Deutsch
Link value & metrics

PageRank

Also known as: Google PageRank, PR

PageRank is a method for scoring the importance of a web page from the links that point to it. Larry Page developed it at Stanford, and it was one of the ranking systems used when Google launched. Google still lists it among its core ranking systems.

How does PageRank work?

The patent “Method for node ranking in a linked database” (US 6,285,999) describes the idea. The application was filed on January 9, 1998 and the patent was granted on September 4, 2001, with Stanford University as the original assignee. Instead of judging a document by its own text alone, the method uses the links between documents. A link from an important document counts for more than a link from an unimportant one, and the abstract describes a probability factor that models random navigation across links.

PageRank is calculated per page. It is not a score for a whole website.

PageRank simplified: how links pass on valueFive pages A to E with links. Page C receives links from A, B, D and E and has the highest value. A receives links from C and D and is almost as large. A splits its value across its two links to B and C. B links to E with rel=nofollow, and this link is not counted in the model. D and E receive no counted links and keep only the base value. Circle sizes are calculated with the 1998 formula by Brin and Page and are for illustration only.nofollowA1.85B0.93C1.92D0.15E0.15Circle size = value in the model(illustrative, calculated forthis example)Link: a page’s value is splitacross its outgoing links.rel=nofollow: not counted inthe model. Google has treatednofollow as a hint for rankingsince September 2019, and forcrawling and indexing sinceMarch 1, 2020.Model after Brin/Page 1998, d = 0.85:PR(A) = (1 − d) + d · (PR(T1)/C(T1) + … + PR(Tn)/C(Tn))Google still uses PageRank in an evolved form. Its current weighting is not public.
Simplified PageRank model: a page passes on its value in equal parts through its outgoing links, and many or strong incoming links lead to a high value. Circle sizes are illustrative, calculated with the 1998 formula. Google uses PageRank in an evolved form; its current weighting is not public. Sources: Brin, Page (1998): The Anatomy of a Large-Scale Hypertextual Web Search Engine; Page, Brin, Motwani, Winograd (1999): The PageRank Citation Ranking; Google Search Central: A guide to Google Search ranking systems; Google Search Central: Qualify your outbound links to Google; Google Search Central Blog: Evolving “nofollow” (Sep 10, 2019).

What is the role of PageRank today?

Google’s guide to ranking systems puts PageRank under link analysis. It says Google has various systems that understand how pages link to each other, and that PageRank was one of its core ranking systems when Google first launched. According to the guide, how PageRank works has evolved a lot since then and it continues to be part of the core ranking systems. Google does not publish weights or other details.

Tools that show a “PageRank” number display their own estimates (assessment). Link strength scores such as Domain Rating come from the same family of ideas but are calculated by tool vendors.

Example

Page A is linked from three pages that each receive many links. Page B is linked from three pages that nobody links to. Under the PageRank idea, A ranks higher than B, although both have three links. The example illustrates the principle and is not a measured result.

FAQ

Is PageRank still a ranking factor? Google names PageRank as part of its core ranking systems and does not disclose its weight.

Can I look up the PageRank of a page? Google does not offer a lookup in its documentation. Third-party numbers are estimates.

Sources

  1. Method for node ranking in a linked database – US Patent 6,285,999patents.google.com
  2. Guide to Google Search ranking systems – Google Search Centraldevelopers.google.com
  3. How Google Search ranks results – Googlegoogle.com

Last updated: