Review section in beauty commerce allows customers to submit their assessment, examination, or complaints about products they have used. Sephora, one of the largest beauty commerce stored 2,909 products with at least 100 reviews per product. Importance-Performance Analysis (IPA) can assist stakeholders in diagnosing which part of the product is still lacking and setting priorities when making decision and developing next strategy. Usually, surveys are needed as the primary data to conduct this analysis, but it is time consuming and not so efficient. So instead of sending out questionnaires, product reviews from the website can be utilized. Keywords extractions were used to get importance score and the performance score was obtained from the sentiment analysis using text reviews and star ratings. The learning model using DistilBERT showed 83% of accuracy and the attributes of the product that needed immediate improvement are the product result, skin types of the product, and the product absorption.