While fluency in second language speaking can be a challenging construct to measure, it is important to identify the discourse features that contribute to it. This small-scale classroom research project examined the effect of formulaic language sequences on fluency as measured by computer-based speaking tasks of young English learners. Thirty-six speech samples were collected as a part of standard instruction in grades 5-8 in a medium-sized public school district in the Southeastern US. The speech samples were analyzed using Praat speech analysis software to identify the mean length of fluent run for fluency and coded formulaic language sequences for discourse function. Findings indicated that the use of formulaic sequences is a significant predictor of fluency in the data set (p = 0.015) and that the most frequently used formulaic sequences were those used for clarification and to compare and contrast. Finally, the article discusses pedagogical implications for second language instruction, specifically for improving fluency on standardized computer-based speaking assessments.