The system that for years decided which pages topped Google's results rests on a simple idea: a link from one site is a vote for another. In 1998, Larry Page filed the patent "Method for node ranking in a linked database", known as PageRank, assigned to Stanford University. Recent reports speak of the patent expiring, but the available sources do not confirm this: they only give the filing date and the assignment to the university.

Psst!Add AlexTech.ai asPreferred Source on Google

From doctoral research to a search engine

In 1996, Page and Sergey Brin, PhD students at Stanford, built a crawler called BackRub to map inbound links. Page saw that treating every link as a vote had a limit: a link from a personal blog should not weigh as much as one from a major portal. He therefore introduced a recursive model based on a "random surfer" who clicks links at random until reaching a dead end.

official image of Google Search's foundational PageRank patent has expired. What will be the impact

Google's New Logo

In 1998, Page, Brin, Rajeev Motwani and Terry Winograd published "The PageRank Citation Ranking: Bringing Order to the Web". The paper described a way to give each page a global importance score computed from the link structure of the web. According to its abstract, the method measures the attention devoted to pages objectively and mechanically, and the authors showed how to compute it over large numbers of pages.

The search engine emerged almost by accident. Comparing results ranked by importance with those of engines of the time, such as AltaVista and Lycos, which mainly counted keywords, the two researchers found the link-based approach far better. The prototype was made available to the Stanford community, and the name Google, derived from "googol", was registered in September 1997.

The first phase: nobody wanted to buy

In 1998, Page and Brin offered the technology to portals such as Yahoo!, AltaVista and Excite, asking between 750,000 and one million dollars. The offers were rejected: portals wanted to keep users on their pages to show ads, while Google's engine quickly led users to the answer they sought. In August 1998, Andy Bechtolsheim, co-founder of Sun Microsystems, wrote a $100,000 check after a demonstration. The check was made out to "Google Inc.", a company that did not yet legally exist; it was incorporated on 4 September 1998.

The system worked well while the web was relatively clean. In 2000, Google made PageRank visible through a toolbar scoring pages from PR0 to PR10, and the result was an industry built on buying and trading links. In the following years, link spam, link farms and automated comments forced Google to correct the method, introducing the nofollow attribute in 2005 and a trust system based on sites considered reliable.

official image of Google Search's foundational PageRank patent has expired. What will be the impact

Image: azweb.com.vn — https://azweb.com.vn/pagerank-la-gi/

In 2011 came Panda, which penalised thin or copied content, and in 2012 Penguin, which targeted artificial link profiles. The public PageRank toolbar was shut down in 2016. Since then, Google has paired link counting with machine learning models such as RankBrain and BERT, which evaluate a page's meaning and the user's intent.

What changes now

PageRank is no longer the sole criterion, but it remains one of the founding ideas of web search research. Academic work building on it shows that the method has been extended to other graphs, and variants such as non-backtracking PageRank or personalised PageRank are under study. For those working on site visibility, the most concrete point remains this: a link counts not for its quantity but for the quality of the source offering it.