Should a Social Enterprise Publicly “Toot its Virtuous Horn”? An Artificial Intelligence Approach

Social enterprises are receiving increasing recognition as businesses that address global challenges for a better and more sustainable future. However, social enterprises can struggle to communicate their values to obtain long-term viability. Little is known about measuring an enterprise's promotion of social values, which we call socially-oriented language, nor when such language helps or harms social enterprise performance. In this research, we draw on Artificial Intelligence techniques, in particular natural language processing techniques (NLP), to measure socially-oriented language on the basis of textual data. Furthermore, we explore the relationship between levels of socially-oriented language and business performance. Our findings suggested that social enterprises under-promote their social values compared to traditional firms, and they do not gain performance advantages when they do. This research breaks important ground in social enterprises and social entrepreneurship research by providing novel insights into sustainability issues in organizational identity and its communication with business performance. Moreover, it provides scholars with a method for measuring socially-oriented language that traditional approaches cannot measure.