Referring Domains Versus Total Backlinks

Every backlink tool shows you both. Most reports lead with the bigger one. It’s the wrong choice, and the reason is structural rather than a matter of taste.

Referring domains counts the number of distinct websites linking to you. Total backlinks counts individual link instances. The second number is always larger, sometimes by two orders of magnitude, and almost all of the excess comes from templates.

Where the extra links come from

Total backlinks inflates through repetition, and repetition is usually mechanical:

• Footers and headers. One link in a global template on a 40,000-page site can register as 40,000 backlinks.
• Sidebars and blogrolls. Same mechanism, same multiplier.
• Pagination and archives. A link in a post that also appears on the category page, the tag page, the author page, and eleven paginated index pages.
• Feed and syndication copies. The same article republished across a network of scrapers.

None of those represent forty thousand editorial decisions. They represent one, replicated by a content management system.

Referring domains is immune to all of it. One site, one count, however many times its template repeats.

Why the distinction matters beyond tidiness

Volatility. A single template change on one large linking site can move your total backlink count by tens of thousands overnight. That produces a dramatic chart with no underlying event. Referring-domain counts move in ones and twos, which is roughly the rate at which link acquisition actually happens.

Comparability. Total backlinks is the number most affected by the counting-rule differences between tools, so it’s the one that varies most between vendors. Referring domains is far more consistent across tools — not identical, since crawls still differ, but the 4× gaps mostly vanish. (Covered in why two tools report different backlink counts.)

Interpretability. “Two hundred and six sites link to us” is a sentence a stakeholder can reason about. “Fourteen hundred backlinks” invites a follow-up question — from how many places? — that you then have to answer anyway.

Manipulability. Total backlinks is trivially inflatable. Referring domains requires distinct sites, which is a meaningfully higher bar. Any metric that’s easy to inflate tends to attract inflation, which degrades its usefulness for everyone.

When total backlinks is the right number

It isn’t useless. It’s the correct metric when the instances are what you care about:

• Detecting a sitewide link. A single domain contributing thousands of rows is a fact you want to know, and you can only see it by looking at instances.
• Anchor-text distribution at the instance level. Though be careful: a sitewide footer with one anchor will dominate the distribution and tell you nothing. Most anchor analysis is better done one-per-domain.
• Link reclamation. Finding every instance of a broken or unattributed mention requires the full list.
• Spotting scraper networks. Thousands of new backlinks from a handful of domains, appearing all at once, is a recognisable pattern in the instance data.

The rule of thumb: use referring domains to describe the profile, use total backlinks to investigate it.

The ratio is a diagnostic

Divide total backlinks by referring domains and you get an average links-per-domain figure. On its own it means little, but as a tripwire it’s useful.

Low ratios — a handful of links per domain — suggest a profile built mostly of individual editorial links. High ratios suggest templates are doing the heavy lifting.

A hypothetical: two sites both report 200 referring domains. Site A shows 1,100 total backlinks, a ratio of about 5.5. Site B shows 96,000, a ratio of 480. Site B’s profile is dominated by a few sitewide placements; its 200 domains are not comparable to A’s 200 domains in what they represent, even though the headline referring-domain figure is identical. (Illustrative figures, not measurements.)

That’s the whole value of the ratio: it tells you when the two headline numbers are describing different kinds of profile. It is not a quality score, and a high ratio isn’t automatically bad — a genuine long-running partnership that lives in a site’s footer is a real link. It just isn’t a thousand real links.

Reporting practice

Lead with referring domains. It’s the number that best answers “how many people link to us,” it’s the most stable month to month, and it survives tool changes best.

Show total backlinks as secondary, with the ratio if it’s unusual.

Chart referring domains, not backlinks. Backlink charts produce cliffs and spikes that generate questions you’ll spend an hour answering, and the answer is nearly always “a template changed on one site.”

Segment referring domains by something, because the raw count treats a national newspaper and an abandoned directory as equals. Banding by DR is the common approach — though read what Domain Rating actually measures before you average anything on that scale. Banding by whether the link is in editorial content, and counting dofollow separately, are both more informative than a bare total.

State the direction over a period, not a snapshot. “206 referring domains, up 14 this quarter, 3 lost” is a report. “206 referring domains” is a fact without a claim attached.

The number you actually want

None of these is the thing you’d measure if you could measure anything. What you’d want is something like “how many independent, discerning sources chose to point at us, weighted by how much anyone trusts them” — and no tool exposes that, because most of it isn’t observable from a crawl.

Referring domains is a rough count of the independent sources. That’s a genuinely decent proxy for the first half of the sentence, and it’s the closest thing to a raw observation your tools produce. Which is exactly why it’s the one to build reporting on: it’s the number with the fewest assumptions baked in before you get to it.