What a connectivity rank tells you about a domain (and what it doesn't)

A connectivity rank is a domain’s position among the 133,241,980 domains in Common Crawl’s domain-level hyperlink graph, ordered by how close each one sits to all the others through links. Position 1 is the most connected domain in the snapshot, and a higher number means a less connected one. The rank tells you how embedded a domain is in the link structure of the web. It does not tell you how many people visit it, whether its content is good, or where it appears in search results.

What harmonic centrality measures

The rank is computed with harmonic centrality. Picture the web as a map of domains, with an arrow wherever one domain links to another. For a given domain, take every other domain and count the fewest links needed to get from there to here. That count is the distance.

A domain one link away contributes 1, two links away contributes 1/2, three links away 1/3, and so on. A domain with no path at all contributes zero. Add up the contributions and you have the domain’s harmonic centrality. Domains that many others link to directly, or reach in a few steps, score highest.

Using reciprocals is what makes the measure workable on a real web. Most domains cannot reach most others, and an unreachable domain adds nothing instead of breaking the sum.

A three-domain example

Take a made-up web of three domains:

• example.org links to example.com.
• example.net links to example.org.
• example.com links to nobody.

Two things show here. example.com comes first without linking anywhere, because the measure counts paths arriving at a domain, not leaving it. And the indirect route from example.net still adds half a point. The real snapshot does the same sum across 133 million domains.

Why the rank is computed on domains, not pages

The graph behind the rank is domain-level. Every link from any page of one domain to any page of another becomes a single connection, however many links there are. Subdomains such as www.example.com and blog.example.com fold into example.com.

That choice suits the question in three ways:

• Links inside a site drop out. A domain cannot raise its own position by linking its pages to each other.
• Uneven crawling matters less. Common Crawl fetches many pages from some sites and few from others, and at domain level one crawled page linking out is enough to record the connection.
• People ask about domains. Who runs this, what does it connect to. A page-level figure would shift with every new URL.

How to read the position number

Divide the position by the size of the snapshot, 133,241,980 domains in cc-main-2026-jul-aug-sep, to see what share of the graph sits ahead of a domain:

A connectivity rank of 1,332,420 means about 1% of the domains in the snapshot are more connected and about 99% are less. Far down the list, many domains have nearly identical scores, so a gap of a few thousand positions there can mean little.

A snapshot is a picture of the link structure at the time of the crawls behind it. Links added since then are not counted. A domain that changed hands keeps the position its previous owner built. When a newer snapshot replaces this one, the total number of domains changes too, so a position can move even if nothing about the domain did.

What a connectivity rank is not

• Not traffic. A link is not a visit. A domain used mainly through an app, a newsletter or an internal network can have a large audience and few links, and the reverse happens too.
• Not quality or trustworthiness. Sites link for many reasons, criticism included, and links can be bought or arranged. The rank says nothing about whether a site’s content is accurate.
• Not a search ranking. Link-based centrality belongs to a family of measures search engines have used to estimate structural importance. A search engine’s results, though, depend on the query and on many signals from its own crawl. A connectivity rank does not predict where a page appears for any search.

Why a new or small site may have no rank yet

There are two reasons a domain can lack a connectivity rank. It may not be in the snapshot at all: it is newer than the crawl, the crawl did not reach it, or no crawled page on another domain linked to it. Or it is in the snapshot but StatsMyWeb has not yet imported its data, since the graph is large and is being loaded in stages.

Either way, the domain report still runs its live checks on every visit: DNS records, HTTPS reachability, headers, the certificate and the rest. You can open the domain report for example.com and change the domain to see this for any name. The page is kept out of search engines until a rank exists, for the reason given on the methodology page. A missing rank is not a verdict on the site.

How to see which domains link to a domain

The number alone does not show why a domain sits where it does. The Web Graph Explorer for example.com does: it lists the domains that link to a domain and the domains it links to, taken from the same Common Crawl snapshot as the rank.

Read the inbound side with the example above in mind. A link from a domain that many others reach brings all of them within two steps, so a few inbound links from well-connected domains can count for more than many from isolated ones. The outbound side does not change the domain’s own score, but it shows where the domain sends its visitors.

Common Crawl publishes its web graph openly, so the figures can be checked against the source data. The about page describes where StatsMyWeb’s data comes from. For the live side of a domain report, our guide to HTTP security headers explains one of the checks in detail.