Mining Text Descriptions of Marketing AI Startups: Predicting Venture Capitalist Funding
(A2021-94462)

AUTHORS

Kaushik Jayaram, The University of Georgia; Sundar Bharadwaj, Terry College of Business, University of Georgia

ABSTRACT

Marketing AI startups provide textual descriptions of the applications they build to acquire urgently needed initial funding from venture capitalists (VCs). In this study, the authors have two goals. First, they identify what text descriptions predict the likelihood of acquiring VC funding. Second, they examine if textual description has the power to predict funding beyond the contingent factors (e.g. financial, demographic) commonly used in models predicting VC funding. The authors use text-mining and machine learning to process and analyze textual descriptions of 1681 marketing AI startups from CrunchBase database. They find that successfully funded startups’ text descriptions are more likely to include words related to customer engagement, personalization and segmentation. They further observe that successfully funded startups are focus on application’s past success. Ability of textual information to predict VC funding has important implications for struggling marketing AI startups.