Domain authority

Why it matters

Domain authority is the number most often quoted in a link building pitch, and the one most often misunderstood by the person receiving it. Understanding what it measures changes what you are willing to pay for.

It is genuinely useful for one job: comparing two sites quickly. It is not useful for predicting whether your page will rank, and it is not a target to optimise toward.

How it works

The score runs from 1 to 100 and is logarithmic, which is the single most important property and the one that causes the most confusion. Moving from 20 to 30 is far easier than moving from 70 to 80. The top of the scale is occupied by the largest sites on the web, so the distance between 80 and 90 represents an enormous gap in linking.

It is calculated primarily from the link graph: how many other sites link to this one, and how strong those linking sites are in turn. It is a prediction of ranking ability, trained to correlate with actual rankings, which is why it tends to move with them without causing them.

Several vendors publish their own version under different names. They are not interchangeable, they use different crawls of the web, and a site will usually score differently in each.

What Google documents

Google publishes a guide to its ranking systems. It names BERT, crisis information systems, deduplication, exact match domain handling, freshness systems, link analysis systems and PageRank, local news, MUM, neural matching, original content systems, removal-based demotions, passage ranking, RankBrain, reliable information systems, the reviews system, site diversity and spam detection. There is no site authority score in that list.

What the guide does say is that site-wide signals and classifiers are used. So the idea that a site's overall standing affects its pages is not fiction. What does not exist is a single published number that represents it, which is the gap a third-party score fills with an estimate.

Worked example

"What is a good domain authority" has no absolute answer, which is why the honest version is relative. A practical reading:

Compared to the sites already ranking for your target terms.
If the first page is populated by sites in the 20s and 30s
and you are at 25, the number is not what is holding you back.
If they are in the 70s and you are at 15, links are a real gap.

A new site typically begins in the single digits and climbs slowly. A score in the 20s to 30s is normal for an established small business site. Scores above 60 generally belong to sites with years of accumulated references, large publishers, or both.

The useful comparison is never against the whole web. It is against the ten results you are trying to join.

Common mistakes

• Treating it as a Google ranking factor.
  It is a third-party prediction. Google's documented ranking systems contain no such score.
• Chasing the number directly.
  Buying links to raise a score violates Google's link spam policies, and the risk attaches to your site. The score may rise while the rankings fall.
• Comparing across vendors.
  Different crawls, different models, different scales.
• Reading it as a quality score.
  It measures linking, not accuracy, usefulness or trustworthiness. A heavily linked page can still be a poor answer.
• Ignoring relevance.
  A link from a smaller site in your field is usually worth more than one from a larger site with no relationship to your subject, and the score cannot see the difference.

What to use instead

For deciding whether links are your constraint, compare your profile to the pages currently ranking rather than to a league table. For judging whether a specific link is worth pursuing, ask whether a relevant audience would plausibly follow it. That question maps closely to what Google's own guidance on helpful content and link spam describes, and it does not require a score at all.

Related terms

• Keyword difficulty
• SEO agency
• SEO services