Building a Smarter Planet. A Smarter Planet Blog.

Jeffrey Nichols, IBM Research Staff Member

By Richard Silberman, Writer/Researcher, IBM Communications

Jeffrey Nichols is putting a new twist on Twitter that could change the way businesses use social media to identify, engage and market to customers.

“What we actually see from Twitter is just the tip of the information iceberg,” said Nichols, who manages the social media and crowd research team at IBM Research in Almaden, Calif. “Below every tweet there’s a lot more information that people have that they’re not sharing.”

Determined to extract hidden information from social media, Nichols is developing strategies to ask questions directly of targeted strangers over Twitter. The crux of Nichols work is to move from the ubiquitous reactive approach to social media, where marketers follow and respond to what people are saying, to a proactive model where they can reach out to individuals to collect specific information.

“The number one thing we’re trying to do is to get beyond social sentiment, which is basically an aggregation of whether the crowd feels positive or negative about something,” Nichols said. “Direct engagement with individuals over Twitter allows us to understand people and their opinions a lot better than we can with a coarse measurement like sentiment.”

Soliciting information from strangers over Twitter

Nichols, who studies human-computer interaction in the User Systems and Experience Research (USER) group at IBM Research, has led several experiments that validate the viability of engaging individuals over Twitter.

Two years ago, Nichols launched TSA Tracker, which collected and displayed wait times at airport security checkpoints. Nichols and his team identified people who were at airports, based on their Twitter status updates, and asked them to share their estimated wait.

“We got a much higher response rate than we expected — above 40 percent — which suggested that this is pretty fertile area to explore,” Nichols said.

Nichols’ next study, being presented at a social computing conference this week, not only measured response rates, but also analyzed the quality of the information respondents provided. In this two-part study, Nichols sent questions via Twitter to individuals who own a particular tablet computer — and also queried people who visited Los Angeles food trucks. He then compared the Twitter replies to reviews on reputable online sites to see how closely they align.

“Once again we got a response rate around 40 percent, which shows that people seem pretty happy to respond to questions over social media,” Nichols said. “The most exciting outcome, however, is that over 70 percent of the responses were really high quality answers, and in many cases people gave us more information than we asked for.”

Modeling personalities based on Twitter text

Targeting the right people on Twitter is, of course, essential to getting high response rates and good data. To that end, Nichols and his team are currently exploring ways to analyze personality traits based on the text in social media messages. By modeling and understanding personality types, it will be possible to reach out to the right individuals, in the right way, at the right time, to elicit the best response possible.

“If you just send a person a message out of the blue and it doesn’t offer any value to them, then that’s spam and they’re not going to respond to that,” said Nichols, who is developing analytics to infer who will likely respond to unsolicited questions on a particular topic.

“If we choose to reach out to someone, we want to know there’s a very high probablilty that they’re going to react positively,” Nichols said. “Response rates of 40 percent are great, but wouldn’t it be awesome if we could get that up to 80 percent?”

New tools for Enterprise Marketing Management

Nichols’ fascination with crowds has its roots in the massively multiplayer games he once played. His interest in harnessing massive social communication and using Twitter in new ways evolved from there.

In one of his first social media experiments, Nichols and a team member used only the status updates posted to Twitter to generate summaries of sporting events comparable to those created by professional journalists. Encouraged by the results, Nichols has been on a steady path of innovation ever since.

“Going forward we hope to take what we’ve done so far, make it more business-relevant and bring it to market,” Nichols said. He is currently working to incorporate some of his techniques into IBM’s Enterprise Marketing Management (EMM) solution to support a very focused, individualized and interactive approach to marketing.

“I think our approach to using Twitter will become a routine part of doing business in the next couple of years,” Nichols said. “It’s not only going to enable companies to hone valuable relationships with individuals, but also to reach beyond and expand their established customer base.”

Technorati Tags: Enterprise Marketing Management, human-computer interaction, Jeffrey Nichols, twitter