Domain authority scores for link prospecting

Any authority style score starts from link or popularity data about the whole web, ranks every domain, and rescales the rank to a friendly range such as 0 to 100. The number says how prominent a domain is in that data. It does not say whether the site converts visitors, whether its editors answer email, or whether a link from it would be editorially valuable.

Because every tool picks its own data and its own formula, a score from one tool is not comparable with a score from another. Pick one source, keep using it, and compare domains only against each other inside that source.

Treat the number as a sorting key. It is good at separating a household-name site from an unknown one, and weak at separating two mid-sized sites. Read the pages before you pitch.

The LookValid score is our own calculation from two open datasets, and it is not a ranking published by anyone else. The first is the Common Crawl web graph. Common Crawl publishes host-level and domain-level graphs, and the domain graph is built by aggregating the host graph at the pay-level domain, using the Public Suffix List ¹. Its tooling computes PageRank or harmonic centrality over such graphs ². We read each domain's position in the domain graph for both measures.

The second is the Majestic Million, a list of the one million domains with the most referring subnets, released under a Creative Commons Attribution 3.0 license ³. We read the domain's rank in that list.

Each position becomes a signal from 0 to 100 on a log scale of its list: signal = 100 x (1 minus log10(position) divided by log10(list size)). Position 1 scores 100 and the last position of the full list scores 0. A list that does not hold the domain contributes 0. The final score is the rounded weighted sum.

We keep the best 5 million domains of the Common Crawl ranks. A domain below that cut, and below the Majestic list, comes back as found: false with a reason, which is free. That means the domain ranks below both lists, not that it has no authority. A subdomain with no rank of its own is matched to its parent domain, and the response says so in matched_domain.

The table below comes from a captured example response for wikipedia.org. You can check the arithmetic yourself: 0.4 x 85.9 + 0.2 x 81 + 0.4 x 82 = 83.36, which rounds to 83.

The response also returns the formula, the weights and the Common Crawl release name, so every score can be recomputed from the raw positions.

• Score the whole prospect list once, then sort. Start with the top third and read those sites first.
• Set a floor, not a target. Dropping domains below a threshold removes noise without pretending that 61 is meaningfully better than 59.
• Keep the reasoning. The signals object shows which list put the domain where it is, which helps when a site is strong in one list and absent from the other.
• Check relevance by hand. A high score on an unrelated topic is rarely a good link to chase.

One lookup costs 10 credits when it finds the domain, and a repeat of the same lookup inside the cache window is free. This call scores a single domain; wrap it in a loop for a list.

1curl -X POST https://YOUR_API_HOST/v1/domain-authority \
2  -H "Authorization: Bearer $LOOKVALID_API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{"domain":"wikipedia.org"}'

Try it without an account in the domain authority checker, or read the domain authority product page for the full response.