Organizational charts are common in companies. They depict the reporting relationships and should accurately represent who is talking with whom, who is in the know of everything, and who are the people at the periphery of the organization. However, often the real communication network is different to the organizational chart. Knowing how information really flows in your company is crucial for getting many projects off the ground and running.
What are organizational charts?
The goal of an organizational chart is to depict in a picture who needs to be reporting to whom, and who should be given task assignments to whom. Often they look like a pyramid, with the CEO at the top, and then fanning out with more and more managers filling out the ranks. Front line employees are not always mentioned.
An organizational chart can be useful when you want to investigate the formal structure of a company. Before re-organization it is helpful to know how reporting should be done. Organizational charts can also help in talent management and promotion. They visualize the path to the top.
Organizational charts have less value when individuals look at them and assume that they represent real communication lines. It is wrong to look at an organizational chart and think that all information flows according to these formal relationships. Organizational charts might be pretty close to reality for military organization and organizations with an overly bureaucratic culture.
What is a social network ?
A social network, or an organizational social network as it is also called, is a chart that depicts how information is shared in reality. It depicts the employees and with whom they are communicating. Depending on how the data is being collected, it can show task related communication, mentoring relationship, or general communication. Most commonly, a people-to-people network is constructed and used to recommend changes in the formal structure. However, it is also possible to create a task-to-people network to map workflow and discover areas of tensions between people and tasks. This would offer a new perspective to discover sources of inspiration and engagement among employees.
Once you have your network you can begin by analyzing the figure and subsequently calculate specific metrics. These metrics will provide more information about the network positions of individuals. Knowing the network positions it is possible to better understand the power and influence individuals have in the organizations. Several metrics exist to describe a network, and the positions of the members in the network.
For example, degree describes how connected someone is. Briefly, the more connection, the better it is. More connections means that more people are talking with you. This is generally good, until you reach too many connections and can’t get your work done for the amount of people who come to you to talk. For certain job functions it is important to have many connections, while for others this might be detrimental for performance. What I described here, is called degree centrality.
Now imagine two middle managers: Sue and Sarah. Both want to increase their budget by, let’s say, $ 10, 000 and have strong ideas how the money will be spent. Both do not have a direct connection to the CEO. 10 people in the company communicate daily with the CEO. Sue has many direct connections. She has a high degree centrality. Of those, a few have links to employees who communicate daily with the CEO. Sarah, on the other hand, only has a handful of direct connections. She has a low degree centrality. These direct connections communicate with many different people in the company, also with those 10 employees who have direct link to the CEO. Assuming that Sue’s and Sarah’s ideas hold the same benefit and risk for the company, who do you think will get the money ?
While Sue has more direct connections and therefore can easily win these 10 people to support her idea, her connections aren’t well-connected and hence she will have to work harder to spread her idea, and gather support for it. While it might look like Sarah has less power, as she only has a few direct connections, these connections are well-connected. This gives Sarah a greater chance that her idea is spread throughout the company and thus she can easier gather support for it.
The case demonstrates the difference between degree centrality and closeness centrality. Many other measures of centrality exists, such as betweenness centrality or eigenvector centrality. The point is that depending on what needs to be achieved, influence and power should be measured differently.
What you need to know before analyzing your social network
To analyze the communication patterns in your company you can run a simple survey or use electronic traces your employees leave. For example, information can be extracted from the to/from/cc fields of emails and use this as a basis. Of course, such an approach would not capture any informal chit-chat that happens between employees. Survey data and electronic data can be augmented using workflow processes, or physical proximity between employees..
Before analyzing your network, be sure about the implications and work with your employees to ensure that everybody understands the purpose of creating the social network map. Remember that a network will show who is talking with whom. Write down how the network will influence your decisions about promotion, professional development, project assignment etc. Finally, be explicit who will have access to the full data, and who will be able to see only parts of it (e.g., a network without names, only networks of their unit).
Tools to analyze your social network
Kumo is a tool that lets you easily create various network visualizations. You can hand draw them, or import a spreadsheet that contains information about the people in your network and how they are connected.
Socilyzer is a simple tool to collect and analyze survey data and traditional research questions.
Ucinet is great for analyzing and visualizing social network data if you don’t have a programmer or data scientist in house. It has a simple interface, and includes many network metrics. It is regularly updated.
If you are able to collect the data with another tool (e.g., SurveyMoneky or Typeform), you can analyze your csv files in R or Python. I recommend R for advanced network analysis as the tools and models already exists (igraph or statnet package). Of course Python is also an option using the igraph or networkx modules.
While R and Python are great for analyzing and visualizing social network data, they require some experience in those languages. If this isn’t an option for you Gephi makes it possible to visualize networks.