Skip to content
SEOCheckGet $100 free

Metrics

Three authority scores, three different answers

A domain can be 62 in one tool and 38 in another. Neither is wrong. They are measuring different things.

SEOCheck research team ·

Three carved stone pillars in a forest with a person inspecting one closely
Short version

Google does not use DR, DA or Authority Score. They are third-party estimates of link strength. Useful for sorting a list. Poor as a verdict on a site.

The usual three, side by side

Based on each vendor's public descriptions, simplified. Our own number is included for contrast.
MetricFromMain ideaWatch out for
Domain Rating (DR)AhrefsStrength of the linking-domain profileNot Moz DA, despite the look-alike
Domain Authority (DA)MozPredicts ranking likelihood from a modelModel changes shift old scores
Authority Score (AS)SemrushBlend of link and traffic signalsTraffic estimates can swing it
DR (ours)SEOCheckLinking-domain count on a log scaleOne input only, monthly refresh

Why the numbers do not match

Each vendor crawls a different slice of the web. They also weigh links differently. One counts domains, another adds traffic, a third trains a model against ranking data. So the same site lands in three places.

Our formula is public and short. You can reproduce it with a calculator. See the DR checker page for the worked table. That is its main virtue and its main weakness, because it ignores link quality entirely and only counts distinct domains.

The practical rule is to pick one source per spreadsheet. Never average across vendors, since doing so gives a number that means nothing to anyone, including you.

There is also the matter of time. Moz and Ahrefs rebuild their indexes on different schedules, and ours refreshes about monthly. A site that gained a big link on the 3rd looks different on the 4th in a daily tool and on the 30th in ours. Plan comparisons around that. For a list-level approach see the pipeline guide.

Carpenter choosing a mallet from a workbench while ignoring other tools

Which score fits which job

The three scores answer different questions, so the right one depends on the job in front of you. Picking the score first and the job second leads to arguments about numbers that were never meant to settle the question.

Start with what you need to decide. Then pick the source that answers it, and write that choice at the top of the sheet.

Sorting a long list

When you have hundreds or thousands of domains and need a rough order, any consistent score works. The important part is consistency. Use one source for the whole list, so the order means the same thing from top to bottom. Sorting is the job a score does best, so let it do that job and no more.

Judging a single site

For one site, a score is a starting point and nothing more. Look at the pages, the topic and the links the site earns. A score that looks odd for a single site often hides a single reason, such as one large old link, which the page view will show quickly.

Reporting to a client

If a client already reads one vendor score, keep reporting it. Adding a second score in the same chart only creates a question nobody can answer. Use the other source in your working files, label it clearly and keep it out of the headline. A short footnote with the source and the date is usually enough, and it saves a long email later.

Explaining a score to a non-technical reader

A client or a manager rarely needs the formula. They need to know what the number is for. Say that it counts linking domains, that it is a rough filter, and that it is not the same as the score they already see elsewhere. Two sentences are usually enough.

Avoid using the number as a grade. A label such as strong or weak invites an argument about the number instead of the decision.

Using two scores on purpose

Sometimes two sources are worth keeping side by side. One can be the broad filter you run on every list, while the other is the number a client already reads. Keep them in separate columns, label each one, and never merge them into a single figure.

Write the reason for each column at the top of the sheet. That note stops the next person from averaging them by accident.

Myth versus reality

Myth: high DR means it ranks

Reality: scores describe links. Rankings depend on relevance, intent, content and many other signals.

Myth: low DR means worthless

Reality: a small, focused site can send a very relevant link. Sort by score, then read the page.

Myth: you can raise it fast

Reality: on a log scale each new domain adds less than the last. Chasing the number invites bad links.

Myth: one tool is the truth

Reality: they are estimates with different blind spots. Trust trends more than absolute values.

Why some domains look so strong

Some domains carry scores that look impossible for a new project. Their links were earned slowly. Many years, across many unrelated sites, is how such a profile is built. A score that counts linking domains rewards exactly that kind of history.

It helps to picture the work behind the number. A large domain that has been cited for a decade has accumulated links from thousands of places, each one added by someone else.

There is no magic threshold

People often ask at what authority level content starts to show up. No vendor publishes a rule that works everywhere. Rankings depend on the query, the intent and the page, and a score cannot fix a page that does not answer the question.

Watch your own trend instead

Your score matters most as a line on a chart. Check it monthly with the same source, and note what you published or earned in between. A rising line with a clear cause is better evidence than any single value you can compare to a rival.

What a log scale hides

On a log scale the early links count for a lot and the late ones count for little. So a domain can look far stronger than a newcomer while adding relatively few new referring domains each month. That is why chasing the number directly tends to disappoint.

Old domains and new projects

A new project can look weak next to an old site on every score, and that is an honest result, not a verdict on the project. Compare the new site with others of its age, and track its own line over time. The gap to the old giant will not close quickly, and it was never meant to.

Look up a domain

Free lookup using our domain action. Compare the result to whatever your other tool says and note the gap. Details in the DR checker guide.

Gardener kneeling to inspect tree roots and soil in a garden bed

A quick routine to check domain quality

Start with the score, but only as a filter. Throw out anything very low if the job needs reach. Then open the actual linking pages for the ones that survive. Reading beats guessing.

Ask four plain questions. Does the site publish real content? Would a person land there on purpose? Is the topic near yours? Does the page that carries the link get any attention? A yes to three of four beats any number. Use these questions whenever you need to check domain quality by hand.

For long lists, do the filtering in code. A single bulk_dr call scores up to 2,000 domains for one credit, so the sort is almost free. Then read the top fifty by hand. Quick. See the competitor teardown for the same logic applied to rivals.

Turn scores into a repeatable check

A check you can repeat is worth more than a clever one-off. The steps below are plain on purpose, so anyone on the team can run them and reach the same result.

Keep the sheet short. Every extra column is a column someone must keep filled in.

Set the filter before the list

Decide the cut-off before you open the list. If you choose it after seeing the scores, you will tend to pick a line that keeps the domains you already like. Write the number in the sheet and leave it there for the whole project.

Read five pages by hand

After the filter, open five pages from the survivors at random. Check whether they are real pages, whether they are on topic and whether the link would sit inside the text. Five is enough to catch a pattern, and small enough to finish in one sitting.

Write down why you kept or dropped a domain

One line per domain is enough. Say what made you keep it, or what made you drop it. A month later, that line tells you whether your rule still makes sense, and it saves you from relitigating every row.

The link quality checklist gives a longer list of questions for that line, if you want more structure.

Retest the filter each quarter

A cut-off that made sense for one list may be wrong for the next. Once a quarter, look at which domains the filter kept and which it dropped. If the drops include results you would have wanted, adjust the line and write down why.

Our position

Use scores to sort, humans to decide

We built a simple number because a simple number is scriptable. We will not claim it replaces Ahrefs DR, Moz DA or Semrush Authority Score. It answers a narrower question, cheaply.

  • Good: ranking 2,000 domains by size of link profile.
  • Weak: judging trust, topic fit or spam risk.
  • Never: mixing vendors in one column.