Dlya Jolly, senior product manager at Google, is presenting some of the search and display giant’s real-time marketing diagnostics at the Brand Marketers Summit, and she’s making a pretty good case for why Google’s algorithms are so good at sniffing out the meaningful signals sent by online users that make them potential prospects for Google’s brand advertisers.
They are so good now, she says, that they can now be deployed at the “scale of TV.”
She says it was a two-year initiative to build Google’s infrastructure and to fine tune its algorithms to get the kind of fidelity necessary to do that.
“We were actually able to get there,” she boasted, about attaining TV targeting scale, noting, “It’s easier to do all this stuff if you’re going to find 10,000 users. It’s harder to do this stuff if you are trying to find the 30 or 40 million moms in the U.S.”
Jolly said the new consumer brand targeting features will be officially launched next Thrusday, and will give marketers -- small and large -- the ability to target consumers based on their own intent signals at scale across all the platforms Google touches: “Display, video and mobile,” not to mention search.
Amazingly, she says it can be done in real-time, noting that it only takes about “two seconds” from the time a brand is served and “making meaning of it.”
That may sound simple, but it’s anything but, and there’s a lot of complex data processing -- and parsing -- going on behind the scenes of data’s algorithms.
The trick, she says, is understanding how to read and interpret the signals based on an individual’s behavior. For example, she noted her husband is not a frequent online researcher, so when he goes go online to research a product, that one-time experience “means a lot.” Versus her own behavior, which is incessantly online, and may not carry as much weight per browsing session.
“Basically, being able to look at your Web behavior and being able to normalize it, is another capability of an algorithm,” she explained.
If you want to learn more about the increasing role algorithms are playing in parsing user data and turning them into meaningful signals come to the Algorithm & Blues panel I’ll be moderating at OMMA RTB in Los Angeles on July 25th.