Get Started

Detect PBN Backlinks: What Each Signal Really Proves

Detect PBN backlinks as a matter of degree, because no tool proves ownership. This guide grades the evidence on the linking page, the surrounding group of sites and the ownership trail, so you can say how sure you are before you act on a link.

Detect PBN Backlinks: What Each Signal Really Proves

You can detect PBN backlinks only as a matter of degree. No tool can prove that a link comes from a private blog network, but a page nobody reads, a cluster of linking sites with shared ownership and a link placed for payment together make a strong case. Check the linking page first, the surrounding group of sites second, and the ownership trail last. Then say how sure you are: suspected, probable or cannot say.

That last step is the one most guides skip. A rival’s links or a seller’s offer rarely get past “probable”, and being honest about that saves you from acting on a guess.

This guide covers three jobs: checking a profile you own, reading a rival’s, and judging a vendor’s offer before you pay. The signals are the same in each, but what you’re allowed to conclude from them changes a lot.

What Counts as Evidence That a Backlink Is From a PBN?

Start by pulling three words apart that get used as one. A link can be spammy, it can be manipulative, and it can come from a PBN. Ahrefs makes the first distinction in its piece on toxic backlinks: spammy links are the junk that any site attracts without trying, while manipulative links are the ones someone placed to move rankings. A PBN link is a narrower claim again, because it says who placed the link and how.

That matters because most of what you can see from the outside only supports the first word. A link from a scraper page is spammy, but nobody paid for it and no network built it. If you call it a PBN link you’ve made a claim your evidence can’t carry.

So what does evidence look like here? It comes in three levels, and each one answers a different question.

The page level asks whether this one linking page looks like it exists for readers. The group level asks whether several linking sites look like they belong together. The ownership level asks whether anyone can show they share an owner.

One more piece of context before the signals. In 2021, Google’s John Mueller said that PBN schemes are “pretty obvious to the appropriate tools & basic scripts”, as Search Engine Roundtable reported at the time. Read that carefully. It’s a statement about Google’s tooling and what Google can see. It doesn’t say you can match that from a browser tab and a backlink export.

How Do You Identify PBN Backlinks on a Single Linking Page?

Open the page that carries the link and read it like a visitor would. You’re looking for signs that the page has a reason to exist other than hosting the link. None of these signs is proof, and the table below says how each one misleads.

Ask these questions in turn. Does the page show any sign of readers, such as comments, shares or traffic that a tool reports? Does the site cover one subject or wander across everything? Is the post on a topic that fits the rest of the site, or does it sit oddly among unrelated articles? Is your link one of very few outbound commercial links on the page, or one of many? Do the about and contact pages exist and say anything?

Signal Where you look What it suggests How it misleads Basis
No traffic evidence for the page A traffic tool’s page report Nobody is meant to read it New pages and niche pages show little traffic too Editorial judgement
Post off-topic for the site The site’s category and archive pages Written to carry a link General-interest sites cover many subjects honestly Editorial judgement
Your link is the main outbound link The post body The post exists for the link Plenty of short, genuine posts cite a single source Editorial judgement
Many unrelated outbound anchors Other posts on the same site Links are for sale across the site Resource and directory pages look the same Editorial judgement
Thin or missing about and contact pages The site’s footer and menu No real publisher behind it Small personal blogs often skip them Editorial judgement
Low authority score A backlink tool A reason to look closer Any site attracts low-quality links, which is Ahrefs’ point about spammy versus manipulative Ahrefs, toxic backlinks article

Notice that every row says “suggests” and none says “shows”. Even three or four of these together only move a single page to “suspected”. They don’t get it further, because every one of them also describes an ordinary neglected blog.

Can You Detect PBN Backlinks From a Backlink Export?

Partly. An export won’t tell you whether a link is from a PBN, but it shows you patterns that a single page can’t. The place to start for your own site is Google’s Links report in Search Console, which lists the sites that link to your site. Google suggests confirming that you recognise those sites and, if unknown, possibly spammy ones appear, asking for removal or, as a last resort, disavowing.

Tool exports add more columns. Sort them and look for groups: many referring domains that look alike in age, subject, anchor wording and where the link sits in the page. A group like that is a lead. One site that happens to match is not.

Be careful with the labels tools attach. A “toxic” or spam score is a prompt to go and look, nothing more, and Ahrefs’ own article says that low-quality links are something every site attracts. If you treat the score as a finding, you’ll end up distrusting links that arrived honestly.

If you’d rather not sift an export alone, a managed review of an existing link profile covers the same ground with someone else doing the sorting. The grading below applies either way.

What you can reach depends heavily on whose links they are. The table compares the three jobs.

Job What you can access What you can’t see Realistic ceiling
Your own profile The Links report, tool exports, your own records of what you bought or were offered Who owns the linking sites and why they linked Probable for a group; suspected for a single link
A rival’s profile Tool exports and the public pages that link Payment, intent, and any report only the site owner gets Probable at best
A vendor’s offer Sample URLs the vendor gives you Anything outside the samples Probable on the samples; cannot say without them

How Do You Tell Whether Linking Sites Belong to the Same Network?

This is where “suspected” can become “probable”. A single odd page could be anything. Several odd pages that share features begin to look like one operation.

Look for three kinds of shared trace. Sites that link to each other in tight loops are one. The same template, the same author names or the same writing habits across supposedly separate sites are another. The third is shared ownership, which is the hardest to get and the easiest to over-read.

On infrastructure, hosting and name server patterns can group sites, but each one has innocent explanations, since thousands of unrelated sites share a host. The full list of infrastructure footprints goes through what each type can and can’t tell you, so this article won’t repeat it.

Ownership is the more interesting problem, because the data has changed. Registration lookups used to run on WHOIS. They’ve been moving to RDAP, which ICANN describes as delivering WHOIS-like data in standardised formats. The format itself is defined by the IETF in RFC 9083. ICANN also says that from 28 January 2025, gTLD registries and registrars are no longer required to provide WHOIS, except for .com, .name and .post.

The practical result is that a blank or redacted registrant is normal now. It proves nothing in either direction. When a registrant name does appear, it’s worth following, and ownership change over time shows how to read old records. Hosting moves leave a trail too, and how hosting and DNS history shows moves covers it. Treat both as supporting evidence, since a domain can change hands for dull reasons.

Because ownership is now so hard to see, group evidence carries more of the load. A cluster of sites that share a template, link to each other and sell links to the same kinds of buyers is more convincing than any one registrant string. Even then, you’re describing a pattern. You aren’t naming an owner.

It helps to remember what you’re up against. Google’s 2021 webspam report is about Google’s own systems, and the numbers below are what it reported that year. None of them are about PBNs, and the report gives no figure for PBNs or for link spam specifically.

What Google reported about its spam systems for 2021 Horizontal bar chart of three figures from Google’s 2021 webspam report: hacked spam reduced by 70 percent, gibberish spam on hosting platforms reduced by 75 percent, and scammy results reduced by 40 percent. Two further figures are shown as notes: SpamBrain identified nearly six times more spam sites than in 2020, and Google aims to keep more than 99 percent of searches spam-free. The report gives no figure for PBNs or link spam specifically. What Google reported about its spam systems for 2021 Reductions Google reported, in percent 0% 25% 50% 75% 100% Reduction in hacked spam 70% Reduction in gibberish spam on hosting platforms 75% Reduction of scammy results 40% Nearly six times more spam sites identified by SpamBrain in 2021 than in 2020 More than 99% of searches kept spam-free, which is the goal Google states in the report No figure here is about PBNs or link spam: Google’s report gives none for either. Source: Google Search Central Blog, “How we fought Search spam on Google in 2021”, 21 April 2022. developers.google.com/search/blog/2022/04/webspam-report-2021
What Google reported about its spam systems for 2021. Google’s report gives no figure for PBNs or link spam specifically. Source: Google Search Central Blog, 21 April 2022.

What the chart supports is narrow. Detection at Google’s end is automated and has scaled, with the report saying SpamBrain identified nearly six times more spam sites than in 2020. That’s a reason to expect your own reading to be a grading exercise, not a match against a known list.

Can You Check a Competitor’s Backlinks for a PBN?

You can read them, and you can form a view. You can’t confirm one. That’s the whole honest answer, and the rest of this section is about living with it.

Everything you see about a rival is outside-in. You have their public links, your tool’s export and whatever the linking pages show. You don’t have their Search Console, their invoices or their emails, so you can’t see whether any link was paid for. Payment is the part that makes something a link scheme, and it’s invisible to you.

That caps a rival check at “probable”, and only when a group of signals lines up. Here’s a constructed example, generic and not drawn from any real site: a rival’s export shows a run of links from sites that share a template, cover unrelated subjects, and link to each other. You’d call that probable network activity. You wouldn’t call it proven, and you wouldn’t say it in public.

There’s a second reason to stay modest. A strong link profile and a weak one can both include links that were placed for money, and neither tells you whether the links help the rival rank. Mueller’s 2021 remark, in the same Search Engine Roundtable report, was that there is “nothing unique about PBNs, they’re just link schemes like others have been for many years.” Seeing one doesn’t tell you what Google does with it.

A rival check tells you what kind of competition you’re facing and whether a tactic is common in your niche. It shouldn’t drive an accusation.

How Do You Check a Vendor’s Offer Before You Pay?

A vendor’s offer is the one case where you can ask for evidence. That changes the job, because you can then run the page and group checks above on whatever they hand over.

Ask for sample URLs of live placements, then treat those samples the way you’d treat a rival’s links. Read the pages. Look at what else the host sites link out to. Check whether the samples look like they belong to one group of sites, and whether that matches what the vendor told you. If the offer says “real sites with real audiences” and the samples have no readers, you’ve learned something.

How a vendor responds is evidence too. A seller who gives samples and explains the nature of the sites is giving you something to test. A seller who won’t give any is leaving you at “cannot say”, and that is the correct grade, because you can’t observe what you haven’t been shown. A refusal doesn’t prove a PBN. It does mean you’d be paying for something you can’t check.

Keep your standards consistent across vendors. If a seller describes the offer as a PBN, they’re telling you what it is, and your job is to see whether the description matches the samples. For a reference point on what a seller should be willing to put in writing, read what a transparently described PBN offer lists and hold other offers up against it.

What Does a Positive Finding Mean, and What Should You Do Next?

It means you have a grade, not a verdict. The confidence ladder below turns each grade into a sensible next move, and it deliberately has no “confirmed” row. No outside test confirms a PBN. If someone tells you their tool does, ask what evidence it shows you.

Level What you’ve seen What it doesn’t prove Sensible next step
Cannot say A low score, a missing contact page, or nothing at all because the samples weren’t given Anything about intent or ownership Leave it alone, or ask the seller for samples
Suspected Several page-level signs on one or two pages That a network exists, or that anyone paid Note it, look for the same signs elsewhere, monitor
Probable Page-level signs repeated across a group of sites that share traces Who owns the sites, or what Google does with the links For your own links, ask the seller or request removal; for a rival or vendor, decide whether it changes your plan

Before acting, count the cost of being wrong. If you treat honest links as manipulative, you may remove or disavow things that were helping you. Google’s position, as reported by Search Engine Roundtable in 2023, is that links people make to manipulate rankings are mostly ignored, so a hit isn’t a penalty forecast.

If a finding reaches “probable” on your own profile, the first move is to ask whoever placed the links to take them down. Disavow is the last resort, and only the way Google’s own help describes it. For the practical steps, see removal and disavow steps.

If you came here deciding whether to buy links at all, you may find that placements on real, read publications are easier to evaluate, because the audience and the editor are visible. And if what you really want to know is how risky the whole category is, the risk question, answered in full goes through Google’s own rules.

Frequently Asked Questions

Is there a tool that tells me for certain a link is from a PBN?

No. Tools report metrics and patterns, and those can prompt you to look. The verdict is an inference, which is why this guide grades findings instead of passing or failing them.

Does a low domain rating mean a site is a PBN?

No. Ahrefs notes that low-quality links are attracted naturally by any site. A low score is a reason to take a closer look, and that’s all it is.

Can Google tell when a link comes from a PBN?

Google’s John Mueller said in 2021 that PBN schemes are “pretty obvious to the appropriate tools & basic scripts”. That’s about Google’s own tooling. It doesn’t mean an outside reader can do the same from a browser.

Should I disavow links I think are from a PBN?

Only in the way Google’s own disavow help describes, and as a last resort after asking for removal. Google is reported to say that most manipulative links are ignored. Don’t build a rule like “disavow everything suspicious” out of a guess.

Why can’t I see who owns a linking site anymore?

Registration data has moved to RDAP, and from 28 January 2025 gTLD registrars are no longer required to offer WHOIS except for .com, .name and .post, according to ICANN. Redaction is normal, so ownership is one signal among several.

Sources

  1. Google Search Central Blog, How we fought Search spam on Google in 2021: https://developers.google.com/search/blog/2022/04/webspam-report-2021
  2. Google Search Console Help, Links report: https://support.google.com/webmasters/answer/9049606
  3. Ahrefs, Toxic Backlinks: SEO Woe or a Load of Baloney?: https://ahrefs.com/blog/toxic-backlinks/
  4. Search Engine Roundtable, Google on PBNs and detection (13 May 2021): https://www.seroundtable.com/google-pbns-are-pretty-obvious-31426.html
  5. Search Engine Roundtable, Google on manipulative links being mostly ignored (14 April 2023): https://www.seroundtable.com/google-most-links-ignored-35214.html
  6. ICANN, Registration Data Access Protocol (RDAP): https://www.icann.org/rdap
  7. IETF, RFC 9083: JSON Responses for the Registration Data Access Protocol (RDAP): https://www.rfc-editor.org/rfc/rfc9083
Madison Houlding
Madison Houlding
November 9, 2026, 16:1874 views
Madison Houlding
Madison Houlding Content Manager

Madison Houlding Content Manager at Hetneo's Links. Madison runs editorial across the link-building space, auditing campaigns, writing the briefs that keep guest posts from sounding like ad copy, and turning analytics into next month's roadmap. Loves a clean brief, hates a buried lede.

More about the author

Leave a Comment

Comments (0)