Our research objectives are to provide a theoretical discussion on how software may impact user performance in ways contrary to designers’ intentions and users’ desires, and to empirically evaluate user performance impacts that derive from ostensibly performance-enhancing software features. We propose that dyadic procedure is associated with higher levels of user performance when compared to monadic procedure. Using word-processing software utilization as the research context, we test the proposition on data from 46 participants. Contrary to expectations, the results suggest that dyadic procedure may decrease the accuracy of users’ work. We conclude that software design features that are intended to improve user performance may have opposite effects, which raise questions about these features’ utility and desirability.