PageRank

PageRank is the original algorithm at the heart of Google Search, named after Larry Page and described in the 1998 Stanford paper "The Anatomy of a Large-Scale Hypertextual Web Search Engine." It assigns each page on the web a numeric importance score based on the structure of the link graph: a page is important if other important pages link to it.

The mathematical formulation models a random surfer who follows links at random with some probability and teleports to a random page otherwise. PageRank is the stationary distribution of that surfer — the long-run probability of finding them on any given page. This is the foundation that every modern third-party authority metric, from Domain Authority to Domain Rating to Trust Flow, approximates in its own way.

Google retired the public Toolbar PageRank display in 2016. The internal signal did not go away — it continued to be used and refined, and Google has confirmed as much in court filings and engineer interviews. What changed is that the public no longer has access to a per-page PR number, which is why third-party proxies became the standard operator tools.

Several practical consequences fall out of how PageRank works. First, links from strong pages carry more weight than links from weak pages — a single editorial mention from a top news site beats hundreds of low-tier directory listings. Second, link equity flows recursively: a page can be strong without direct inbound links if pages internally linking to it are themselves strong, which is why internal-linking architecture matters. Third, the dampening factor (typically 0.85 in the original paper) means equity dissipates with every hop, so deep pages many clicks from the homepage receive less internal PR than shallow pages.

Modern Google ranking blends PageRank with hundreds of other signals — content relevance, E-E-A-T, freshness, intent match. But the link graph remains a load-bearing input, and understanding PageRank's logic is still the right mental model for how authority and Link Equity propagate.