This research focuses on the information overload problem on the Internet, and proposes a potential remedy to the overload encountered while searching the Web. We developed a system that makes use of clustering and visualization for browsing the results of a typical web search. We built two different (full and fisheye) zooming capabilities into our system, and empirically compared their success with each other as well as with the traditional non-visual presentation method through an experiment. We hypothesize that the visual systems will lead to higher success than the text-based system, and that the fisheye zooming system will lead to higher success than the full zoom system. The results of our exploratory test provide partial support for our hypotheses. This empirical support and the comments made by the participants in the experiments suggest that our design ideas are promising, and it is worthwhile to further investigate the use of clustering and visualization mechanisms for reducing information overload.