machine learning to make ads more relevant

Advertisers have been targeting their messages to relevant media for as long as advertising has existed. Kids' TV channels are awash with toy commercials, breaks during wedding shows are full of ads for honeymoon destinations and so on. But now, NBCUniversal has adopted a new machine learning tool that makes the placement of ads on TV more exact, relevant and timely.

The Context Intelligence Platform analyzes scripts, closed captions and visual descriptions to identify the perfect moment for particular ads to appear. So if you're watching a USA Network show that features a party scene, for example, it won't be a coincidence if the following ad is for an alcoholic drinks brand. According to NBCU head of marketing and advertising creative Josh Feldman, trials of the platform have resulted in a 19 percent jump in brand memorability, 13 percent in likability and 64 percent in message memorability.

The system is due for official release early next year, with three to five advertisers currently working with NBCU on beta-testing. No details have been given on the brand names involved, but Feldman said they fall under a range of consumer categories including retailers, studios and packaged goods.

"Before there was machine learning, there was common sense," Feldman said. "When my team created something for an advertiser that was a really heartfelt piece of marketing, we would make sure that the spot ran on heartfelt-type programming as opposed to, say, slapstick comedy. We've been doing this on a manual basis for a long time -- but now we're going to be able to do it at scale."