Talking to Valdis
To continue from a previous post, Valdis Krebs uses a software called Inflow to analyse networks -- within companies, among civic groups, among geographical locations -- you name it, and Valdis will often find a network component important to everyday functioning.
Valdis has been doing this for over twenty years -- much longer than the term social network has been in common usage. It's not a gimmick. It's a very smart approach to identify areas that until now have been considered intangibles.
What Inflow Measures (and Why)
Valdis designed Inflow to identify "the most popular social metrics." "We tend to focus on centrality -- how central is a node in the network and how centralized is the network as a whole."
There's reachability which Valdis explains as something that factors in awareness. "It's what the sociologists call prestige," Valdis said. "But Google uses similar metric for page ranks -- in other words, who's pointing to your web page and who's pointing to those who are pointing to your website. So you have to think broadly about it as well as narrowly."
Some metrics can be seen in cluster analysis. This process identifies which nodes are more closely connected to each other than they are to the rest of the network.
"There are small world networks, althought this is more popular with academics than with business people.
"You can also talk about structural equlivalence that show which nodes in the same network have the same connections. If you and I have the same connections, we are structurally equivalent. That could be good news -- we can act as substitutes for each other.
On the other hand, Valdis adds, "It can also be bad news because we're both fighting for the same resources, satisfy the same customers, or when it comes down to reduce the network, one of us can be let go.
People who are structurally equivalent and don't know each other often find themselves in conflict. So network analysis can often identify why people don't get along in ways that are not personal, and something can be done about it."
More on Valdis and measuring networks in the next post.
No comments:
Post a Comment