What domain authority actually looks like across 118 million sites
We computed the authority of every domain in an independent index of 117,963,409 sites and 7.1 billion links. Three findings stand out — and one of them contradicts advice you have probably been given.
Published 5 August 2026 · Source: Common Crawl cc-main-2026-may-jun-jul · Every figure is computed, not estimated
The full distribution
Every domain in the index, grouped by authority score (0–100). The web is not a bell curve — it is one of the steepest power laws you will find in any dataset.
| Authority | Domains | Share of web | Relative size |
|---|---|---|---|
| 0–9 | 97,782,937 | 82.9% | |
| 10–19 | 17,034,488 | 14.44% | |
| 20–29 | 2,655,549 | 2.25% | |
| 30–39 | 413,980 | 0.351% | |
| 40–49 | 64,537 | 0.055% | |
| 50–59 | 10,060 | 0.0085% | |
| 60–69 | 1,569 | 0.0013% | |
| 70–79 | 244 | 0.0002% | |
| 80–89 | 38 | — | |
| 90–100 | 7 | — |
Finding 1: the web is almost entirely long tail
97.8 million domains — 82.9% of the entire index — score below 10. Across all linked domains the median number of referring domains is two. Not two hundred. Two. And 13% of domains have no discoverable inbound links at all.
At the other end the air is thin: only 289 domains score above 70, 45 above 80, and 7above 90. The gap between “a normal website” and “an authority site” is not a gap — it is a cliff.
The practical read: if your site has 20 referring domains you are already ahead of roughly 97% of the web. Worth knowing before you conclude your link building has failed.
Finding 2: more links does not mean more authority
Look at two adjacent bands. Sites scoring 20–29 have a median of 22 referring domains. Sites a full band higher, at 30–39, have a median of only 15.
A stronger band with fewer links. This is not a rounding artefact — the mean moves the opposite way (80 for the lower band, 186 for the higher), which tells you the stronger band is carried by a few domains with extremely well-placed links, not by link volume.
The mechanism is simple once stated. Authority here is harmonic centrality — how close a domain sits to the rest of the graph. A link from a domain that is itself central carries the weight of everything behind it. Three links from central hubs beat three hundred from the periphery, and the data shows it plainly.
Counting referring domains as a KPI will mislead you. Fifty links from weak directories can leave authority flat, while five placements on genuinely central sites can move it. Measure the authority of the sites linking to you, not the number.
How many links each tier actually takes
For domains that do climb, here is the median (and mean) number of referring domains at each authority tier — real figures from the index, not a rule of thumb.
| Authority | Median referring domains | Mean |
|---|---|---|
| 40+ | 649 | 1,215 |
| 50+ | 5,098 | 8,819 |
| 60+ | 12,801 | 28,754 |
The most authoritative domains on the web
The 20 domains closest to the centre of the entire link graph. Note the top spot: googleapis.com outranks google.com, because it is referenced from the code of a vast share of all websites. Infrastructure beats content when you measure by links.
| # | Domain | Authority |
|---|---|---|
| 1 | googleapis.com | 100 |
| 2 | facebook.com | 96 |
| 3 | google.com | 94 |
| 4 | instagram.com | 93 |
| 5 | googletagmanager.com | 91 |
| 6 | youtube.com | 90 |
| 7 | twitter.com | 90 |
| 8 | gstatic.com | 89 |
| 9 | linkedin.com | 88 |
| 10 | gmpg.org | 88 |
| 11 | cloudflare.com | 87 |
| 12 | gravatar.com | 87 |
| 13 | apple.com | 86 |
| 14 | wikipedia.org | 86 |
| 15 | jsdelivr.net | 85 |
| 16 | pinterest.com | 85 |
| 17 | wordpress.org | 85 |
| 18 | youtu.be | 84 |
| 19 | goo.gl | 84 |
| 20 | x.com | 84 |
Authority by top-level domain
.com holds 49.6% of the entire web (58.5 million domains). But size and authority are different questions: among TLDs with over 200,000 domains, .jp has the highest average authority (7.05), ahead of .io, .ru and .org — while the sprawling .com average is dragged down by its long tail.
| TLD | Domains | Avg authority |
|---|---|---|
| .com | 58,462,385 | 5.18 |
| .org | 6,685,720 | 6.38 |
| .de | 5,024,647 | 4.86 |
| .net | 3,768,720 | 5.85 |
| .uk | 3,195,700 | 5.55 |
| .ru | 2,322,243 | 6.53 |
| .fr | 1,671,711 | 5.15 |
| .jp | 1,166,503 | 7.05 |
| .info | 1,117,250 | 5.55 |
| .au | 1,043,117 | 5.77 |
| .ca | 887,190 | 6.02 |
| .io | 348,257 | 6.65 |
Method
Source data is the Common Crawl domain-level web graph, release cc-main-2026-may-jun-jul: 117,963,409 domains and 2.77 billion domain-to-domain edges (7.1 billion across the two releases we hold). Authority is harmonic centrality, published by Common Crawl; we map its rank onto 0–100 with a logarithmic curve, so score = 100 × (1 − ln(rank) / ln(N)). Referring-domain figures are distinct linking domains per target, computed directly over the edge list.
Two honest caveats. First, this is a domain-levelgraph — one referring domain may link from a thousand pages, and that counts once. Second, our score is not Moz's DA or Ahrefs' DR; those run on proprietary crawls with different coverage. The shape of the distribution should hold; the exact numbers will not match.
Check your own domain against this
Every number above comes from the same index that powers our free tools. Look up any domain and see exactly where it falls in this distribution.