How did Google decide which web page deserves to come first?

Google treated every link as a vote, and weighed each vote by who cast it. PageRank, the algorithm Google Search first used to rank pages, counts the number and quality of links to a page, on the assumption that more important websites are likely to receive more links from other websites. A hyperlink counts as a vote of support, and a page linked to by many pages with high PageRank gets a high rank itself.

That definition is recursive: a page's rank depends on the rank of the pages that link to it. The circle is broken by repetition: the computation makes several passes through the collection, called iterations, each bringing the approximate values closer to the true ones. Another way to see it is an imaginary surfer who clicks links at random and sometimes jumps to a random page instead: PageRank gives the likelihood that this surfer arrives at any particular page.

PageRank was developed at Stanford in 1996 by Larry Page, whose surname it shares, and Sergey Brin, as a research project for a new kind of search engine, different from the then-dominant AltaVista. Brin's idea was that information on the web could be ordered by "link popularity". The trick was borrowed: PageRank was influenced by citation analysis, developed by Eugene Garfield in the 1950s. Today PageRank is not the only algorithm Google Search uses to order results, but it was the company's first and it is the best known.