• We use text mining to identify harmful e-commerce terms of service (ToS) clauses. • We present a prototype tool that scans ToS and assigns severity rankings. • We identify the impact of customer unfriendly clauses on business performance. • Businesses with higher market share prefer less consumer-unfriendly ToS terms. • The effect of market share on customer satisfaction is partially mediated by ToS. E-business user agreements are seldom read but almost always accepted by users. We present an in-depth analysis of e-user agreement issues using textual data on Terms of Service (ToS) agreements from more than 250 e-commerce companies. Our text mining analysis identifies several terms in ToS that pose significant risks to consumers. We propose a prototype text mining process that scans ToS agreements on behalf of consumers and assigns an overall severity score based on the potential threats. We supplement this with Latent Semantic Analysis (LSA) for topic modeling in our corpus of ToS documents, to identify user-unfriendly clauses from ToS documents. Next, our empirical analysis reveals three novel findings. First, the presence of hostile clauses in ToS negatively impacts customer satisfaction but is associated with higher survival of firms. Second, firms with higher market share prefer less consumer-unfriendly terms in their ToS. Third, the impact of market share on firm performance is partially mediated by ToS severity. Overall, this study analyzes the linkage between market share, ToS severity, and firm performance, and provides significant insights on ToS policies to IS scholars, legal experts, and firms.
Read full abstract