Nathan Gilliat posts an insightful analysis of the question of defining influence. The central argument comes in these segments:
On the other hand, everyone is an influencer online. Big news stories--and big company problems--don't always flow from the most influential bloggers. Part of the beauty of social media is that everyone has a voice. With a sufficiently compelling story, any voice can start the conversation. Call it the Long Tail, call it Chaos, or just call it a really big job to monitor. The problem is that any mention of a company (including products, promotions and people) could be the first appearance of something that's going to be a big problem. (Or, being optimistic, the beginning of a great opportunity, but those tend to be less urgent than the problems.) The macro approach catches the problem as it gains traction in the larger conversation; what's needed is a micro approach to provide early warning, too.
How do you do micro reputation monitoring? You can use the do-it-yourself approach with vanity feeds and a feed reader; the problem is, that doesn't scale. It works for companies who aren't mentioned a lot (which is a lot of companies, just not the household names). Some of the high-end monitoring services may be able to catch the single-post indicator of an approaching storm, but I suspect that this is an area where a human reader is still most effective.
We recognize that everyone has influence in conversational environments--I've been writing about this for many years--but how ideas spread does have recognizable patterns. Some folks are influential from the get-go on a particular topic based on their regular attention to the subject while others can gain influence based on a well-written posting that gains traction. Yet other folks perform a different role, spreading messages. This is why we talk about the people who shape the discussion and those that amplify it.
So, is it possible to take a do-it-yourself approach to finding influencers at the micro level? Yes, but it is also possible to do a spreadsheet by hand, yet Lotus 1-2-3 and Excel made that a waste of time. Marketers have varying resources, so we help the allocate them as effectively as possible.
Our system doesn't just expose the top influencers, it can show them all, and some of our early customers, notably David Churbuck at Lenovo, used the tool to look at the other end of the scale to start and work up. If you look at Churbuck's posting about BuzzLogic, he talks about contacting Rick Klau about a problem he wrote about on his blog--Rick was not the top influencer, but the returns on addressing him were significant. David recognized this and acted. Smart marketing just got easier, but it still requires smart marketers.
We want to allow people to use their time more effectively. Our first customers are marketers, but this is a platform that can serve functions, including showing how ideas change from the grassroots as well as through the work of media and think tanks.
UPDATE: Bloggers Blog raises some interesting questions, as well, mainly that our metrics are flawed:
Many of the aggregators don't even report back the number of people who are reading individual blogs so it is impossible to determine exactly how many people are reading a specific blog. The idea of using "contextual relevance" and "frequency" may also be problematic -- a blog may still be very influential in a given field even though the blogger changes subjects and goes off-topic for several posts.
We are looking at popularity, using metrics from several sources to estimate readership. But I particularly take issue with the idea that influence is somehow static, as expressed in the statement that influence would not diminish if someone turned away from a subject for a time. Influence does change, and we're looking at it in detail to determine how it is is changing. Someone who frequently posts on a topic would not lose a lot of influence if they stopped posting about that topic for "several posts," but if they did stop for months, they would. Our system takes all that into account.
I also addressed some of this in response to a posting by Matt Hurst.
