In April you may have tweeted how much you hate doing taxes. Sometime later you may have been browsing the Web and noticed ads for TurboTax popping up. That probably wasn't an accident.
That's because TurboTax-maker Intuit was testing New York-based startup LocalResponse’s new Intent Targeted ad product, which uses publicly available social data to target mobile and desktop display ads.
“We found over a million people that tweeted the week before tax day ‘I hate doing taxes’ [or] ‘Taxes suck’...and when these people actually came to the Web, the desktop Web, we actually displayed to them a TurboTax banner ad,” said LocalResponse CEO Nihal Mehta. “This is revolutionary because literally we’ve done tons of research in the space, and we don’t think there’s anybody else doing this.”
Creepy? Yeah, a little. But LocalResponse limits itself to only publicly available content. So while most tweets are fair game, a Facebook status post is shielded, provided users’ privacy settings block anyone from being able to see it.
Here’s how it works: LocalResponse combs through public social media content from the likes of Twitter, Facebook, Pinterest, Instagram and Foursquare and identifies a social ID, like a Twitter handle, connected to that content. The company then partners with third-party data providers to connect that social ID to a cookie that will identify the user the next time he or she visits a website and logs in with his or her Twitter account.
At that point a cookie is dropped on the user’s browser, so that if he or she navigates to a site where a Twitter log-in isn't required, LocalResponse can recognize that person's cookie within an advertising exchange—and then deliver an ad relevant to that user's social media posts.
Mehta said LocalResponse is able to target up to 100 million unique users on a monthly basis through its partnerships with numerous data providers and online publishers. Company president and co-founder Kathy Leake added that LocalResponse expects to encounter zero privacy issues with this kind of targeting because it only uses publicly available social data such as tweets that can already surface from searching on Google or Bing.
LocalResponse says it can also employ natural language processing, which enables it to infer things such as a person's location. For example, a user could tweet, “There’s a sale at Retailer X,” with Twitter geo-location settings turned on, and LocalResponse could identify the words “at” and “Retailer X” and correlate that with the longitude and latitude of the tweet’s location to determine that the user was in Retailer X's store when the given tweet was posted.
And if Retailer X is a LocalResponse customer, the brand could access the company’s self-serve dashboard and launch a display campaign targeted at that user. That's about as precise as it gets, according to Leake.
“If you’re broadcasting intent—whether that’s sentiment or your location—that’s pretty much the strongest signal there is. It’s not look-alike targeting. It’s not traditional segmentation or behavioral targeting or contextual targeting. It’s actually intent that you yourself have declared,” asserted Leake.