Network Analysis - A Use case (Telecom Business)
Social Network Analysis is a well-researched
area of computer Science and Mathematics. Any structure which can be
represented in the form of edges and nodes is a network. Nodes are connected to
each other by edges (sometimes called links). When we study any network, our
objective is to evaluate the importance of each node and link. Network Analysis
can be used in the areas of genetic engineering, social engineering, marketing,
fraud detection, crime detection and economic research etc. Network analysis has
been widely used in genetic engineering. Recently the application of
network analysis has gained momentum in the field of business marketing and
analytical sciences. In this blog, I will briefly discuss the verticals
of businesses in which network analysis can be used to gain insights of
customers. I will exclude the discussion on the application of network
analysis in any of the biological fields as I do not have concrete knowledge of this
area.
1. Telecom Sector: The telecom sector is a great example of real life networks. A mobile number represents a node and conversation with a person using the phone represents an edge. There are millions of SIMs being used by the users and they call to many people in a day. There are examples where a person calls three-four times to the same person in a day but rarely to someone else. He/she calls a person once or twice in a week. So the frequency of conversation between two people varies. This means that the strength of closeness of the two people also varies from person to person. If the frequency of the conversation, average talk time per conversation, use of any other services such as SMS are high between two people, then the link is strong.
There are cases where a person receives
calls/SMS from many people. This is generally the case with a influential person.
Therefore, the node represented by him/her is a very strong node in the
network. There are some exceptions also. An auto/ taxi driver will receive
calls from numerous people, but he is not an influential person. There are
also cases where a person calls many people. These cases can be analysed
using social network analysis, and stronger links and influential nodes can be
easily identified.
A telecom network is a large network. A
large network leads to the formation of communities. For example, in a university
we have a close friendship with a few students whereas we may know more students. We might have met with them once or twice but we meet regularly with
some students. We unconsciously create a friend circle. Similarly, in the
telecom network, we have frequent conversations with a few people. Hence, we form
a community. Communities created in network can be easily identified
by using graph theory algorithms.
Using the attributes of a node, we
classify a person as leader or follower. A leader means he has more than
average number of connections and also influential in the network.
Leaders and followers are defined using business logics. Based on the position
of a person in the network marketing/retention strategies are decided for him.
A leader is high value customer as he can influence other also.
Using network analysis we can identify the communities based on which a plan can be offered to the customers belonging to the same community. This will in turn grow the relationship with the customers and will increased the sense loyalty toward s the service provider.
Comments
Post a Comment