Uncovering People's Patterns Pays Big for Social Marketers



  • People Pattern assembles a personal portrait based on your social media comments.
  • They can replicate in a few days the same personal analyses companies attempt to compile in six months.
  • Identifying patterns creates psychographic groups that advertisers can target more effectively.

L-R: Jason Baldridge, James Scott, and Ken Cho

As you’re heading out for this year’s holiday shopping, it might not be just Santa Claus who’s looking over your shoulder. It might also be entrepreneur Ken Cho, MBA ’03, along with University of Texas at Austin professors Jason Baldridge and James Scott.

Instead of a workshop at the North Pole, their 18 employees are watching from a converted warehouse in East Austin. And instead of naughtiness and niceness, their database holds detailed profiles on more than 100 million consumers: their likes and dislikes, what they do, the cities they live in, and whom they call friends.

People Pattern, started up in March 2013, could represent the next frontier of direct marketing. Google and Facebook serve up ads based on what you click online, but People Pattern takes things a step further. It assembles a personal portrait based on what you say on Facebook, Twitter, and roughly 20 other social media platforms. 

“If you’re on Twitter, and you tell me you listen to NPR and you like Obama, I may know what foods you’re likely to eat,” says Scott, chief statistician at People Pattern and assistant professor in the department of Information, Risk, and Operations Management at the McCombs School of Business.

“One might say there are patterns to people,” says chief scientist Baldridge, who doubles as associate professor of linguistics.

Their goal is to transform market research by bringing it into the era of social media and big data. Why invest in surveys and focus groups when consumers already volunteer their opinions online? Says Baldridge, “We’ll replicate in a few days the type of personal analyses that companies will spend millions of dollars on and that take six months.”“Why invest in surveys and focus groups when consumers already volunteer their opinions online?”
The trick is to continuously sweep social media for the latest posts, then crunch them as though they were stock prices or crop reports. That’s where Baldridge comes in. He teaches computational linguistics, using computers to analyze what people write. His computers can identify the same person’s postings over a variety of social media, sort out real people from phony spam accounts, and classify their various interests.

To that raw material, Scott applies statistical techniques. He remixes individual profiles into psychographic groups to find patterns that might otherwise be missed.

Cho, the firm’s CEO, found his first pattern while he monitored posts about an exhibit on Edgar Allen Poe at UT’s Harry Ransom Center. He noticed a lot of chatter among Goth girls. Upon closer examination, he found they were following heavy metal singer Marilyn Manson, who had done a painting of Poe. Cho got the HRC to target ads at Facebook followers of Goth bands, and foot traffic at the exhibit tripled.

He realized that social marketers had been tracking comments but not the people who were making them. “I started to see a shift in importance,” says Cho, “from understanding the conversation to understanding who’s behind the conversation.”
“[Cho] realized that social marketers had been tracking  comments but not the people who were making them.”
Once People Pattern has identified a psychographic group, it sifts through its databases to find others who might belong. It advises clients on where to reach them and even what words to use.

For the woman-oriented cable channel WE tv, People Pattern tried to build interest in a new legal drama called “The Divide.” By analyzing tweets from the show’s existing followers, it found a majority were white females aged 18 to 34. They tended to follow TV news and series like “Scandal,” “Gladiators,” and “Girls’ Night Out.” 

profile sampleThe firm designed a Twitter campaign around lines like, “Need a new show for #girlsnight?” and tweeted ads to 77,000 other users who fit the profile. The result was an engagement rate of 7.3 percent, compared to 2.2 percent for tweets that were not tailored to that audience. In a week’s time, says Cho, views of the show’s trailer jumped eightfold.

He acknowledges that not everyone’s comfortable with having their online selves in a database, but he points out that people already granted permission to Facebook and others to exploit their personal data. Some firms use more intrusive methods — like Google, which scans Gmail messages to build user profiles and target ads. 

“All the data we’re working with are things people said in public, which each individual is already comfortable putting out there,” says Baldridge. He adds that over time, users are becoming more conscious about keeping embarrassing information offline.

Could such a database lead to a future like the movie “Minority Report,” in which Tom Cruise flees past store windows while they flash him personalized ads? It could, says Baldridge, if you linked it with other databases, like mobile users and GPS signals. 

But consumers have an easy out, says Cho: “On the cutting room floor, they forgot the scene where Tom Cruise was opting out of all the ads.”

The key to acceptance, he says, is that consumers get back information they can use, not random spam. “Believe it or not,” he says, “marketers want to be responsible. They want to make sure the information is valuable to end-users. Your brand is valuable. If you tarnish it with spam, you devalue your brand.”“Consumers do have an easy out,” says Cho. “On the   cutting room floor, they forgot the scene where Tom Cruise was opting out of all the ads.”


Faculty in this Article

James Scott

Associate Professor, Information, Risk, and Operations Management McCombs School of Business

James Scott is an associate professor at the McCombs School of Business in the department of Information, Risk, and Operations Management (IROM)....

Jason Baldridge

Associate Professor Department of Linguistics

Jason Baldridge holds a Ph.D. from the University of Edinburgh and is a computational linguist at The University of Texas at Austin. His research...

About The Author

Steve Brooks


In a quarter-century as a journalist, Steve Brooks has won two Neal awards for excellence in trade reporting and a Press Club of New Orleans award...

Leave a comment

We want to hear from you! To keep discussions on-topic and constructive, comments are moderated for relevance and for abusive or profane language.
Login or register to post comments