PageRank.

PageRank was developed at Stanford University in the late 1990s and was one of the core ideas with which Google set itself apart from other search engines. Instead of judging pages only by the occurrence of words, Google looked at the structure of the web: a page that many other pages link to is probably important. A page that important pages link to is even more important.

Mathematically, PageRank describes the probability that a random surfer clicking on links will land on a particular page. Each page distributes its own value across the pages it links to; the more outgoing links a page has, the smaller the share per link. A damping factor takes into account that the surfer stops clicking at some point and starts again.

For a long time, Google displayed a public PageRank value from 0 to 10 in a browser toolbar. This led to a trade in links, with pages sold according to their PageRank. Google discontinued the display in 2016. Internally, though, Google continues to use link signals; Google employees have confirmed that PageRank, in an evolved form, remains one of many signals. Commercial metrics such as Domain Authority attempt to replicate this value, but they are the tool providers' own estimates.

Two lessons from the PageRank principle remain relevant in practice. First: links from respected, thematically relevant pages are worth more than many links from insignificant pages. Second: internal links distribute weight too. Pages that many links point to within your own website signal their importance to Google; pages without an internal link are hard for crawlers to find and receive hardly any weight.