Google Search is being routinely gamed by fake blogs and has been for years

Since its earliest days, the core of Google’s search algorithm - and its greatest innovation - has been ranking a page by the amount and quality of the links pointing to it.

For almost as long, businesses have fought to climb to the lucrative top positions on search result pages, with some resorting to unethical methods which, if noticed, could see Google penalise their sites, or drop them entirely.

Despite Google’s efforts, for many years businesses have been paying to have the internet flooded with hundreds of fake links to their sites which, if done correctly, can be impossible to detect.

It’s called a PBN and it’s here to stay

A private blog network, or PBN, is a constellation of seemingly legitimate sites all linking to a page that a business with questionable ethics wants to promote. Businesses can set these up themselves, or they can choose from an increasingly sophisticated set of cheating-as-a-service providers, who’ll happily charge hundreds of dollars in return for setting up fake links.

PBNs are pervasive. There are forums dedicated to them with millions of views, businesses specialising in providing hosting just for PBNs, and some offering to do the complete process: wire them money and they’ll point links from their networks to your site.

Judging from the posts in black hat SEO forums, there are thousands of purchasers routinely receiving links in exchange for hundreds of dollars.

A prominent marketplace for buying and selling money making websites, known for being a source of trustworthy, legitimate businesses, goes so far as to prominently display whether or not a business sale comes with a PBN included. Many do, and these businesses sell for figures in the tens of thousands.

How is a link from a fake blog valuable?

But what value is a link from a fake blog? Google determines the value of a link based on the reputation of the site, and that reputation is calculated by, as you might expect, the quality and amount of links it has coming in.

It’s a bit like if links were rivers, and sites were lakes. The more water coming into a lake, the more it has to send out to others. The New York Times is a massive lake, with 714 million links (really) which bestows upon it the power to greatly enhance the profile of any site it links to, propelling the lucky recipient up the results page.

One would assume a fake blog would be an empty basin, with a very unhealthy 0 links. But that’s not the case.

How to fake a network

With advertising dollars fleeing the news sites for the greener pastures of Google and Facebook, media companies are dropping like flies. When a site goes out of business their domain name registration might lapse and become available again for purchase.

These dropped domains, as they’re called, come with a very valuable commodity: their links. All an unscrupulous marketer has to do is snap it up and replace the content pages, then drop a link or two to the site they’re trying to promote. Dropped domains with lots of links are quickly bought just as soon as they drop, and dozens of sites have sprung up to help marketers snatch them.

But what about the content? After all, apart from links, content is the other crucial ranking factor. Easy. You can pay someone to write a decent article for the price of a meal or two. Or you can use AI-infused tools to “spin” articles - a term used to refer to modifying someone else’s content enough that the search engines think it’s unique. They do this with varying levels of success, and some services offer to make manual tweaks to the automatically altered articles to improve their readability or uniqueness.

Also, there’s evidence that the amount of domains - not just the amount of links - to a page affects search rank. Meaning that even these low quality links help, just by virtue of coming from a variety of domains.

Massive industry

Businesses offering PBN services and their customers are commonplace. On a well known and very popular black-hat SEO forum, a post entitled “❇️ GROUND BREAKING PBN PACKAGE ❇️ AGED AUCTION DOMAINS❇️ HIGH PA/DA BLOGS ❇️ 50% OFF” has 4,000 replies and 490,000 views. That’s just one thread on one site, and there are hundreds, if not thousands, more.

These aren’t just shoddy, poorly-written, obviously fake links being bought - after all, these people are in an arms race with all the Silicon Valley human and artificial intelligence of Google itself. This is a mature industry and ecosystem, all built around faking the web and fooling Google Search.

Every package is advertised with just how they pull the wool over Google’s eyes, proudly proclaiming that their servers are in different countries, their domains aged many years with unique whois records, their sites using unique software and themes. They even promise to drip feed the links so everything seems above board, over time.

Let’s look at a real PBN

I’m hesitant to link directly to a PBN or mention one lest Google smite us, but you can see an example of someone complaining to Google about one here (click the “more” link to see the links pointing to them): https://support.google.com/websearch/thread/4239775?hl=en.

A quick examination of the links to the site shows there are many that are extremely suspicious. Some links appear suddenly in unrelated articles, which should be easy enough for Google to detect. But others appear on sites where they don’t look totally out of place.

The site appears to have been snapped up as a dropped domain - there are lots of powerful links from the likes of PBS and Wikipedia to what appears to have been a legitimate site with music reviews and interviews many years ago. Someone who follows these links these days would find a 404 page littered with links like “Top 10 Dynamic Vocal Mics for Live & Home Recording” - easily monetizable search engine targeted content.

They also have a spammy link as a citation from Wikipedia where an article about the delay audio effect unit tangentially mentions “additionally, several companies make new analog delays” with a cite link to a page called “10 Best Analog Delay Pedals of 2019”. So apparently link spam on Wikipedia is still a valid technique in 2019.

Somewhat disturbingly, on a high traffic programming site there’s a page entitled “How to create a simple REST API in PHP” containing a sentence suggesting if the reader is interested in making YouTube videos, they should follow a link to read camera reviews, which goes to the spammy site. The sentence is completely out of the blue, and doesn’t belong, but is unlikely to be picked up automatically.

Google’s response

Google began to target PBNs in 2011 when it took action against JCPenney for one of the first large-scale link schemes. There was a big drive to kill the technique in 2014.

Since then it’s been a constant battle and core part of their efforts to combat spam, with search updates frequently sending waves of terror through black-hat communities, but they adapt their techniques and keep up their shady business.

The G warns developers not to “…promote your site (or let anyone do so) using policy violating techniques […] posting links on link farms and ‘private blog networks’” but is otherwise surprisingly silent on the topic.

Given that in theory a PBN can be completely undetectable, it’s hard to imagine the practice disappearing. All Google can do is improve the intelligence around their detection, but with improving AI content generation tools, the ability to spin up websites in minutes, and black-hats successfully avoiding detection for years, it’s hard to be optimistic that Google Search can become a fair platform.