What Is PageRank?

How PageRank Works

PageRank started from a simple observation about the web: a link from one page to another is a kind of endorsement. If page A links to page B, A is vouching for B. PageRank turns that intuition into math by asking a probability question — if a person started on a random page and kept clicking links forever, how often would they land on any given page? The pages they land on most are, by definition, the pages the web points to most, and those get the highest scores.

That “random surfer” is the mental model at the center of the algorithm. Authority is not counted, it flows. A page receives authority from every page linking to it, and it passes that authority along through its own outbound links. Because a page’s score depends on its linkers’ scores, which depend on their linkers, the calculation runs iteratively — over and over across the whole link graph — until the numbers stop changing and settle into a stable state.

Two consequences fall straight out of the model and still shape SEO thinking. First, a link’s value depends on the linking page’s own authority — one link from a widely-cited page can outweigh hundreds from obscure ones. Second, outbound links split authority — a page linking to fifty others passes each of them a smaller share than a page linking to five. This is why a single backlink from a strong, focused page is worth so much more than a link from a link-stuffed footer.

The PageRank Formula

The original paper expresses a page’s rank as:

PR(A) = (1 − d) + d ( PR(T1)/C(T1) + … + PR(Tn)/C(Tn) )

Reading it in plain terms: page A’s rank is a small baseline value plus a share of the rank of every page (T1…Tn) that links to it. Each linking page Tn contributes its own rank PR(Tn) divided by C(Tn), its total number of outbound links — so a page hands out its authority in equal slices across everything it links to. The term d is the damping factor, set to 0.85 in the paper, representing the probability the random surfer keeps clicking rather than teleporting to a random page. The remaining (1 − d) chance of a random jump keeps the whole system stable and guarantees the scores converge.

Example of PageRank

The definitive worked example is the algorithm’s own debut. In April 1998, Larry Page and Sergey Brin presented The Anatomy of a Large-Scale Hypertextual Web Search Engine at the Seventh International World Wide Web Conference in Brisbane, describing the prototype search engine — Google — they had built at Stanford. The companion paper, The PageRank Citation Ranking: Bringing Order to the Web, laid out the scoring model in detail. Together they made a claim that sounded audacious at the time: you could rank the entire web better by analyzing its link structure than by matching keywords on a page.

The specifics are what make it a real example rather than a story. The authors defined the random-surfer interpretation explicitly — the surfer “keeps clicking on links, never hitting back, but eventually gets bored and starts on another random page” — and pinned the damping factor at 0.85. They then ran the iterative computation across their crawled link graph and used the resulting scores to order search results. The approach worked so well that it became the foundation of Google Search, and the method is protected by a Stanford-assigned patent naming Larry Page as inventor.

The aftermath is the part every SEO lives with. Once links carried measurable ranking value, an entire economy grew up around acquiring them — and around gaming them. Google’s response shaped the modern link landscape: the nofollow attribute arrived in 2005 to let publishers pass a link without passing PageRank, and Google eventually retired the public Toolbar PageRank score in 2016 so people would stop treating a single green bar as the goal. The visible number is gone, but the mechanism it described — authority flowing through links — is exactly what every Domain Rating, URL Rating, and Domain Authority score on the market is still trying to approximate.

Frequently Asked Questions

The Bottom Line

PageRank is the founding insight of modern search: treat a link as a vote and weight each vote by the authority of the page casting it, then let that authority flow across the whole web until it settles. Google no longer publishes the score, but the principle — authoritative links pass authority — still underpins link-based ranking and every third-party authority metric built to approximate it.