Content Marketing Glossary

  • What is PageRank?

What is PageRank?

PageRank

PageRank is the algorithm developed by Larry Page and Sergey Brin at Stanford in 1996 that originally powered Google Search. It ranks web pages by treating hyperlinks as citations - a page with more incoming links from authoritative sources is deemed more authoritative itself, recursively. The algorithm was novel enough to become the foundation of Google, which launched in 1998 and within a decade made the company worth hundreds of billions.

Google retired the publicly-visible “PageRank score” browser toolbar in 2016. The underlying algorithm - or a descendant of it - reportedly still influences rankings internally, but the score is no longer exposed, and Google has repeatedly said the signal is now one of hundreds rather than the dominant one it was in the early 2000s.

How PageRank actually worked

The core idea: treat the web as a graph, pages as nodes, links as edges. A page inherits authority from the pages that link to it, weighted by how authoritative those linking pages are. Iterate the calculation until the scores converge.

Three properties of the original algorithm:

Link equity flows both outward and inward. A page passes some of its authority along each outbound link. Linking to many pages dilutes the authority each receives.

Links from high-PageRank pages are worth exponentially more. One link from a page with 100x the authority of another is worth 100x more, not 2x.

The “damping factor” (typically 0.85) models random-surfer behaviour. The theoretical user doesn’t follow links indefinitely; eventually they teleport to a random page. This prevents authority from accumulating infinitely in tightly-linked clusters.

The paper Page and Brin published in 1998 (“The Anatomy of a Large-Scale Hypertextual Web Search Engine”) remains the clearest exposition. It’s short, readable, and one of the more consequential academic papers of the past 30 years.

Why PageRank alone stopped working

By the mid-2000s, SEO-industrial link farms were gaming PageRank at scale. Paid link networks, automated comment spam, wiki vandalism, reciprocal-link exchanges, and subdomain link-pumping all targeted the algorithm directly. Google responded with a cascade of updates:

Florida (2003). First major anti-spam update.

Penguin (2012). Demoted sites with manipulative backlink profiles. Affected 3.1% of search queries - a huge footprint.

Panda (2011), then ongoing core updates. Shifted ranking weight toward content quality, user engagement, and site-level trust signals rather than pure link counts.

By 2016, when Google killed the public toolbar score, the metric had become misleading even for legitimate SEOs. The number was updated sporadically, didn’t reflect internal Google scoring, and was actively harmful as a strategic focus.

What replaced PageRank (and what’s left of it)

Three notes on the current state:

Google still uses a PageRank-descended signal. Google has confirmed this multiple times. The modern version is more sophisticated - topical, weighted by trust signals, spam-filtered - but the underlying “links as citations, with weighted authority” intuition is still in the algorithm.

Third-party metrics approximate it. Moz’s Domain Authority (DA), Ahrefs’ Domain Rating (DR), Majestic’s Trust Flow and Citation Flow. Each uses a proprietary algorithm on their own crawl data. Useful for directional work; none is used by Google.

The E-E-A-T overlay. Google’s Experience / Expertise / Authoritativeness / Trustworthiness framework is less a direct-ranking mechanism and more a quality-rater guideline. But it shapes the broader algorithmic direction - authority is no longer purely link-derived; it’s also content-derived and signal-derived.

What PageRank means for Penfriend’s approach

Penfriend doesn’t build links. Link-building is off-page work - outreach, PR, contributor programmes, genuinely useful content that earns citations. We don’t do any of that directly.

What we do: produce on-page content that’s good enough to earn links. PageRank - and its modern descendants - only matters if the content at the destination deserves to be cited. A site full of thin, generic pages won’t earn links regardless of SEO tactics; a site with substantive, well-researched, opinionated long-form content will. The output we aim for is content worth linking to. Whether anyone actually links is downstream of that.

The honest diagnostic: if your site is producing 40 pages a month and none of them are getting cited externally, the problem is usually quality, not quantity. More content doesn’t fix that; different content does.

An example

A marketing-consulting firm had been optimising aggressively for Domain Authority for three years. They bought guest posts, ran reciprocal-link campaigns with peer agencies, and celebrated every DA point gained. DA climbed from 18 to 46 over three years.

Organic traffic during the same period: grew 12%. Meaningful pipeline from organic: grew 3%. The DA metric had been going up; the business value had not.

A new head of marketing audited the link profile. Most of the “high-DA” incoming links were from shared guest-post networks - technically real, practically discounted by Google’s spam filters. She killed the link-buying programme, redirected the budget into original industry research and contributor placements in recognised publications. DA flattened, then slowly grew to 52 over the next 18 months. Organic traffic grew 140% in that window. Pipeline from organic doubled.

The DA metric had been the symptom the firm was optimising for. The underlying cause - actual authority earned through quality work - was what finally moved the business.

Related terms