Organizing Suggestions in Autocompletion Interfaces

Even though Autocompletion is widely used in search interfaces with different variations, few guidelines exist on how to present autocompletion suggestions. In this paper, we describe two user studies that shows some types of organization strategies help users search faster and easier in a known-item search task. We studied the effect of grouping suggestions in an autocompletion interface to select terms taken from a thesaurus. In the first study, we explored ways of grouping term suggestions from two different thesauri (TGN and WordNet). The results of the first study suggested that the best organization strategies are different when using different thesaurus. Users found Group organization may be appropriate to organize location names from TGN, while Alphabetical works better for WordNet. We then carried out a follow-up study, where we compared three different organization strategies (Alphabetical, Group and Composite) for location name search tasks. The results of the second study indicate that in general, autocompletion interfaces help improve the quality of keyword. We also found that Group and Composite organization help users search faster, and users perceive the suggestion organization as easier to understand and use than the Alphabetical. 2000 Mathematics Subject Classification: 1998 ACM Computing Classification System: H.1.2Human factors