Social network analysis

In a networked economy, organisational advantage can depend on relationships as well as internal knowledge. Social network analysis maps connections among employees and external stakeholders—such as contacts within a supplier—and represents them as nodes for actors and ties for relationships. The map helps managers see how information, resources and influence may move through the system.

When to use it

Use network analysis when the organisation needs to understand its reach, identify people who can mobilise support or locate relationships relevant to a product launch, commercial campaign or crowdsourcing exercise. The method produces a snapshot and can be complex and time-consuming, so define the decision and network boundary before collecting data. It is especially useful for identifying where relational strength is concentrated, which experts could be consulted and where dependence on a small number of people creates risk.

Origins

Social network analysis draws on graph theory, sociology, psychology and anthropology. Jacob L. Moreno used sociometry and sociograms in the nineteen-thirties to represent interpersonal relationships. Anthropologist John A. Barnes later applied the term social network systematically in a study of a Norwegian community. Research at Manchester, Harvard and elsewhere developed methods for examining cliques, centrality and structure, while advances in computing established the field as an interdisciplinary analytical practice.

What it is

A network consists of nodes, representing people, teams, organisations or other actors, and ties, representing a defined relationship among them. Analysis can examine degree, density, centrality, brokerage, clusters and path length to understand how information, influence or resources may travel. A network may be directed or undirected and its ties may be weighted or unweighted. Every result depends on what counts as a node and a tie, so a contact, a trusted advice relationship and a contractual dependency should not be treated as equivalent.