Separate the link from its source domain
A backlink is a hyperlink from another website to a target page. A referring domain groups one or more of those sources by a domain boundary. Many links can therefore come from one domain. But a row in an export is not automatically one physical hyperlink: a provider may combine repeated links, summarize a page or report different observations of it. Name the measurement unit before using the word total.
Ahrefs · Backlinks / Referring domains
Follow the diagram from observations to domains
The diagram uses A only and makes the grouping visible. Expand paper.example to see multiple source documents and the observations assigned to them. The navigation switch adds twelve distinct source pages within that existing domain. It does not duplicate a CSV row. Observation and document counts rise while the domain count remains unchanged. This is an explicit hypothetical navigation pattern, not a conclusion about a real publisher’s template.
The diagram uses export A. Its 9 observations collapse to 7 source–target pairs, 6 documents, 5 hosts and 4 grouped domains. Add twelve fictional navigation rows from paper.example: observations become 21 while domains stay at 4. A repeated export row differs from a separate link on another page.
| Counting unit | A | A + 12 |
|---|---|---|
| Export observations | 9 | 21 |
| Source–target pairs | 7 | 19 |
| Source documents | 6 | 18 |
| Source hosts | 5 | 5 |
| Referring domains | 4 | 4 |
A duplicate row is not a sitewide link
An identical exported record is removed once in this method. Two observations with different anchors, attributes or fragments remain distinct evidence even if the source–target document pair is the same. Pair grouping intentionally hides those differences for one question, but the original CSV and draft preserve them. Repeated navigation on different pages can be separate links; repeating one row during export does not create a new link on the web.
Hosts, domains and publishers are different concepts
paper.example and shop.paper.example are two hosts but one grouped domain here. Public suffix rules also cover multi-part suffixes and selected private hosting services, so a simple last-two-labels split is unreliable. The chosen boundary is not a corporate ownership register: separate domains can have one owner, and hosted tenants can be independent. Do not call a count of domains a verified count of independent editorial publishers.
URLs use browser normalization. Observation identity preserves fragments and exact anchor/attribute text. Pair and document units remove fragments but keep queries, paths and protocols. Domains use the Public Suffix List including private suffixes; that is a grouping rule, not proof of shared ownership. No redirects or canonical URLs are resolved.
Use the denominator that answers your question
Use source documents to discuss where links were observed, pairs to compare document relationships and domains to describe breadth under an explicit grouping rule. A domain count does not measure editorial quality, topic relevance, traffic or ranking contribution. No universal ideal backlinks-to-domains ratio follows from this exercise. For change over time, preserve target scope, dates, filters and collection limits, and review membership as well as totals. The companion comparison shows why 4 domains can mean different sets in two exports.
Compare coverage without turning it into quality
Record the provider, report, index/live-history selection, filters, export limits and collection time in each source’s scope. Match the intended target first. The tool cannot verify that the supplied files actually use comparable provider settings. A collection date is not a link publication date.
Only in A/B means absent from the other supplied file, not absent from the web or proven new/lost. Overlap is not accuracy, precision or recall: there is no verified complete universe here. Equal totals can contain different domains.
Why backlink counts differ: compare scope before judging accuracy
Download the example and apply the method
Use Source URL and Target URL; Anchor, Rel and Nofollow are optional. Common English header aliases are accepted. Up to 2,000 records and 1 MB per CSV; JSON drafts up to 5 MB. Unsupported or invalid URL records are excluded with their record numbers. Other columns are retained in the draft but do not define an observation.
Files and notes stay in this browser. Compare supplied exports without contacting a backlink provider or fetching source URLs. Save a JSON draft to retain the exact inputs; nothing is saved to an account.