When I present research, I’ll sometimes hear people say things like, “I knew that,” or, “This doesn’t surprise me.” Recently, I presented new research about the adoption and use of social media in America, at Blogworld in New York. There are three facts from that study that I suspect will surprise you—and which, taken together, should cause you to completely rethink how you use social media data. I’ll buy you a drink if they don’t. The study, titled The Social Habit 2011 (you can download the complete study for free here), was based on a random, representative, and projectable study of 2,020 Americans ages 12 and older. It was the 19th in an ongoing series of reports we’ve issued since 1998. First, let’s take the relative popularity and usage of Facebook and Twitter: Now, you may or may not be surprised by this. Certainly you knew that Facebook was the 10-ton gorilla of social networks; it is, in fact, the great outlier. You might also have deduced that Twitter was necessarily smaller. The fact, however, that Twitter is six to seven times smaller than Facebook—and this has been corroborated both by our previous tracking studies, and in work done by the equally reputable Pew Internet and American Life Project—does surprise most people I run into, particularly if they spend lots of time on Twitter, which does tend to bias the sample. Second, we asked those Americans who follow brands, products, or services on one or more social networks, which social network they use the most for brand-following behavior. The answer should not surprise you, though the magnitude of the disparity might: So, while Facebook is already six to seven times larger than Twitter, the disparity in brand-following behavior is even larger, with more than 12 times more Americans who follow brands on social networks saying they do so most on Facebook than those who do so most on Twitter. Finally, we asked all social networking users what social site or service is the most influential to their buying and purchase behavior. The No. 1 answer? None of them. No. 2, however, should again be no surprise: Though 24 percent of social networking users named Facebook as the “most influential” to their purchase decisions, no other site or service broke 1 percent, including Twitter (which was at just 1 percent, actually). So, for this particular question, Facebook was named 24 times more often. The biggest disparity of all. I’m not denigrating Twitter—it’s a very important service to me, and (if you do the work to determine this) it might be very important for your business. Or it might not. But consider this: Most of Facebook’s user data (and, even worse, an indeterminate amount of Facebook’s user data) are not exposed to sites and services that measure sentiment, buzz, and influence. So, all of the new sites and services that measure these things—including Klout, Crimson Hexagon, and Radian6—rely heavily on Twitter, the Internet’s great easy button, as their most easily accessible source of unstructured social media data. Think about this, however. If 1 percent of social media users actually tell you that Twitter is the social site that most influences their buying decisions, and services like Klout measure social media influence predominantly through algorithms based on Twitter (they could hardly dispute this), just how far apart are the measures you derive from these services from actual reality? The answer itself isn’t scary. That you don’t know the answer is scary. That answer is knowable. It isn’t rocket science, though it is work to figure it out, no question. (It’s the work I do every day for brands.) If you aren’t doing that work—the real work of social media measurement—then I would submit you don’t know anything about the real impact of your social media measurement efforts. But I’ll happily buy you that drink if I’m wrong. Tom Webster is a 20-year veteran of opinion, media, and marketing research, and the principal author of Twitter Users In America, The Social Habit, The Podcast Consumer Revealed, and other widely-cited studies of consumer technology usage. He is currently vice president, strategy, for Edison Research, most widely-known as the sole providers of U.S. Election exit polling data to all major media outlets. A version of this story first appeared on the blog BrandSavant.