Discourse Marker (DM) So is essential for participants to understand an ongoing dialogue. Despite the important role, the lack of effort discloses the academic context of spoken monologic discourse. This research investigated the effects and functions of DM So in monologic data of spoken corpus of academic lectures. The MICASE (Michigan Corpus of Academic Spoken English) data was annotated with the help of the MonoConc corpus program. Selected frameworks were used to examine the data to show their original functionality. The consistency was achieved through coding agreement between coders .82, showing the reliability of its findings. The results presented seven functions of DM So in the lecture corpus, namely topic developer (TD), point sequencer (PS), main idea unit (MIU), exemplification, inferential conclusion, and rewording. These included three newly established functions because they indicated different meanings from those in frameworks, namely PS, exemplification, and rewording. Therefore, this present research suggested that the use of DM So is considered to have a key role in signaling the lectures’ macrostructure to facilitate the student-listeners to process the given information.