The concept of UX smells has been recently studied as a systematic way to detect predefined user interaction issues and fix them with cataloged solutions. Most of the existing literature about UX smells focuses on desktop web applications, while there are only a few works ad- dressing the mobile web. Although specific UX smells for mobile interac- tions have been proposed, there are no objective evaluations to determine their impact on the perceived UX. In this work, we evaluated 6 mobile UX smells (3 from the literature and 3 new proposals) with respect to efficiency in use. We conducted an online evaluation with 72 participants in 3 real websites, each one with a set of specific mobile UX Smells. In this evaluation, we compared each website to a refactored version of it- self, i.e. with proposed fixes for each of the smells. To do this, we ran a between-subject experiment in which participants completed 10 every- day tasks on the websites while we measured their efficiency in terms of task completion time and number of user interaction events. As a comple- mentary post-hoc analysis, we also grouped temporally close interaction events into interaction bursts, providing an additional efficiency-related perspective. All the captured metrics were compared in the default ver- sion of the websites vs. their refactored counterparts. Results showed that in most cases (15/20), either the time to complete the task or the amount of interaction events were higher in the presence of UX smells. Moreover, in 7 of the cases, the observed differences were statistically significant (p<0.05). The burst-based analysis was consistent with these trends.