Social media has amply demonstrated its abilities as an outbound tool, a means for government to push information out to its constituents. Municipalities tweet emergency information during crises. Politicians rally friends on Facebook.
Even as these active uses of social media come into their own, newer passive uses are evolving. Rather than shout, government agencies listen: They harvest the chatter, sifting for relevant mentions that might help them to better respond to crises and emergencies.
During the Deepwater Horizon oil spill, Tulane University students used Ushahidi, an open source software platform, to aid in the massive cleanup effort. Participants in the Oil Spill Crisis Map project helped visualize data on maps by harvesting reports generated through email, text messaging, Twitter and other social media platforms.
Governments increasingly are looking to that kind of social media monitoring to help keep tabs on the mood and activities of those who receive services.
Social media monitoring is a relatively new practice in state and local government. It had its start in the retail arena, where uses can be fairly sophisticated, as businesses labor to target their marketing and advertising to likely subsets of buyers.
Still, headway is being made in governmental and quasi-governmental organizations. One of the clearest examples comes from the American Red Cross, which in March 2012 established a digital operations center at its national headquarters. Built on technology platform Radian6, the center’s console displays a running stream of social media mentions based on keywords of interest to the agency. “Red Cross” alone draws some 4,000 mentions a day.
“We are trying to give the public a seat at the table of our operation,” said Wendy Harman, social strategy director for the American Red Cross. At first the system was set to watch just for names of the agency, but it has evolved. “Now we can also see mentions of a given disaster, anybody talking about an earthquake, a fire, a tornado, a bus crash,” she said. “Anything like that, we can see it in a visual format that is easy to digest.”
This has practical implications.
“It is becoming an expected part of situational awareness. Every morning during a disaster, we produce a report that is distributed widely, with graphs and data points, so everybody is on the same page as to what the affected public is saying,” Harman said. “We know what people are going through, we know where there are pockets of need and we know what to watch out for during the course of a day of responding.”
In a time of tight budgets, some are achieving similarly promising results with less expensive technology. Take, for example, the Clark Regional Emergency Services Agency in Vancouver, Wash., where planners use the Twitter app TweetDeck to follow conversations.
Most staff members in the agency run two computer screens: one for daily work and one for watching tweets that are filtered according to relevant keywords and hashtags. “Anytime something begins to trend or starts getting popular, you’ll see the TweetDeck screen start moving a little bit faster and that will catch our eye. It may be an earthquake or maybe a celebrity has died,” said Cheryl Bledsoe, emergency management division manager.
During the 2009 swine flu outbreak, the monitoring system helped city managers control the situation. As citizens tweeted their concerns about the vaccine’s validity and availability, managers were able to make correct information available. Such capabilities give responders a clearer view of the situation during a crisis. “You are getting the news on the ground: ‘This is what I am seeing,’” Bledsoe said. “So it validates a lot of what you are seeing on the emergency response side of the house. It helps us piece together what it looks like outside our offices.”
Some at the municipal level are getting the job done with even less sophisticated tools. Administrators at the Washington, D.C., public library simply monitor neighborhood email lists for mentions of their operations.
“This way we get the most specific information about what people think about us,” said the city’s chief librarian, Ginnie Cooper. “Someone says there is a really good children’s librarian in this place and the moms love them. Or maybe there is a complaint of some sort. Then that becomes something we can act on.”
One sure way to gauge the rising popularity of a given technology is to tally the number of vendors vying for a spot on the playing field. In the case of social media monitoring, that number has been rising steadily and today includes such aspirants as Alterian, Social Radar, Radian6, HootSuite and U.K.-based Brandwatch. To these one may add existing big-name players Dell, SAS and IDC. While offerings and prices vary widely among these firms, all are looking to claim a place at the table.
Vendors in this space all offer monitoring tools of the type likely to be used by state and local governments. These instruments generate streams of observed mentions and conversations — although sometimes the effect can be to overwhelm their users. Even when agencies can afford the tools necessary to track social media, they may not always have the resources to utilize that data. The Red Cross relies on a corps of volunteers to sort through its social feed, with individuals assigned to track certain functions like food, shelter and so on.
The effort to sift data can be a burdensome task, but it’s often worth the investment, said Maribel Sierra, director of social media services at Dell.
“As government, you are trying to get the pulse on any conversations that are happening out there, that are going on right now,” she said. “There are tons of those conversations, and if you just stop and listen, you can get so much value.”
For those who cannot tackle the logistical burden of listening — who cannot spare the personnel or drum up the volunteers — technological fixes are available. Today’s more sophisticated tools can slice and dice social chatter for greater accuracy. Suppose an agency wanted to hear conversations about Apple. With fine selection, software can discern an iPhone mention from a discussion of apple pie. Dell learned this when its engineers tried to track themselves and kept running into a singer with the same name. “We always knew when he was on tour,” Sierra said.
Other tools can filter by sentiment and tonality. By weighing various adjectives, for instance, these systems can deliver just the posts that are positive or negative, said Jodi Blomberg, SAS state and local government principal technical architect.
Even with such precise searches, it still can be challenging to sift the wheat from the chaff. No matter how precisely one cuts the search criteria, there still is a lot of talk going on in social media and the prospects of being overwhelmed by data are significant.
The fix, some say, is to outsource the whole project. In addition to selling the software, some vendors will deliver monitoring and analysis as a service.
“You don’t need to make the investment in the tool, which is where the bulk of the dollars go, and you don’t need to go through the learning curve,” Sierra said. “You can leverage the learning that we have acquired through years of doing of this for ourselves.” Dell’s social media services cost between $5,000 and $10,000 for setup along with a $6,000 to $20,000 monthly fee.
Sharing the Load
Even when state and local agencies cannot harness the power of social media monitoring themselves, there have been cases in which nongovernmental organizations have been able to mine the chatter and share it productively with partners in the public sector.
Take for instance Humanity Road. Founded in March 2010, this Virginia-based organization maintains relationships with various emergency management entities, whose efforts it supports by detecting and passing along meaningful social media communications in times of crisis. This public-private cooperation rests on the backs of the organization’s 75 volunteers.
After the 2010 Haitian earthquake, Humanity Road volunteers tracked official messages about hospital vacancies. Social conversations clued them in that news about hospitals’ status was not reaching the right people, and organizers were able to communicate accurate information directly from hospitals to the Coast Guard, to give rescuers a more complete picture of the situation on the ground.
Humanity Road has forged similar cooperative efforts at the state level. When Hurricane Irene flooded the Poconos in Pennsylvania, the volunteer corps picked up on a microblog that was posting names and contact information for those trapped. The organization contacted county authorities, who otherwise might not have known about those individuals.
Government authorities might not be overanxious to act on the word of a nongovernment agent claiming to have urgent news drawn from Twitter or the blogosphere. But it works for Humanity Road because the organization has made an effort to forge those ties in advance.
“Because we work an event publicly, because we retweet [government] information, and because we attend conferences of the National Emergency Management Association, we have developed some recognition,” said President Christine Thompson.
For all its virtues — greater awareness, improved responsiveness — social media monitoring remains an imperfect means of advancing government interests. In the most immediate sense, analyzing social media chatter means, mostly, trolling through junk. Thompson said that in her group’s experience, social media is a 90 percent ratio of noise to signal, and most casual users would probably agree.
At the same time, social media monitoring creates a new level of accountability, said Harman of the American Red Cross. Every time the agency acts on a tweet, it confirms in the public mind the notion that the Red Cross will act on information shared through Twitter.
“It means we take on a whole new level of responsibility, because it raises the expectation that we are doing something about each thing that a person takes the initiative to put out there,” Harman said.
However, not all of the news is positive: Technical hurdles remain. Automated language recognition is not perfect, especially when it comes to catching mood and tone.
Perhaps the most significant drawback — and it’s a drawback throughout all of social media: There’s no way to know who is talking. Anonymous sources must be viewed with skepticism, especially when those voices are helping to direct policy or apportion resources during a crisis.
Despite all the effort, possible costs and technical imperfections, social media monitoring still has many benefits for government officials looking to improve their performance. In the current climate, anything that brings the servants close to those who are served can only be a good thing, said Ruthbea Yesner Clarke, research director for IDC Government Insights’ Smart Cities Strategies program.
“We have incredibly low voter turnout in this country, incredibly low participation in our democracy,” she said. “If you can use this to increase engagement, then maybe you can increase public interest in their local governments. Confidence then goes up because there is a fuller understanding of the initiatives that might be happening in your city, and expectations may change because people have a far better understanding of how things are working.”
This holds true not just at the institutional level, but also for individuals in government. “For elected officials, it’s a chance to understand: What are people saying about me? Who am I not reaching? Who is not understanding me? And what is the public perception of what I am trying to do?” Clarke said.
Taken altogether, it seems likely that the collective input of tens of thousands of citizens ought to give some new level of insight into the way things are, and might be, run.