In content creation, customer behavior insights are very important as they help creators find and create the content that drives sales. To comprehend customer needs, content creators need not just generalized information but also specific information, which can be different across markets and cultures. This information also needs some standards so it can be analyzed systematically. This paper aims to obtain customer insight into web content. Inside the web content, one possible source of this information is the tags based on customer feedback and the related entities. In this case, the product review data were collected and analyzed. However, manually analyzing feedback is a time-consuming activity. In this work, we formulated the topic analysis problem specialized for material and product sourcing, which could benefit product analysis and development. Technically, we also compared different text processing and classification methods, which set the benchmarks for reviewing the model performance in the future.