I've been having a series of discussions with colleagues and others about getting the most accurate and complete assessment of an issues-based online conversation. This isn't the same thing as a keyword search for a product name or a company name, and it isn't the same as categorizing blog posts as "positive," "negative" or "neutral." It's about revealing the layers of nuance in a multi-faceted issue that can't break along the lines of "for" or "against."
Let's say you're a company executive with an interest in what the federal government may change in the regulation of health care. No one opposes "reform," but everyone has a different idea of what "reform" means. For example, there are those who oppose a "public option" in the health care debate but don't use those words. Most of those who support a "public option" don't agree on the details. For doctors, the idea of reform centers around how and what a "payer" (i.e., an insurance company or the government) compensates a doctor for particular services. For consumers, the concept of "reimbursement" is essentially absent from the discussion - it centers on how much you pay an insurance company and what you have to pay for after that.
Go to Google and type in "health care reform" and you get 19,600,000 results, plus a list of sponsored links that provide language that has probably been so focus-grouped that all the meaning has been filtered out of it. Further, the "top results" provide a lot of positional rhetoric but not a lot of analysis or projection.
Go to uber-human-search engine Mahalo and type in "health care reform" and Mahalo asks you a question - "Is 'health care reform' a person, place, thing, event, how to, or other?" Of course, you could argue that "health care reform" is "other," but how helpful is that? And by the way, who likes having their question answered with a question?
Now realize that the people actually making the decisions about America's health care system have never seen about 19,599,800 of the 19,600,000 of the search results you found at Google. Further, references to the most important discussions these people have had about health care - if you can find them on the Internet at all - are very likely buried between search results 15 and 16 million. They're not in the newspaper, they're not on tv or radio.
The most critical discussions are taking place in closed rooms among participants who don't want to be watched. I don't mean that in the conspiracy-theory, crypto-clearance, oh-my-God-they're-plotting-to-take-over-your-body way. I mean that many of the sensitive and complicated discussions about health care (and many other issues) require a quiet and deliberative environment.
There's a difference between "quiet" and "secret," however. These are not secret discussions; in fact, most of the details are available to anyone who wants to know them. It's just that there aren't many people with the education, experience and patience to pay attention to them - or even realize that's what they're looking at. There's an even smaller subset of people in that group who know how to translate health-policy-speak into plain English. And an even smaller subset of those who operate online.
If we apply the measures most commonly used in the social media world, most of these people aren't "influential" at all - if they write a blog or appear online, not many people read it or link to it. They write pieces that aren't optimized for search engines. Many barely show up on the radar. But they wield enormous influence in the areas that truly matter. They walk in the same circles as those who make the decisions that affect everyone. Sometimes they ARE the people who make the decisions.
So to me, this is where social media search must evolve. We need to move from keywords and hashtags to PEOPLE. Google and Technorati and all the shiny "dashboard" tools PR firms build that just look at Google and Technorati only take you so far. I need to do the simple keyword search, sure. But if I want the real picture, I have to find the people who really know, who have really read up on this, who influence the process and the outcome of a given issue. I have to know them ahead of time.
It's very difficult to automate that. It's easier to find people who truly know where to look.