Loading...

Enterprise Communication Analysis

Catch the wind of change with Graphlytic

We are overwhelmed by data of any kind. We try to leverage them, more or less successfully. Smaller companies strive to operate them in Excel kind of tables or in any cloud-based application. Larger companies store data in warehouses and data are presented in the form of reports with tables and graphs. With growing amount of data, the phenomenon of relationship data arises. So, what if we need not only data structured in the tables but we wish to work with many relationships between single data elements?

As you may guess, the graph visualization is the answer :)

And, what if we say we can transfer the existing company data elements into relationship data that can bring unexpected results for managing the organization smarter?

How to do it?

We will look at overall company communication - to bring new findings. We will use enterprise communication data to examine the behavior of the employees, and the company as a breathing unit. We can use any kind of communication events such as call logs, e-mails, company chat or social media tools logs, as well as company phone and e-mail books. These events could be enriched with data from any company tool as CRM, ERP etc. There is a general rule for graphs – the more data we use the better visible and easier to identify behavior patterns are. Having this prepared in a tailored data model we will follow the patterns of collaboration among employees and the teams in a graph visualization tool Graphlytic. And, as with any live element we start to observe that behavior is changing in time - we will use timeline filter slider for that. But not only this. We will measure the communication since the beginning which will allow us to see and manage trends of collaboration.

This way we can find informal leaders, the talents, or the connectors of teams. By changing the collaboration patterns, we can easily unveil potential problems within organization or among single units. And by having this ability we can influence and intentionally change and improve collaboration patterns to improve the overall company results.

To start with the basics let us visualize a simple example of phone calls communication among four employees within a given period. Graph below is composed of employees (nodes in graph terminology) and arrows representing certain amount of calls (which is defining relationship) among them. The direction of the arrows describes the direction of communication. The thicker and more of a red color the arrow is the higher intensity of the communication was. Having it weighted using clustering coefficients for example allows us to set a bottom line for measuring and evaluating the collaboration development.

Core concept of enterprise communication visualization

Now, what if we apply this approach to a larger scale – real enterprise – and will focus on overall communication patterns and trends? Can we look upon it as a social network? Yes, we definitely can. In the picture below we can see an example of a small company or part of a corporation depicting graph visualization of communication between the departments. And the communication describes the collaboration perfectly.

After a short fine-tuning we are able to see the dynamics of collaboration within each team – who is the heavy communicator, which team players act as a solid team connector and who are to improve, what the trends are, or how the organizational change influenced the behavior and collaboration within and among teams.

Unveil sophisticated fraud patterns much easier using graph visualization

In this example we can see a heavy communication between the product owner and SW development team. Is it strong enough? Does she communicate with the right positions within the SW development team? How did it change after last organizational re-shuffle? Is the collaboration of the product owner sufficient with the sales team? And this way we can go on limitless to find potential bottlenecks or uncover the real rising stars in the company. So, jump in and simply follow the step-by-step improvement process based on previously unseen and unmeasured data.

Summary