There is a pressing need in the modern business environment for business-supporting software products to address countless consumers’ desires, where customer orientation is a key success factor. Consumer preference is thus an essential input for the requirements elicitation process of public-facing enterprise systems. Previous studies in this area have proposed a process to capture and translate consumer preferences into system-related goals using the Consumer Preference Meta-Model (CPMM) used to integrate consumer values from the marketing domain into objectives of information systems. However, there exists a knowledge gap between how this process can be automated at a large scale, when massive data sources, such as social media data, are used as inputs for the process. To address this problem, a case in which social media data related to four major US airlines is collected from Twitter, is analyzed by a set of text mining techniques and hosted in a consumer preference model, and is further translated to goal models in the ADOxx modelling platform. The analysis of experimental results revealed that the collection, recognition, model creation, and mapping of consumer preferences can be fully or partly automated. The result of this study is a semi-automated method for capturing and filtering consumer preferences as goals for system development, a method which significantly increases the efficiency of large-scale consumer data processing.