Social media offers valuable information that can improve day-to-day operations and strategic planning for businesses and governments. The ease of accessing social media, however, belies the difficulty of obtaining the right information and interpreting it correctly.
The digital marks we all leave with our activities on the Internet increasingly can feed business and government decisions. Many tools exist to mine the Internet for things like public opinion, buying history and intentions, the preferences of specific audience cohorts, extremist messaging, and breaking events. These nuggets lie in a vast universe of information that is otherwise useless for business and government purposes and serve as noisy interference for those seeking actionable content.
Considerable hyperbole often accompanies discussions about social media—usually beginning with how ubiquitous it is, how much it’s growing, and claims that important events could have been predicted if only we had “listened” to social media. Social media can be harnessed to provide decision advantage but if—and only if—it’s used wisely and in concert with other information inputs.
Effectively leveraging social media—collecting the right information and drawing accurate conclusions from it requires three capabilities:
An effective paradigm for collecting and analyzing social media information to advance your mission begins with articulating the question(s) you are trying to answer. This may seem like a self-evident starting point, but framing the right question is crucial to targeting the search for information and finding the relevant answer. And it is better to ask several key questions than to try to condense your interests into one uber-question.
Having defined the right question(s), the next step is determining where the answer(s) lie among the universe of available information.
This search is mapped by defining media environments, which identify the processes and sources by which people receive and transmit information and the demographics and usage patterns that characterize communications of specific audience segments.
This process is accomplished by combining subject matter expertise with resources on Internet usage available on the Internet. Your goal is to reduce the size of the pond you fish from to increase the likelihood of catching the fish you want.
Two competing traps to avoid in identifying your data needs are: (1) “I want it all,” which is casting a net so wide that it becomes difficult to understand, reconcile, store, and query data; and (2) “I have it all” (or availability bias), which is incorrectly assuming that you possess all relevant data and then drawing conclusions based on that incomplete information. Identifying and eliminating what you don’t need can be easier that clarifying what you do need, and is a good first step in defining information requirements.
Some important considerations:
- Social media enjoys an impact disproportionate to the number of users (which varies widely but may only average 20-30 percent of a country’s population). Social media users tend to be among the most influential in a society and traditional media tends to magnify social media by reporting on them.
- Social media, even aggregated, is not the same as polling—it is not a representative sample. Knowing how social media reflects society is key to understanding its influence.
- Baselining is critical—you cannot determine whether a social media response is higher or lower if you do not know the norm. How can you manage the opportunities and vulnerabilities associated with change if you do not first understand what defines change?
Once you have articulated what you seek and have determined the best place to look for it, social media collection tools can be applied to discover and analyze valuable information (skipping the first two steps and relying on a collection tool alone can be costly and biased).
The scope and the diversity of the Internet make it essential that you rely on the right social media collection tool—a tool that fits your mission. Tool vendors often try to pitch their tools to diverse customers, but success in one arena cannot guarantee success in another. Tools that lack the agility to cover multiple social media platforms risk providing limited information that can lead to inaccurate conclusions and judgments, and they can become obsolete when preferred platforms change. Because much of social media is in the vernacular, the tool you select must be competent in languages relevant for your business, and even better if it can accomplish cross-lingual searches where one term, entered in English, for example, is searched by all related terms simultaneously in multiple languages (I am aware of only one tool that can achieve this). Subject matter expertise is essential to understanding slang, sarcasm, idioms, and other native artifacts that can alter meaning. Finally, it is advisable that someone in your company understands how the tool works (without compromising vendors’ proprietary information). Avoid “magic box” solutions that promise answers but do not explain the “magic.”
Having uncovered the rich data appropriate to advance business objectives, you are now ready to analyze them. A number of analytic tradecraft standards borrowed from the intelligence community can improve the rigor and accuracy of social media analysis. These include techniques to deal with the volume, anonymity, and diversity of social media information; the veracity of reporting; determining which posts and topics may go viral (expand rapidly) or have legs (a story that endures); understanding how social media and other information interact; and anticipating which messages and threads will be impactful. Failing to employ these techniques makes you vulnerable to drawing the wrong conclusions from social media, which can have adverse consequences on business and decision making.
Choosing the right indicators is critical to building confidence levels about the reliability of what you are seeing in social media. Among the more important indicators are: source(s) track record, use of appropriate language and slang, confirmation with verifiable first-hand knowledge, the timing of posts, and metadata (data about data). Although only two to five percent of social media messages are geo-tagged, geo-inferencing—determining proximity based on content and metadata—can narrow down source location.
Today there is a consensus that social media offer valuable inputs to business and government decisions, and to measuring the efficacy of those decisions. In response, there is a push to obtain the capabilities to capitalize on these benefits. Keys to success include understanding where to find the information you need amid the universe of noise, identifying the right collection tool suited to your mission, and mastering the analytics needed to integrate social media with other information you possess and to make sense of it to advance your mission.
Patrick O’Neil is an Executive Vice President with The Radiant Group, a premier provider of intelligence and technical services to government and corporate customers, and the former Director for Analysis at the Open Source Center.