<p>Perceived quality reflects the subjective perceptions of consumers regarding a product and holds significant importance for manufacturers seeking to enhance quality. In recent years, social media has emerged as a new platform for consumers to share their perceived quality experiences. However, existing studies have overlooked the information usefulness inherent in each data point, thereby creating obstacles for manufacturers in processing impactful information. To address this gap, this paper proposes a two-stage approach to quantifying perceived quality based on the usefulness of information. Firstly, the categories of usefulness related to perceived quality are identified through a combination of deep learning and the knowledge adoption model. Subsequently, multiple usefulness categories are considered to quantify the information pertaining to perceived quality. In the empirical study, the proposed method was validated using an automobile dataset obtained from Autohome. The results demonstrate that the method acquires more effective perceived quality information. To showcase the practical relevance of the proposed method, the current study further explores the application of the perceived quality from three perspectives, namely price variability analysis, attribute importance analysis, and competitiveness identification. This contribution further advances research in both perceived quality quantification and the exploration of information usefulness.</p>