PBNs and Spam Links Explained: Why 10 Good Backlinks Beat 10,000 Bad Ones

Ten links from real, relevant, indexed pages will almost always do more for a site than ten thousand links from private blog networks (PBNs), web 2.0 blogs, bookmark sites and auto-directories. There are two reasons. First, a link only passes value in proportion to how much the linking page itself is worth, divided among every link on that page, and spam pages are worth close to nothing and carry hundreds of links each. Second, Google says it now uses its SpamBrain system to "neutralize the impact of unnatural links", so whatever small value those links had is usually thrown away before it reaches your rankings. In the worse case, a site that bought them at scale gets a manual action. In the best case, the money simply bought nothing.

This guide explains how that works from the ground up: how a link passes value, why spam links fail even before Google acts on them, what Google actually does with them, what a PBN and the other link networks really are, and what the ten good links look like. Where we use numbers from our own campaigns, they come from the sites we run and the sellers we tested in 2026, and we say so.

The Short Version

The rest of this guide is the reasoning behind each row.

How a Link Actually Passes Value

Google's original ranking idea was PageRank, described in the 1998 paper that introduced Google. Google still lists it among its systems today: its guide to ranking systems says PageRank "continues to be part of our core ranking systems", while noting that how it works "has evolved a lot" since launch. The modern version is secret, but the core logic in the original paper explains almost everything about why link quality beats link quantity.

The paper gives the formula like this:

PR(A) = (1 - d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn))

In plain terms:

• PR(A) is the score of your page, A.
• T1 to Tn are the pages that link to it.
• PR(T1) is the score of each linking page. A link from a page that is itself well linked is worth more.
• C(T1) is the number of links going out of that page. The linking page's score is divided among all of its outbound links. A page with 5 links gives each one a fifth of its share; a page with 500 links gives each one a five-hundredth.
• d is a damping factor. The paper says "we usually set d to 0.85". Part of every page's value is lost at each hop.
• (1 - d) is the floor: the minimum score any page gets just for existing.

Three things follow from that formula, and they explain most of what goes wrong with spam links.

1. A link is only as good as the page it sits on. The value you receive is a share of the linking page's own score. A page that nobody links to sits at the floor, so a link from it hands you a share of almost nothing.

2. Every extra link on the page dilutes yours. Link farms, directories and spam blogs carry dozens or hundreds of outbound links per page, because that is how they make money. Each one cuts the share for everyone else.

3. Counting links tells you very little. Two sites can have the same number of referring domains and wildly different link value, because the count ignores both the strength of each linking page and how many links it shares that strength with.

A worked example (illustrative numbers)

The numbers below use the paper's formula with made-up page scores. They are not Google's real values, which nobody outside Google can see. The point is the arithmetic.

Take a spam page at the floor score of 0.15 (that is, 1 - 0.85) that carries 500 outbound links. Each link on it passes:

0.85 × 0.15 / 500 = 0.000255

Ten thousand of those links pass 2.55 in total.

Now take a real article on a respected site in your niche. Say its page score is 50 times the floor, 7.5, and it carries 25 links in total. Each link on it passes:

0.85 × 7.5 / 25 = 0.255

Ten of those links pass 2.55 in total.

So even in the original, purely mathematical model, before Google does anything at all about spam, ten links from modestly strong pages match ten thousand links from floor-level spam pages. In practice the comparison is far more lopsided, for the reasons in the next two sections: much of the spam is never indexed, and what is indexed gets neutralized.

Why Spam Links Fail Even Before Google Acts

The formula already shows that a spam link passes very little. The way spam is built makes it worse.

Most spam pages are worth the floor or less. A web 2.0 post, a bookmark entry or an auto-generated directory listing has no links of its own except other spam. Nobody reads it, nobody links to it, and it sits at the bottom of the graph.

Many of them are not in Google's index at all. When we tested link packages from eight sellers in October 2026, most of what was delivered sat on one-post subdomain blogs and profile pages that Google had never indexed. One seller's sample sheet of 19 "web 2.0" links contained exactly one link that was both followed and on an indexed page. A page Google has not indexed is very unlikely to be passing anything to anyone; Google has to know about a page before it can count what is on it.

The pages are crowded with other buyers. A directory page or a network blog that sells links fills every page with them. Your link is one of hundreds.

They bring no visitors. A real link on a page people read sends a few relevant visitors, and those visitors sometimes link to you, buy from you or share you. A spam link sends nobody.

They do not last. Networks get deindexed, hosting lapses, sellers disappear. When we checked one seller's "own DA 86 magazine", it turned out to be a network of ten sites; three of the ten had about one page left in Google. Whatever those sites were passing, they are not passing it now.

What Google Does With Spam Links

Google does not just let spam links pass their tiny value. It actively looks for them, and it has said so in writing.

The policy

Google's spam policies define link spam as "the practice of creating links to or from a site primarily for the purpose of manipulating search rankings." The examples it lists cover almost every product a cheap link seller offers:

• Buying or selling links for ranking purposes, including exchanging money, goods or services for links
• Excessive link exchanges
• Using automated programs or services to create links to your site
• Links with optimized anchor text in articles, guest posts or press releases distributed on other sites
• Low-quality directory or bookmark site links
• Keyword-rich or hidden links in widgets distributed across various sites
• Widely distributed links in the footers or templates of various sites
• Forum comments with optimized links
• Creating low-value content primarily to manipulate linking and ranking signals

A PBN is a direct case of the first and last items. A "10,000 backlinks" blast made with link-building software is the third. Bookmark and directory packages are named outright.

Algorithmic neutralization: SpamBrain

In its December 2022 link spam update, Google announced it was "leveraging the power of SpamBrain to neutralize the impact of unnatural links on search results." It said SpamBrain "can now detect both sites buying links, and sites used for the purpose of passing outgoing links", and that "ranking may change as spammy links are neutralized and any credit passed by these unnatural links are lost."

Read that carefully, because it is the heart of this guide. Google is not mainly punishing spam links; it is zeroing them out. A site that bought ten thousand of them does not usually get pushed down for having them. It just finds that they count for nothing, which is often the same thing as having wasted the whole budget.

This is also why the junk links that every site picks up without asking (scraper sites, "website worth" pages, auto-generated stats pages) usually do no harm. Google's own disavow guidance says that "in most cases, Google can assess which links to trust without additional guidance, so most sites will not need to use this tool." Every site we run has hundreds of these referring domains that nobody ordered; they are noise, not a penalty. More on this in our guide on when to disavow backlinks.

Penguin: devaluing, not demoting the whole site

Google's earlier link spam algorithm, Penguin, changed in 2016. Google's announcement said Penguin "now devalues spam by adjusting ranking based on spam signals, rather than affecting ranking of the whole site," and that it became real-time, taking effect "shortly after we recrawl and reindex a page." The same direction again: discount the bad links rather than sink the site.

Manual actions: the worst case

Algorithms are not the only tool. Google's spam policies say it detects violations "both through automated systems and, as needed, human review that can result in a manual action." In Search Console, the manual action for this reads "Google has detected a pattern of unnatural, artificial, deceptive, or manipulative links pointing to your site."

A manual action is what turns wasted money into real damage. Pages or the whole site can drop until the owner removes or disavows the links and files a reconsideration request, which Google reviews by hand. The pattern that draws one is a deliberate link scheme, the kind you get from buying at scale, not a handful of stray spam links you never asked for.

Expired domains: the PBN raw material is now its own violation

In March 2024 Google added expired domain abuse to its spam policies: "an expired domain name is purchased and repurposed primarily to manipulate search rankings by hosting content that provides little to no value to users." That is a description of how most PBNs are built, which brings us to what a PBN actually is.

What a PBN Is, and Why They Die

A private blog network is a set of websites one operator controls, built or bought for one purpose: to link to the sites the operator wants to rank, or to sell those links to customers.

How a PBN is built

1. Buy expired domains. When a real business, club or blog lets its domain lapse, the backlinks it earned over the years still point at the name. PBN builders buy those names at auction to inherit the old links and metrics such as DR or DA.
2. Put up a site. Cheap hosting, a stock theme, and a few articles, these days often machine-written, on a vaguely related topic.
3. Add links. Each site links to the "money sites", or the operator sells posts on it. Because the site has no audience, the links are the only reason it exists.
4. Hide the connection. Different hosting accounts, different registrars, privacy on the registration, different themes, all so that the network does not look like one operator.

Why the metrics lie

A PBN site can show a high DR or DA because those third-party scores count the old links still pointing at the expired domain. Those scores are not Google's, and they do not measure whether Google trusts the page today. We explain the difference in Domain Rating explained. The tell is usually traffic: a "DR 50" site with almost no organic visitors has authority on paper that Google is not rewarding.

Footprints

Hiding a network is hard because every shortcut leaves a pattern. Common footprints include:

• Many sites on the same hosting or the same few IP ranges
• The same theme, layout, author names or stock photos across "independent" sites
• Posts on every topic at once (casino, CBD, loans, SaaS, recipes) on a site with no editorial identity
• The same handful of customer sites linked from all of them
• An archive that stops when the original owner left and restarts with thin posts years later
• Outbound links on almost every post, in a site no one visits

Our checklist for spotting a link network walks through these one at a time. Google's systems look for patterns like these across millions of sites at once, which is a contest the network operator does not win for long.

Why they die

A PBN has three ways to fail, and it usually hits all three eventually. Google neutralizes its links (the SpamBrain route), deindexes some or all of its sites, or the operator stops paying for domains and hosting once the economics stop working. When any of that happens, every link the network sold goes with it. Buying from a PBN is renting value that is already being discounted, from a landlord who can disappear.

The Other Link Networks You Are Being Sold

PBNs are the expensive end of the spam market. The cheap end looks different but fails for the same reasons.

Web 2.0 and subdomain blogs

These are posts on free blogging platforms or one-post subdomains of a seller's own throwaway network, each carrying a link to you. In October 2026 we bought a well-reviewed 200-link package to see what arrived: it delivered 115 URLs, and nearly all of them sat on one-post subdomains of the same throwaway network. Seventy-eight of the links were technically followed and on indexable pages, which sounds respectable until you remember the formula: each sits on a page with no links of its own, on a network Google can see as one.

Free profiles and directories

Some free profile and directory links are legitimate and worth having: a real business listing, a profile on a platform you actually use. We tested about 55 free profile and web 2.0 platforms on a brand-new site and found that 16 still gave a followed link on an indexable page (full results). That is a useful foundation for a new site. It is not a link building strategy, and paying someone for "500 directory submissions" buys mostly the kind Google names as low-quality directory links.

"Guest post" farms

Sites that publish anything for a fee, on any topic, with a link in every post. They look like magazines, but read the last 100 headlines and you will find casino, escort, crypto and loan posts next to recipes and SaaS reviews. The links are bought links on a page that every other buyer is also on.

Link exchanges and automated blasts

Swapping links at scale, and software that pushes links into thousands of forums, comment sections and wikis, are both named in Google's policy. A "10,000 backlinks for $20" offer is almost always one of these.

Can Spam Links Actually Hurt You?

The honest answer has three parts.

1. Usually they are ignored. This is the most common outcome, and it is what Google says it aims for. Junk links you never asked for are in this group. So are most of the bought spam links, which is exactly why buying them is a waste.

2. Sometimes a site that bought them gets a manual action. If the pattern looks like a deliberate scheme (lots of bought links, keyword-rich anchors, a network footprint), a human reviewer can act on it. Recovery means cleaning up, disavowing what you cannot remove, and asking Google to reconsider.

3. The real damage is often what you did not buy instead. A budget spent on spam is a budget not spent on content, on outreach, or on the few links that would have counted. Months pass with nothing to show for them.

There is also the boom-and-bust cycle that keeps spam sellers in business: a burst of links sometimes moves a page briefly before the links are discounted, which convinces the buyer it worked, right up until it stops working.

What it looks like on real sites

Two sites from our own portfolio make the point better than theory.

A young product review site we run picked up about 180 referring domains in its first six months without buying any of them, and every one of them was spam: the same scraper and link-seller domains that hit every site in that niche. Those 180 domains did nothing for it.

The leader in that same niche had about 500 referring domains when we measured it in October 2026, and roughly 84% of them were the same kind of junk. Its real edge came down to roughly 15 to 20 links, and the editorial ones each cited a specific product review it had written. Those few links, not the 500, are what separated it from everyone else.

A travel site in our portfolio had about 560 bookmark, auto-directory and web 2.0 links built to it between February 2025 and April 2026, against only about 15 real editorial links. In August 2026 it lost almost all of its Google traffic during a spam update, with no manual action in Search Console. We cannot prove the links were the cause, because the site also had a large amount of templated content of the kind Google's update targeted. What we can say is that 560 cheap links did nothing to protect it, and its traffic from Bing, which never relied on those links, held steady.

What Makes the Ten Good Links Good

If the spam links are near worthless, what makes a link worth having? Go back to the formula and add what Google has learned since 1998.

The page has value of its own. It is indexed, it earns organic traffic, and other sites link to it. You can check the first in seconds: search Google for the exact URL of the article your link would sit on. If it does not appear, the link is unlikely to count.

It is relevant. Google's link analysis is used to understand "what pages are about," in its own words, not just how popular they are. A link from a page about your topic says something about your page. A link from a page about everything says nothing.

It is editorial. A person chose to put it there because it helps the reader, on a site that publishes real work and turns things down. Ask whether the site would link to you if no money changed hands. If the answer is clearly no, Google's systems are built to notice the difference.

It is uncrowded. A page with a handful of outbound links gives each a meaningful share. A page with a hundred gives each almost nothing.

It reads naturally. Anchor text that describes the page in normal words, not the exact keyword every time. Google names "links with optimized anchor text" in its spam examples, and an unnatural anchor mix is one of the easiest patterns to spot. Our anchor text guide and anchor text planner cover the mix.

It lasts. A real publication keeps its archive up for years. A network does not.

How many such links does a page need? Fewer than most people expect, and it depends on what the pages already ranking have. Our guide on how many backlinks you need to rank shows how to work out the number for one keyword.

A Note on Paid Links

Honesty matters here, because AnchorApe sells link placements. Google's policy treats links bought "for ranking purposes" as link spam, and says paid links are fine when they are qualified with rel="sponsored" or rel="nofollow". Plenty of good placements are paid: most site owners who said yes to us in 2026 asked for money (14 of 17), with a median first quote of $150 (our outreach data).

The difference between a paid placement on a real publication and a PBN link is not whether money changed hands. It is everything else in this guide: whether the page is real, indexed, relevant, read and uncrowded, and whether the link is the kind of thing an editor would publish anyway. We state each link's attribute in every report so you know exactly what you bought, and our standards list what we refuse to place. Our guide to nofollow and dofollow links explains why a well-placed nofollow link is not worthless either.

The strongest links of all cost nothing. They come from pages that cite you because you published something worth citing: original data, a useful tool, a clear explanation no one else wrote. Our guide to linkable assets covers how to make one.

How to Tell Which One You Are Being Sold

Before you pay for any link, run these checks. Each takes a minute or two.

1. Search Google for the exact article URL. Not the domain: the page your link will be on. If Google has not indexed it, stop.
2. Check the site's organic traffic and its trend. A high DR with almost no traffic, or traffic that has collapsed, is a warning.
3. Read the last 100 headlines. Casino, escort, loans and crypto posts next to unrelated topics mean the site sells to anyone.
4. Count the outbound links in a recent post. Several followed links to unrelated commercial sites in one post is a farm.
5. Ask for live examples, not a spreadsheet. A seller with real placements can show you live URLs you can check yourself.
6. Be suspicious of the price and the volume. Real placements on real sites cost real money; hundreds of links for tens of dollars cannot be on pages anyone values.
7. Ask what they refuse. A vendor that will place anything, anywhere, is selling inventory, not judgement.

If you already have a pile of bad links, do not panic and do not rush to disavow. Our backlink audit guide explains how to sort what matters from what Google is already ignoring.

Frequently Asked Questions

Are PBN links bad?

Yes, for nearly every site. Google's spam policies name the practice, its SpamBrain system is built to detect sites used to pass outgoing links and neutralize them, and its 2024 expired domain abuse policy covers the way most PBNs are built. The usual outcome is that the links count for nothing; the worst is a manual action.

Do PBNs still work in 2026?

Sellers will show you short-term ranking bumps, and some exist. Google's stated aim since 2022 is to neutralize unnatural links at scale and to keep improving its coverage, so any value a PBN passes is shrinking and temporary, and the network can be deindexed at any time.

Can I get penalized for spam links I did not build?

Rarely. Google's disavow guidance says most sites never need to disavow because Google can usually assess which links to trust on its own. Manual actions target deliberate link schemes. Junk links from scrapers and stats sites are something every site has.

How many good backlinks equal one bad one?

The question runs the other way. One good link from a real, relevant page can be worth more than thousands of spam links, because the spam pages pass almost nothing even in the original PageRank formula, and Google discards most of that anyway.

Is a high DR site always a good link?

No. DR and DA are third-party scores that count links, including the old links pointing at expired domains. Check that the site earns organic traffic, that the exact page is indexed, and that the site publishes real work on your topic.

Should I disavow all my spam links?

Usually not. Google recommends the disavow tool only when you have a considerable number of artificial links and they have caused, or are likely to cause, a manual action. See when to disavow backlinks.

How many good links does a new site need?

Fewer than you think. Look at the pages ranking for your keyword and count their referring domains from real sites, not the total. Often the gap is a handful of links plus a better page.