Oops! The input is malformed!
Originally published September 26, 2013
Developing a model for representing social networks and then analyzing those networks for profitable insight often hinges on characterizing any specific individual’s sphere of influence across that network. And, of course, the sphere of influence and the strength of that influence must be reflected in the social network model. But that also means that you must be able to measure any individual’s sphere of influence, and that highlights the importance of different graph metrics. These metrics are intended to provide some insight into the qualitative characteristics associated with linkages within and across the network. In turn, these metric measures can be evaluated in the context of different types of relationships in the graph that point to the different qualities and criteria for identification of key individuals within the social network.
For example, a good place to start is Wikipedia’s page on social network analysis1 that provides an interesting list of metrics that deserve further consideration to review how their measures can be interpreted for customer centricity.
|Metric Name ||Description||Interpretation for Customer |
|Distance||The number of edges between any two specific nodes.||Characterizes closeness of any pair of individuals.|
|Betweenness Centrality||This measure is associated with a specific node in the graph, and quantifies the number of shortest paths from all nodes to any of the other nodes that pass the specific node.||Finds individuals who are near intersections of close-knit sub-communities. These are individuals whose influence spans multiple groups, and are like “brokers.”|
|Closeness Centrality||This measure is associated with a specific node in the graph, and quantifies the average distance from that node to every other node in the network.||Finds individuals who are near the center of clusters, and indicates people who are well-known within a group and whose opinions are well-respected.|
|Eigenvector Centrality||This measures the connectedness of any individual node to parts of the network that are highly connected, and thus finds individuals whose connections also have many connections. ||This identifies influential individuals with connections to other influential individuals, but not necessarily the breadth implied by closeness.|
|Bridge||Identifies an individual providing the only link between two other sub graphs (either individuals or clusters). ||Finds the key people for ensuring that a message can be transmitted across the breadth of the network.|
|Density||The ratio of direct connections in the network compared to the total possible number of direct connections.||Characterizes the degree to which the entire customer population is tightly connected.|
|This measures the degree to which individuals with a set of characteristics are connected to others with similar characteristics.||This reflects the concept that “birds of a feather flock together,” and helps find whether certain similarities exist across groups of connected customers.|
|Multiplexity||This measures the number of different links between any two nodes.||This helps find sets of customers that have a stronger or more closely knit connection, such as people who are related, work at the same company, and live in the same town.|
|Reciprocity||The degree to which two individuals mimic each other’s sets of connections.||This can identify people who share a set of interests or friend but may not have explicitly connected with each other. This can help with suggestions for alignment with special interest groups.|
|Cliques||A clique is a group of individuals in which each one is connected to all others in the group.||This is useful for finding communities of interest that share similar interests together – good for driving recommendation engines.|
Recent articles by David Loshin