Ahrefs Swaps Domain Rating on 1 December. About Half of All Scores Go Missing

I pulled ten domains through Ahrefs' own lookup page on 6 November 2026. The numbers below come from that run, not from the announcement.

Domain Rating scores a site's backlink profile against every other site in the Ahrefs database. DR4 keeps that job and changes two inputs.

First, Ahrefs now counts links from sites it estimates carry organic search traffic. That shrinks the pool of domains able to lift anyone else's score. Only around half the domains in the database come out of the change with a DR4 score at all.

Second, subdomains get their own rating. help.ahrefs.com no longer borrows the score of ahrefs.com.

Nothing else about the metric moves. Content quality, domain age and traffic estimates still play no part in it. Links remain the only input.

Read the announcement: https://ahrefs.com/blog/domain-rating-dr4/

Ahrefs built a demo page for the preview, so anyone can look up a score before 1 December.

Domain	DR3	DR4	Change
help.ahrefs.com	91	75	-16
play.google.com	100	96	-4
blogs.microsoft.com	96	89	-7
insights.samsung.com	92	74	-18
finance.yahoo.com	94	92	-2
medium.com	94	95	+1
notion.site	92	27	-65
github.io	94	64	-30
facebook.github.io	94	87	-7

My figures and Ahrefs' own table differ by a point or two in places. The scores move between pulls, so treat this table as a reading taken on 6 November 2026.

The split is the point. github.io scored 94 on links pointing at pages hosted under it, and DR4 scores the domain on its own links, so it lands at 64. facebook.github.io keeps 87 because sites link to that host by name.

Run your own lookup: https://detailed.com/dr-v4/

thesharpdigital.com reads 33 under DR3 and 33 under DR4. I checked it twice the same morning.

A flat score tells you one thing, which is that the pool feeding it did not shift enough to move a 33. It does not tell you our links are clean.

Small sites should expect small swings. Domain Rating is relative and logarithmic, so a move from 54 to 55 is easier than a move from 84 to 85.

Ahrefs labels some domains as spam and keeps them out of link alerts. Domains with that label and a DR3 score between 30 and 70 lost a median of 15 to 20 points. Domains without the label barely moved at all.

Across the 1,000 sites with the most estimated organic traffic, 64.3 percent gained, 15.8 percent held and 19.9 percent fell. The average fall was 7.3 points, a rounding error next to the spam drops.

The correlation is the part that justifies the change. Across 10 million hosts, DR4 tracks top ten Google positions at 0.359 against 0.140 for DR3. On root domains alone the old model still wins, at 0.407 against 0.304.

Ahrefs traded root domain accuracy for host accuracy. If you prospect links, that trade helps you. If you report one number for a whole brand, it complicates your week.

Domain Rating is not a Google ranking factor. Google has never used it, so a drop cannot cost you a position on its own.

The score is not a grade on your content either. Ahrefs says so plainly, and SEOs repeated it within hours, including Caleb Trevino.

What a lower score does say is narrow. Ahrefs doesn't see enough links from sites with organic traffic pointing at your page. Nothing about your site got worse overnight.

If you want to know which links count, start with the profile. Our guide on what makes a backlink valuable covers the signals that survive a model change.

1. Look up your root domain and every subdomain you quote in a report.
2. Screenshot each row with the date visible, because the preview is live and your December figure may land a point or two away from today's.
3. Look up three competitors in the same sitting, then change one line in your template to "DR4, checked on [date]".

A drop that hits everyone reads differently from a drop that hits you alone. The version and the date on that template line carry the honesty. Anything longer than this is a data project the metric doesn't deserve.

One sentence covers it: the metric got stricter, the link profile didn't get worse, and here is the new figure next to three competitors.

You can pull the number through the API too. Add version=dr4 to the request and you get DR4, while DR3 stays available for a short period after rollout.

If you want this check run for you before the switch, book a call and bring your competitor list.