Google PageRank: What It Actually Measures Today

Google PageRank is the algorithm Larry Page and Sergey Brin built at Stanford to answer one question: which pages on the web deserve more trust than others, based purely on how other pages link to them. It doesn't read your content, check your keyword usage, or care how fast your site loads. It counts and weighs incoming links, treating each one as a vote, but not an equal vote. A link from a page that itself has many trusted incoming links carries far more weight than a link from an obscure page nobody links to.

The public-facing version of this, the 0-to-10 score millions of SEOs once tracked through the Google Toolbar, is gone. Google froze the underlying data years before officially retiring it in 2016. But the concept it was named after never disappeared; it evolved into the link-based authority signals that still influence rankings today, folded into a system of hundreds of other factors. Understanding the actual mechanism, not the old toolbar number, explains why some internal linking decisions and some outreach targets matter far more than others.

How the Original Algorithm Actually Works

Picture a person clicking links completely at random, forever, with no destination in mind. Every so often, roughly 15% of the time in the original formula, instead of following a link they teleport to a totally random page on the web. That probability is called the damping factor, usually set around 0.85. PageRank calculates, for any given page, the likelihood that this random surfer lands there after an enormous number of clicks. Pages that many other well-visited pages link to end up with a higher probability of a visit, and therefore a higher score.

The formula also divides each page's score across everything it links to. A page worth 100 units that links out to 10 other pages passes roughly 10 units to each of them, before the damping factor discounts it further. The same page linking to only 2 others passes far more authority to each one. This is why the number of outbound links on a page matters almost as much as the number of inbound links pointing at it, since a page's own authority gets diluted the more places it sends votes.

A simplified example makes this concrete. Imagine three pages: A, B, and C. Page A has no incoming links and links out to both B and C. Page B has one incoming link, from A, and links to C. Page C has incoming links from both A and B, and links to nothing else. Even before running the actual formula, it's clear C ends up with the highest score: it receives a share of A's authority directly, plus a share of B's authority, which itself came partly from A. This compounding is why authority concentrates on pages sitting at the receiving end of multiple paths through a link graph, rather than pages that merely have one or two links pointing at them.

A Brief History: From Patent to Toolbar to Retirement

Page and Brin described the algorithm in a 1998 Stanford paper, and Google built its earliest ranking advantage on it. Around 2000, Google exposed a simplified 0-to-10 version of the score through the Google Toolbar, and it quickly became an obsession. SEOs tracked it, link brokers priced placements by it, and entire sites existed to sell high-PageRank backlinks. Google grew uncomfortable watching an industry build a market around gaming a number it had published itself.

For a period in the mid-2000s, a visible link market grew around the score: brokers sold placements on pages by PageRank tier, plugins promised to sculpt PageRank around a site's own navigation, and even the resale value of expired domains was partly set by whatever PageRank the domain had accumulated before it lapsed. Google's own guidelines began explicitly warning against link schemes built purely to manipulate the score during this same period, which is a large part of why the number's public visibility eventually became more liability than asset for the company.

The toolbar score was effectively frozen around 2013, still displaying but no longer updating with any regularity, and Google officially discontinued it in 2016. Google has said many times since that link analysis remains part of ranking, but it has never republished a visible score. Any browser extension or website today claiming to show your PageRank is either showing stale, years-old data, or calculating an entirely different proprietary metric that happens to share the name.

Misconceptions Worth Retiring

A few ideas about PageRank keep circulating in SEO forums and client meetings long after they stopped being accurate.

• "PageRank checker" tools show Google's real score. They show frozen pre-2013 data, or a completely different proprietary metric that happens to share the name.
• A high PageRank guarantees a high ranking. Even in the 1998 paper, PageRank was one input among several; today it is a small ingredient in a system built around relevance and content quality first.
• Reciprocal link exchanges build PageRank efficiently. Google's systems long ago learned to discount obviously negotiated, reciprocal link patterns; the original math assumed a naturally formed web, not a bartered one.
• Buying links is a shortcut to authority. It is a shortcut to a manual action if Google's systems catch it, and the link scheme guidelines exist specifically to catch this pattern.

Where This Kind of Equity Flows Inside Your Own Site

Because authority passes through links, and your homepage almost always accumulates the most external links, since people link to a brand's front door far more often than to a blog post three folders deep, your internal linking structure decides how much of that accumulated authority ever reaches the rest of your site. A page with no internal links pointing to it, an orphan page, receives essentially none of it, no matter how much authority the site carries overall. Search engines can still discover such a page through an XML sitemap, but it starts with almost nothing behind it compared with a page linked from your navigation or from a dozen related articles.

This is why click depth, how many clicks a page sits from the homepage, correlates so strongly with how well pages tend to perform. A hub-and-spoke structure, where a central pillar page links out to every related article and those articles link back to the hub and to each other, keeps authority circulating instead of pooling in one place and dead-ending elsewhere.

• Audit for orphan pages, meaning content with no internal links pointing to it at all.
• Link into the pages you actually want to rank from your homepage and your other most-linked pages.
• Avoid linking every page to every other page indiscriminately; that spreads authority thin without concentrating it anywhere useful.

Nofollow, Sponsored, and UGC: How Attributes Change the Flow

Early SEOs discovered that because a page's authority is divided among everything it links to, you could try to sculpt internal PageRank by adding a nofollow attribute to unimportant links, a footer Terms of Service link, for instance, on the theory that it would concentrate more authority on whatever was left untagged. Google closed this loophole around 2009: a nofollowed link's share of authority is simply lost rather than redistributed to the remaining links on the page. Sculpting internal navigation with nofollow to save PageRank stopped working over a decade ago; it now just tells search engines which links you don't vouch for.

In 2019, Google introduced a sponsored attribute for paid or affiliate links and a ugc attribute for user-generated content like forum posts and comments, alongside the original nofollow attribute. Since 2020, Google has treated all three as hints rather than absolute directives, meaning a tagged link can occasionally still be considered for ranking purposes. The practical guidance hasn't changed much: tag paid links to stay compliant with Google's link scheme policies, and don't expect attribute tagging to meaningfully redirect authority around your own site.

How Link Authority Fits Into Ranking Today

Link-based authority is layered under, not instead of, relevance. Google's publicly available Search Quality Rater Guidelines describe evaluators judging pages on expertise, experience, authoritativeness, and trustworthiness, often summarized as E-E-A-T, alongside how well a page satisfies the specific query behind a search. A page can carry substantial link-based authority and still fail to rank if it doesn't match what the searcher is actually looking for; the reverse is also true, since a page with modest link authority can outrank a much more heavily linked competitor if it answers the query far more precisely.

This is the practical reason link-building advice shifted over the past decade from raw volume toward relevance and context. A link from a well-linked page in an unrelated industry still passes some authority under the original mechanics, but it does very little to establish that your page is a good answer for a specific query, which is the layer that now does most of the work in deciding what actually ranks.

What This Means for SEO Strategy Today

Chasing link volume without regard to quality is chasing a number that stopped mattering the way it once did. A single link from a page that itself sits inside a trusted, well-linked domain, in a topically relevant context, generally does more for a page's authority than a dozen links from unrelated, low-authority sites. That isn't a rule of thumb pulled from nowhere; it's a direct, if simplified, description of how the underlying weighting behaves, since the random surfer is simply far more likely to arrive by way of the trusted page.

In practice, that means outreach and content strategy should target pages that are themselves well-linked and topically relevant, rather than volume-based directory submissions or link exchanges. It also means your internal architecture is a lever entirely within your control, with no outreach required, by making sure the pages you most want to rank are linked to from the pages on your site that already carry the most weight.