The present study analyzed the use of conversation analysis (CA) in online interaction. The aim of the study is to describe the patterns of turn taking, repair, and the adjacency pairs in the online conversation. Employing a qualitative research methodology, data were taken from an online conversation about how to learn English, taking place in YouTube between a professor of English language teaching and the host who is a professor of education. The conversation was mainly in Bahasa Indonesia. The data were first transcribed in Bahasa Indonesia and translated into English. The study found some structures of exchange in the communication and the patterns of turn taking, repairs and adjacency pairs evident from the analysis of conversation. There are some patterns of turn-taking and adjacency pairs along the online conversation. The turn constructional units vary in syntactical forms. In addition, both self-initiated and other initiated-self repairs occurred. The study drew the implication of CA to language teaching. CA contributes to language teaching in terms of offering not only the authentic real-life communication, but also the authentic spoken interaction which will encourage learners to be able to produce authentic utterances.