Search engines used to use just the text on a web page in order to rank a web page. However web publishers began adding keyword phrases in a way to help rank their pages. Not all pages that ranked were relevant, resulting in a poor user experience. Information retrieval researchers began work on new ways to identify relevant web pages, particularly with links. Search engineers noticed that the best pages on the Internet tended to accumulate links. The more links a site had the more important it tended to be. Conversely, the less links a site accumulated the less important it was judged to be.