The final possible answer to why Relevance Feedback isn't conspicuous on the Web deals with the very structure of the Web itself. One of the motivating reasons behind Relevance Feedback is that the only way to get to a document is by querying the retrieval system. On the Web, this is not the case. Users are able to travel from document to document via the hyperlinks contained within. Since documents connected via hyperlinks are generally related in some way, it is a valid conjecture that a relevant document will have hyperlinks to other relevant documents. Some studies have even proposed using the number of hyperlinks pointing to a particular page as a method of augmenting document ranking functions [27, ]. Thus, another open question is whether users are better served by the traditional Relevance Feedback techniques surveyed in this paper versus exploiting the hyperlinks found in the initial set of relevant documents returned via a conventional IR system.