Abstract:
What information does the internal communication between employees of an organization convey? How do these develop and evolve over time as the company grows and formalizes? We study these questions through an in-depth and data-science analysis of four years’ worth of internal Slack public channels communication in a mid-size high-tech firm. We empirically examine four core aspects of employees: Behavior, Text Focus, Text Affect, and Communication networks. We identify two cardinal and complimentary types of behavior (activity and interactivity) using factor analysis and three types of focus (Internal Tasks, External Tasks and People) using topic modelling. We find that members of different divisions maintain different norms regarding their behavior and regarding topics discussed. We extracted the expressed affect in employee messages using sentiment analysis and analyzed their change over time as the firm evolves and formalizes. We then use network analysis tools to identify patterns of exchanges between users and emergent sub-communities within the firm. We find that the growth and maturity of the firm leads to an emergent pattern of an increasing tendency among employees to interact within their own division rather than other divisions and a generation of sub-communities with respect to the divisional structure of the firm.