A method for determining which web page among multiple candidate web pages is similar to a given web page. For each candidate web page, a set of scoring rules is provided to score the components therein. When the given web page is compared against a candidate web page, each component that is found in both the given web page and the candidate web page under examination is given a score in accordance with the set of scoring rules that is specific to that web page under examination. A composite similarity score is computed for each comparison between the given webpage and a candidate web page. If the composite similarity score exceeds a predefined threshold value for a comparison between the given webpage and a candidate web page, that candidate web page is deemed the web page that is similar.