On applying Dynamic Network Analysis (DNA)

Q4h1MI

Image courtesy: morguefile 

Once you begin to study networks it is difficult not to see them everywhere.” ~ Sanjeev Goyal
As I was going through some papers for my PHD, I stumbled upon this interesting representation of organizations called : Meta-Matrix

The meta-Matrix has been introduced by Kathleen M. Carley from the Carnegie Mellon University to represent the entities within an organization and the relationships between them. The meta-matrix is a multi-color, multiplex representation that focuses on people, knowledge/resources and events/tasks.

Meta

What is interesting about this representation is that it sees an organization as a set of dynamic networks and I couldn’t agree more. Changes in one network cascade into changes in the others; relationships in one network imply relationships in another. 

While classical SNA (Social Network Analysis) concepts applied to organizations focus on only a cell or two of the table above, the meta-matrix representation takes into account the different interactions between the components of every organization and considers the overlapping between the networks. This helps create a number of metrics that do a better job in explaining the evolution, performance, and adaptability of the network’s dynamics. 

How can such representation be used?

In her book, Complexity leadership – Volume 1, Mary Uhl-Bien presents interesting applications of Dynamic Network Analysis along with the meta-matrix representation. She states that collecting and analyzing data from the social, knowledge and task networks can help measure the communication density within an enterprise; which is “a measure of how many communication relations exist as compared to the total that could exist. This can provide feedback on the relational coupling and social capital structure of the organization”.

The model can also be used to forecast knowledge diffusion. Collecting data from the knowledge network defines who knows what, the social network would define who is talking to whom and the task network could be used to weight the extent to which each agent talks to each other. Analyzing such data can help when implementing a KM solution fro example…

Although one cannot deny the complexity of such approach (most DNA applications are purely hypothetical), seeing the organizations as a set of dynamic networks is probably the most accurate and near-real representation that has been introduced so far.

So DNA offers huge potential but it is still far from going mainstream. We need effective tools that make the whole theoretical process transparent to the knowledge workers. And most of all, we need to raise the awareness that such models can definitely help organizations better structure themselves and face the rising complexity of their environment.

Lamia Ben