REKOMENDASI PENGEMBANGAN PRODUK TABIR SURYA PT KOSMETIK MENGGUNAKAN AISPA DENGAN MEMANFAATKAN ULASAN DARING

• Main Author: B G P W, Mahesa
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/77816
• Institution: Institut Teknologi Bandung

PT Kosmetik has two popular sunscreens that have received low ratings on the Female Daily website, a site for cosmetic products in Indonesia. The low ratings indicate that the company has not fully understood customer satisfaction. This research uses the asymmetric impact- sentiment-performance analysis (AISPA) to provide product development recommendations for each of the customer satisfaction dimensions (CSD) based on their categories. The formulation of AISPA can utilize online reviews. The advantages of online reviews can enhance the understanding of customer satisfaction. The formulation of AISPA begins with web scraping online reviews from the Female Daily website for the company's products (referred to as Focal 1 and Focal 2), as well as competitor products. The process continues with the data preparation and data processing (CSD identification with latent Dirichlet allocation (LDA) topic modeling, identification of sentiment polarity is necessary using aspect-based sentiment analysis implementing the bidirectional encoder representations from transformers (BERT) model, and measurement of the fulfillment effect of CSDs on customer satisfaction is using ensemble neural network model or ENNM). Finally, AISPA recommendation is identified based on CSD categories. This research collected 41.799 online reviews. From data processing, eight CSDs are obtained: oiliness (MI), aroma (AR), texture (TE), performance (PE), packaging (KE), affordability (HA), acne effect (JE), and white cast residue on the skin (WH). According to the CSD categories, recommendations can be obtained through AISPA. The CSDs “MI”, “PE”, “JE”, and “WH” from Focal 1, and the CSDs “AR” and “WH” from Focal 2 are recommended as "possibly constantly improve". The CSDs “AR”, “KE”, and “HA” from Focal 1, and the CSDs “KE” and “HA” from Focal 2 are recommended as "possibly maintain". The CSD “TE” from Focal 1, and the CSDs “MI”, “TE”, and “JE” from Focal 2 are recommended as "possibly significantly improve". The CSDs “MI”, “TE”, and “JE” from Focal 2 are recommended as "possibly significantly improve".