Tracking the Influence – the Conversation Index

by Jason Preston on August 22, 2007

white paperA few days ago (I know, I know, I’m late to the party again. I prefer to say “fashionably late”…) Jeremiah posted a link to a white paper (PDF) he co-authored, titled Tracking the Influence of Conversations. It covers a lot of important concepts about measurement online, and how metrics are going to change—how they must change—going forward.

It’s well worth a read, especially if you’re new to the scene.

Going through the list of “important attributes” the panel came up with for measurement, I was unsurprised to note how many of them were in many ways completely nebulous. How do you track relationships? or relevance?

One of my favorite ideas is Stowe Boyd’s “Conversation Index,” which is a concept that’s very familiar to me. I started blogging in 2001 on LiveJournal, and on LJ there were two ways to figure out how important someone was: the ratio of friends (so, how many people read you vs. how many people you read) and the ratio of comments (how many comments have you made, how many have you received).

A perfect example of this is TheFerrett, a big-wig livejournaler I’ve read for years. If you look at his user info page, you’ll see that he was 273 friends, and he’s the friend of 2686 people. That’s a pretty good ratio.

Now comments. He’s posted 27,782. And he’s received 194,030. Awesome. This guy is big stuff. He’s got an extremely active readership of almost 3,000 people.

As crude as that data is, it’s a pretty good guideline on Livejournal. I imagine Boyd’s Conversation Index as something similar, tacked out to the blogosphere at large.

Like anything else in this medium, it’s hardly the raw data that companies want. I can give you six-hundred reasons it’s a bad metric right now, starting with the commenting culture in the blogosphere at large–it’s not set up the same way it is in LJ. But it would make for a pretty good guideline as to who “matters” and who “doesn’t”.*

—-
* In reality, everybody matters. Hell I get about 2 comments every month on my personal blog, and I still matter. But maybe that’s just my big ego talking.

{ 4 comments… read them below or add one }

1 Jeremiah Owyang 08.22.07 at 12:57 pm

This is a great discussion, thanks. These attributes were what was ‘voted’ up by the expert attendees. Relationships can also be thought about as network, this is easily measurable within Facebook or social networks. MyBlogLog, to some extent, does this for blogs. Reading WHO left a comment in a website is also just one indicator.

Regarding Relevance, we discussed this, and it can also be though of as context to the conversation at hand. Yes, very ghosty and ambiguous but guess what, so are humans.

Social Media Measurement will vary as well humans!

2 Saurabh 08.22.07 at 2:56 pm

The challenge has always been to put a quantitative face to qualitative metrics. Understanding more about the human element in social media is key; capturing psycho-linguistic drivers which could and should be quantified.

From a measurement perspective, the range of attributes important in social media can be analyzed not simply with regard to the product/process spectrum but also with respect to a number of other linguistic variables. Some of these include, a) function of the message and comments (i.e. diversity of purpose is commonly reflected in diversity of writing style and as captured by terms like engagement and sentiment); b) access to verbal and non-verbal cues (which in turn affects impression generation and formation as captured by influential idea and memes); c) special linguistic features (i.e. language and tags that are unique to each tool and affect how they would be measured and captured); d) profile of participants (i.e. ability to distinguish among participants on the basis of demographics (age, gender, language skills etc. ) which in turn affects the credibility of the communication.)

The key is to “isten”. Keep the conversation going…

3 Saurabh 08.22.07 at 2:57 pm

The key is to “listen”. Keep the conversation going…

4 Glenn Fannick 08.27.07 at 2:29 pm

I was one of the Factiva (Dow Jones, er, News Corp) folks at the session that spawned this white paper. We went in there hoping to find some good fodder for an accurate way to measure influence on the blogosphere. As far as I’m concerned, we’re still looking. I heard one interesting comment from a vendor last week. He said something like there really is no such thing as an influence metric because the diversity of readers’ needs is so great and the breadth of content even greater that trying to come up with a “key” bloggers list is the wrong approach. His comments really come down to the Long Tail paradigm again.

Sponsored links

advertise here