Recently, some have questioned the effectiveness of user-generated tags on several grounds, one of which being its lack of structure. To explore this issue, our study conducts an experiment to investigate whether user-generated tags can be enhanced with a faceted structure particularly on book tagging. This study aims to understand the effect of different tagging modes on the resulting tag sets, particularly in the context of fiction and non-fiction works. Two different kinds of tagging interfaces (with and without faceted template) and two different genres of works (fiction and non-fiction) are manipulated in this experiment. Participants’ tagging behaviors, including tags used, time spent were logged; and their perceptions with the interface was captured by the questionnaires. According to the results of assessment, it was found that the tag sets of faceted template display more distinct tags, more number of assigned tags on average, higher degree of tag similarity, and higher convergence of tags. While the results suggest that the faceted interface generated tags of better quality, it also incurred more user effort. Although it is hard to make clear conclusions based on one single study, the data suggested the usefulness of a faceted template as it tends to generate tags with higher viewpoint exhaustivity as well as consensus. Nevertheless, the actual retrieval effectiveness of the combination of tagging and faceted structure still has to be examined and assessed in further empirical research.