With the boom of online learning, conflicting findings on this topic have been yielded. This work seeks to answer the core research question: During synchronous online discussion on Tencent Meeting, how do different modes of discussion (turn‐taking and free discussion) and the degree of intimacy (friends and strangers) affect undergraduate students' discussion efficiency? Eventually, the data of 32 participants are valid, with four groups for one mode. To evaluate their efficiency of online discussion, this study combines a more subjective content analysis with the exported results from the tests. As revealed by the final results, although they are not supported by all hypotheses, there are notable interaction effects presented. Sum of words in turn‐taking discussions between strangers are much more than that occur in free discussions whereas two different modes (free discussion and turn taking) do not matter for most of the variables for strangers. Moreover, among the 5 elements of content analysis, participation across groups appears slightly statistically significant, indicating that that friends discussing in a free discussing condition produce more swap of turns. Familiarity of the topic, as a confounding variable, has a significant impact on participants score. Combined, the results of this study have educational implications, serving as a reference for teachers designing teaching strategies. Lastly, some limitations of this study are discussed and future research directions are suggested.