PageRank Explained · Google’s Link Algorithm

PageRank is a link analysis algorithm created by Larry Page and Sergey Brin at Stanford University in 1996 as part of their research project that became Google. The algorithm ranks web pages by counting the number and quality of links pointing to a page on the assumption that more important websites are likely to receive more links from other sites.

The core insight: a link from one page to another is treated as a vote of confidence. Not all votes carry equal weight — a link from a highly-linked page passes more authority than a link from a page with no incoming links. PageRank quantifies this recursive relationship: a page is important if important pages link to it.

The original PageRank formula for a page A is:

PR(A) = (1 - d) + d × (PR(T1)/C(T1) + PR(T2)/C(T2) + ... + PR(Tn)/C(Tn))

Where:

• PR(A) is the PageRank of page A being calculated
• d is the damping factor — typically set to 0.85. Represents the probability that a random web surfer continues clicking links rather than starting fresh at a random page
• T1...Tn are pages that link to page A
• PR(Ti) is the PageRank of each linking page
• C(Ti) is the total number of outbound links from each linking page — the PageRank passed per link is divided equally among all links on the page

The formula is iterative — it must be calculated repeatedly across the entire web graph until values converge. In the original paper, Page and Brin ran the algorithm on a crawl of approximately 26 million pages. Modern Google runs an equivalent process across hundreds of billions of pages continuously.

The damping factor (d = 0.85) models user behaviour: an imaginary "random surfer" who clicks links with probability d and jumps to a completely random page with probability (1 - d). Without the damping factor, PageRank on a fully-connected graph would converge to equal values for all pages — the damping factor prevents this by injecting randomness.

The (1 - d) term, typically 0.15, ensures every page receives a minimum baseline PageRank even if no other pages link to it. This prevents sink nodes (pages with no outbound links) from absorbing all the PageRank in a graph.

Each page divides its PageRank equally among all pages it links to. A page with 100 PageRank units linking to 10 pages passes approximately 10 units to each linked page (before the damping factor adjustment). This has several practical implications:

• Fewer links per page = more equity per link. A page with 5 outbound links passes more equity per link than a page with 50 outbound links. Navigation-heavy pages with dozens of links dilute the equity passed to each destination.
• Internal linking redistributes external equity. External backlinks to your homepage flow through internal links to other pages. A homepage that links to a target page passes a portion of its accumulated external link equity to that page.
• Orphan pages receive zero PageRank. A page with no incoming internal links receives no PageRank from the site's link graph — regardless of how many external links the site overall has received.

The original PageRank algorithm described in the 1998 paper is not what Google uses today. Two decades of research, engineering, and adaptation to a web that has grown from millions to hundreds of billions of pages have produced a far more sophisticated system.

• TrustRank (2004). A Stanford/Yahoo research paper extended PageRank by seeding the calculation from a set of manually-verified trusted sites — propagating trust through the link graph rather than just authority. Google incorporated similar concepts into its spam-detection systems.
• Topic-sensitive PageRank. Separate PageRank calculations with different damping factors for different topic categories — a link from a relevant topical page passes more effective authority for that topic than a link from an unrelated page.
• Link spam resistance. The Penguin algorithm (2012) actively devalued links identified as manipulative — link farms, paid links, exact-match anchor text schemes. PageRank in the modern era counts only links Google considers genuine editorial endorsements.
• Retirement of Toolbar PageRank (2016). The public-facing PageRank score (0–10 scale visible in the Google Toolbar) was discontinued. The internal PageRank calculation continues but is no longer exposed publicly — eliminating the direct incentive to game the public score.

Google has confirmed that PageRank — in some form — remains part of its ranking systems. In a 2020 blog post, Google's Gary Illyes described PageRank as still one of the most important signals Google uses. However, it operates as one of hundreds of signals in a machine-learning ranking system, not as a standalone algorithm.

The practical implications for SEO remain consistent with the original algorithm's logic: backlinks from authoritative, relevant pages pass more ranking benefit than links from low-authority or irrelevant pages; internal linking distributes authority through your site structure; pages that receive no links — internal or external — receive no PageRank benefit.