Twitter adalah salah satu media sosial berbasis teks terbesar di dunia dengan berbagai fitur yang menunjang aktivitas komunikasi penggunanya. Seiring berjalannya waktu, pengguna dapat mengirimkan pesan kepada akun menfess sesuai minat dan preferensi mereka, yang kemudian akan mengunggah kembali pesan itu secara publik. Penelitian ini bertujuan untuk menganalisis faktor-faktor usability, social interaction, dan browsing activity yang memengaruhi perilaku impulsive buying pengguna menfess Twitter. Penelitian ini dilakukan dengan mixed-method. Penelitian kualitatif dilakukan dengan melakukan wawancara bersama 10 narasumber pengguna Twitter yang pernah berinteraksi dengan akun menfess Twitter. Data kualitatif yang telah terkumpul diolah menggunakan grounded theory. Hasil pengolahan data kualitatif digunakan sebagai acuan pembentukan instrumen kuantitatif. Penelitian kuantitatif dilakukan dengan menyebarkan kuesioner online dengan 918 responden pengguna menfess Twitter yang pernah melakukan transaksi pembelian setelah mendapatkan informasi dari akun menfess Twitter. Data kuantitatif diolah dengan metode Structural Equation Model Partial Least Square (PLS-SEM). Hasil penelitian menunjukkan bahwa content relevance, navigability, dan presentation memengaruhi perceived usability; perceived usability memengaruhi browsing activity dan social interaction; social interaction memengaruhi informativeness, baik secara langsung maupun dengan mediasi perceived interaction quality; informativeness memengaruhi perceived information quality; perceived information quality memengaruhi browsing activity dan perceived enjoyment; browsing activity dan perceived enjoyment memengaruhi urge to buy impulsively; dan urge to buy impulsively memengaruhi actual purchase secara langsung serta melalui cognitive evaluation. Penelitian ini melengkapi pengetahuan di bidang usability, social interaction, dan browsing activity pada praktik akun menfess Twitter. Hasil penelitian dapat dimanfaatkan oleh pengembang Twitter untuk mengembangkan Twitter sebagai media interaksi berbasis menfess.

---

Twitter is one of the largest text-based social media in the world with various features that support the communication activities of its users. Over time, users can send messages to menfess accounts according to their interests and preferences, which will then re-post the message publicly. This study aims to analyze the factors of usability, social interaction, and browsing activity that influence the impulsive buying behavior of Twitter menfess users. This research was conducted using mixed-method. Qualitative research was conducted by conducting interviews with 10 Twitter user interviewees who had interacted with the menfess Twitter account. The qualitative data that has been collected is processed using grounded theory. The results of qualitative data processing are used as a reference for the formation of quantitative instruments. Quantitative research was conducted by distributing online questionnaires with 918 respondents of Twitter menfess users who have made purchase transactions after getting information from Twitter menfess accounts. Quantitative data is processed using the Structural Equation Model Partial Least Square (PLS-SEM) method. The results showed that content relevance, navigability, and presentation affect perceived usability; perceived usability affects browsing activity and social interaction; social interaction affects informativeness, both directly and by mediating perceived interaction quality; informativeness affects perceived information quality; perceived information quality affects browsing activity and perceived enjoyment; browsing activity and perceived enjoyment affect urge to buy impulsively; and urge to buy impulsively affects actual purchase directly and through cognitive evaluation. This research complements knowledge in the fields of usability, social interaction, and browsing activity in the practice of Twitter menfess accounts. The results of the study can be utilized by Twitter developers to develop Twitter as a menfess-based interaction media.