PageRank History: From the 1998 Paper to the Toolbar’s Removal

The story of PageRank is really two stories. One is about an idea that made Google. The other is about a visible number that the web turned into a target, and what Google had to do about it.

The second story is the one that still matters for site owners, because the same mistake is being made today with different numbers.

Before Google: where the idea came from

Judging importance by references was not new. Academic citation analysis, which counts how often papers cite each other, dates back to the 1950s. In 1976 researchers proposed weighting citations by the importance of the citing journal, the same core idea PageRank would apply to the web.

Google was not alone either. Robin Li built a link-based ranking method for a search engine called RankDex in the mid-1990s and patented it. He later founded Baidu.

1996 to 2001: Stanford and the patent

• 1996: Larry Page and Sergey Brin begin the research project at Stanford that becomes Google, under the early name BackRub.
• January 1998: the PageRank patent application is filed.
• 1998: Brin and Page publish their paper describing the Google search engine and PageRank.
• 4 September 1998: Google is incorporated.
• 4 September 2001: the patent is granted.

The patent was never Google’s. It was assigned to Stanford University, with Google holding an exclusive licence. In return Stanford received 1.8 million Google shares, which it sold in 2005 for 336 million dollars. It remains one of the most valuable technology licences any university has made. The patent has since expired.

2000: the score goes public

On 11 December 2000 Google released its Toolbar, which showed a PageRank score from 0 to 10 for any page you visited. Earlier that year Google had already sorted its web directory by PageRank.

The effect was immediate. A number that anyone could see became a number everyone wanted to raise. A market grew up selling links from high-scoring pages, and scores became a selling point for websites themselves.

Why the public score was so easy to game

• It updated only a few times a year. Sellers could place links shortly before an update and advertise the higher score for months afterwards.
• It had only eleven possible values. Most ordinary pages sat in the same narrow band, so a small change looked significant.
• Google could not see private deals. A paid link looked the same as an earned one unless someone disclosed it.
• Nofollow relied on honesty. It only worked if the people selling links chose to use it.

The score measured something real, but displaying it changed how people behaved. That is the core reason it eventually had to go.

2005 to 2012: fighting the market it created

• 2005: the nofollow attribute arrives, letting sites mark links that should not pass value, mainly to stop comment spam.
• 2007: Google begins visibly lowering the toolbar scores of sites caught selling links.
• 2009: Google stops showing PageRank in Webmaster Tools, and confirms that using nofollow to sculpt internal PageRank no longer works.
• April 2012: Penguin launches to demote sites built on manipulative links.

Penguin’s full story is covered in the Penguin update.

2013 to 2016: the score fades out

On 6 December 2013 Google updated toolbar PageRank for the last time. For more than two years the displayed scores sat frozen and increasingly meaningless.

On 7 March 2016 Google removed the score entirely. Google’s John Mueller explained that a single isolated number had become less useful. The Toolbar itself was finally shut down in December 2021.

2016 to today: invisible but present

• September 2016: Penguin becomes part of core ranking and starts devaluing spammy links rather than demoting whole sites.
• 2017: Google’s Gary Illyes confirms PageRank is still used, among hundreds of other signals.
• 2019: nofollow becomes a hint, joined by the sponsored and ugc attributes.
• November 2022: Google’s new ranking systems guide lists PageRank among its active systems.
• December 2025: the latest version of that guide still does.

How the system works today is covered in what is PageRank, and does Google still use it.

Myths the toolbar era left behind

Several beliefs from that period still circulate, long after Google addressed them.

• Redirects lose PageRank. In 2016 Google’s Gary Illyes said redirects no longer lose PageRank, which is why permanent redirects are safe in a migration.
• Links from .edu and .gov sites carry special weight. Google has said these domains get no special treatment; their value comes from the same signals as any other link.
• Nofollow links pass nothing at all. Since 2019 they are hints that Google may use, though they are still not a route to manipulate rankings.
• A higher score means higher rankings. It never did on its own, which is exactly why Google stopped showing it.

The timeline at a glance

Date	What happened
1996	Research begins at Stanford
1998	Patent filed, paper published, Google incorporated
2000	PageRank shown publicly in the Toolbar
2005	Nofollow introduced; Stanford sells its shares
2009	Removed from Webmaster Tools
2012	Penguin launches
2013	Last public score update
2016	Public score removed; Penguin joins core
2025	Still listed as a core ranking system

What the history teaches

Any visible score becomes a target. Once PageRank could be seen, people optimised for the number instead of for what it was meant to measure, and Google spent the next decade undoing the damage.

The same thing is happening now with third-party scores such as Domain Rating and Domain Authority. They are useful estimates, covered in why you can’t check PageRank any more, but Google does not use them, and buying links to raise them repeats the exact pattern that led to Penguin. If you suspect bought links are holding your site back, see link penalty recovery.

Where PageRank fits among Google’s other systems is covered in Google’s ranking systems explained.