After this lesson you should be able to
- Distinguish backlinks, referring domains and referring URLs correctly
- Read the rel attribute on a link and know what each value means
- Explain what Trust Flow, Citation Flow and Domain Rating measure and who makes them
- Recognise which terms are vendor products rather than search engine concepts
The counting words: backlink, referring domain, referring URL
Start here, because almost every misleading link report in existence exploits confusion between these three.
A backlink is a single hyperlink on another site pointing at your site. If a page links to you three times — once in the navigation, once in the body, once in the footer — that is three backlinks.
A referring URL (sometimes called a linking page or source URL) is a specific page that contains at least one backlink to you. Our example page is one referring URL carrying three backlinks.
A referring domain (also linking root domain, or unique domain) is a distinct website that links to you at least once, regardless of how many pages or how many links. Fifty pages on one site linking to you is fifty referring URLs and one referring domain.
Most tools also report referring IPs and referring subnets — how many distinct servers and network blocks your links come from. These matter when you are assessing whether a set of "different" sites is really one operation, which is what Module 3 is about.
Why referring domains matter and link counts mislead
The distinction above is not pedantry. It is the difference between a number that describes your site and a number that describes one accident of somebody's template.
Take the profile I use as the evidence base for this course: billhartzer.com, 1,301,839 links from 22,260 referring domains, none ever bought. That is an average of 58 links per domain, which already tells you the distribution is not remotely even. It gets more extreme the closer you look:
- A single domain sent 219,159 links — 16.8% of the entire profile — from one sitewide placement. One editorial decision, made once, on a site with a lot of pages.
- The top 100 referring domains account for 84.3% of all links. The remaining 22,160 domains share the other 15.7%.
So a report boasting "1.3 million backlinks" describes a link profile roughly as well as a payroll total describes a company's salaries. Sixteen percent of it is one person's decision. If that site redesigns its template tomorrow, the headline number collapses and nothing meaningful about the site's standing has changed.
The rule that follows: count referring domains, always, and treat the links-per-domain ratio as a diagnostic. A healthy profile usually sits in the low single digits. A high ratio is not automatically bad — it may just mean one generous sitewide — but it always means the link count is telling you about a small number of sources rather than about broad citation.
Anchor text and the attributes on a link
Anchor text is the visible, clickable text of a link. It gives search engines a description of the destination written by somebody other than the destination's owner, which is why it has always carried weight and always been manipulated. Practitioners classify anchors as branded ("Hartzer.com"), naked URL ("https://example.com"), generic ("click here", "this study"), exact-match ("buy running shoes online"), partial-match, and image anchors — where the alt text of the image serves the same role.
Links can carry a rel attribute that qualifies the relationship:
- nofollow — introduced by search engines in the mid-2000s to fight comment spam. It tells engines not to treat the link as an endorsement. In 2019 Google said it would treat it as a hint rather than a strict directive, meaning it may sometimes be considered.
- sponsored — introduced alongside that change, marking links that are paid or part of an advertising arrangement. This is what a compliant paid link should carry.
- ugc — user-generated content: comments, forum posts, profile links. Content the site did not author and does not vouch for.
- dofollow — not a real attribute. It is industry slang for a normal link with no rel qualifier. You will never find
rel="dofollow"in a specification, and writing it into your HTML does nothing.
Links can also be suppressed at page level with a robots meta tag, or built with JavaScript in ways that make them unreliable for crawlers. If a link matters to you, view the page source and confirm it exists as a real anchor element with a real href.
Equity, sitewides and tiers
Link equity (link juice) is the ranking value understood to flow from a linking page to a linked page. It is a model, not a measurable quantity — Lesson 1.1 covers what the model does and does not license you to believe.
Sitewide link: a link appearing on every page of a site because it sits in a template — a header, sidebar, footer, or blogroll. Sitewides produce enormous link counts from one placement, which is where the 219,159-link figure above comes from. Search engines have understood sitewides for a very long time and generally treat them closer to a single citation than to thousands. A paid sitewide footer link across a large site remains one of the most recognizable manipulation signatures there is.
Tier: a level in a deliberately layered link structure. Tier one links point at your site; tier two links point at your tier one links, intending to strengthen them; tier three points at tier two, and so on. The theory is not wrong — a link on a well-linked page is worth more than a link on an ignored one — but in practice tiered structures are built with automated tools at the lower levels and are one of the most heavily fingerprinted patterns in the industry.
Two more terms you will meet: link velocity, the rate at which new links arrive over time, and link decay or link rot, the rate at which existing links disappear. Lesson 1.5 puts real numbers on the second.
The metrics, and who makes them
None of these come from Google. Every one is a tool vendor's model, built from that vendor's own crawl of the web. They are useful for comparing sites within one tool and meaningless when compared across tools.
- Trust Flow (Majestic). A 0-100 score estimating how trustworthy the sites linking to a page or domain are, computed by measuring distance from a manually reviewed set of trusted seed sites. It is a quality-flavoured metric: hard to inflate with volume.
- Citation Flow (Majestic). A 0-100 score reflecting how many links point at a page or domain, without regard to their quality. A volume metric.
- The relationship between the two is the useful part. Citation Flow far above Trust Flow means many links from sources with no trust lineage — the classic signature of bulk link acquisition. Majestic also publishes Topical Trust Flow, which breaks trust down by subject category, and that is often more informative than the headline number.
- Domain Rating, or DR (Ahrefs). A 0-100 logarithmic score of the strength of a domain's backlink profile.
- Domain Authority, or DA (Moz). A 0-100 predictive score estimating how likely a domain is to rank. Note the difference in intent: DR describes a link profile, DA predicts an outcome.
- Authority Score (Semrush) and URL Rating (Ahrefs) follow the same pattern at domain and page level respectively.
All of these are logarithmic or near-logarithmic, so the distance from 20 to 30 is nothing like the distance from 70 to 80. All of them can be gamed, and there is an entire industry doing so, which is precisely why buying links filtered by a metric threshold selects for sites that have optimized for that threshold.
The risk words: toxic link and disavow
Toxic link is a tool vendor's term, not Google's. Various tools compute a toxicity or spam score for each link from features such as the linking site's outbound link count, its content, its anchor text patterns, and its neighborhood. There is no published evidence that these scores correspond to how Google actually treats a link, and I have watched perfectly ordinary links get flagged as toxic by one tool and rated fine by another. Use them to sort a large list for human review. Never use them as a decision.
Disavow is a genuine Google tool. You upload a text file listing URLs and domains you want Google to ignore when assessing your site. It exists because of the link-buying era, and Google has repeatedly said most sites should never use it — its systems ignore spam links automatically, and a badly built disavow file can remove links that were helping you. The realistic use cases are narrow: an active manual action for unnatural links, or documented knowledge that someone bought links at scale for this domain in the past.
One last piece of vocabulary you will hear in sales calls: negative SEO, the idea that a competitor can harm your rankings by pointing bad links at you. It is possible in principle and vanishingly rare in practice, and it is the standard justification offered for selling disavow services to sites that do not need them.
Questions
What is the difference between backlinks and referring domains?
A backlink is one link; a referring domain is one website linking to you regardless of how many links it sends. The gap can be enormous. On the profile used throughout this course, one domain accounts for 219,159 links — 16.8% of 1.3 million — from a single sitewide placement. Referring domains is the number that describes how many independent sources cite you.
Is a nofollow link worthless?
No. Google said in 2019 that it treats nofollow as a hint rather than a strict directive, so nofollow links may sometimes be considered. More importantly, they still send traffic, still build brand recognition, and still appear in the corpus that AI systems and other engines read. A nofollow link from a major publication beats a followed link from a site nobody visits.
Which metric should I trust: Trust Flow, DR, or DA?
None of them alone, and never across tools. Each is one vendor's model of one vendor's crawl. Trust Flow leans toward quality of sources, DR toward strength of the link profile, DA toward predicted ranking ability. Use one consistently for comparison within a prospect list, and pair it with relevance and evidence of real traffic before deciding anything.
Is rel="dofollow" a real thing?
No. There is no dofollow attribute in any HTML specification. A normal link with no rel qualifier is what everyone means by dofollow. Adding rel="dofollow" to your markup accomplishes nothing at all. The term persists because it is a convenient opposite of nofollow in conversation, but seeing it in a vendor's deliverable spec is a signal about that vendor.
What counts as a sitewide link?
Any link placed in a template so it appears on every page of a site — header, footer, sidebar, or blogroll. It generates one link per page, which can mean hundreds of thousands of backlinks from a single decision. Engines have long treated sitewides closer to one citation than to many, and paid sitewide footer links remain one of the most recognizable manipulation patterns.