This paper is concerned with the feature of segmentation, one of the key functionalities of computer-assisted translation software. Due to segmentation, the so-called CAT tools have been adopted globally for their ability to facilitate translation output, quality, and work pace. Despite various related benefits, there are areas that remain much under-researched, such as the effect segmentation itself may have on the translation process and translation production. According to some research, splitting up text and sentence-by-sentence workflow may lead to undesired results. Therefore, this paper attempts to summarize key issues raised by some scholars to date as well as test a hypothesis through a practical translation study.