Domain Rating, Domain Authority, and Trust Flow read the link graph differently.
Ahrefs DR computes a logarithmic PageRank approximation. Moz DA trains a machine-learning model against ranking outcomes. Majestic Trust Flow splits the score into a quality axis (proximity to a curated seed set) and a volume axis (raw inbound link count). The three encode different signal types and correlate but do not converge on identical numbers. None equals what Google actually scores.
Four reads on the single-number authority score.
Vendor metrics collapse multi-axis link evaluation into a 0-100 score for tool convenience. The algorithm evaluates source-domain TrustRank proximity, topical relevance, link position, co-citation neighborhood, and per-URL anchor distribution separately. The single number is a filter, not the signal.
Ahrefs Domain Rating: link-graph PageRank approximation on a 100-point scale.
Ahrefs Domain Rating (DR) computes a logarithmic PageRank-approximation score from the link-graph Ahrefs crawls. The score reads the inbound profile to a domain (referring domains, link-graph position, source-domain DR distribution) and feeds the calculation through a damping factor similar to the original PageRank algorithm. DR scales 0 to 100 with logarithmic distribution: DR 70 reflects approximately 10x the link-graph weight of DR 60. The score correlates with ranking but does not measure all of what Google scores.
Moz Domain Authority: machine-learning score against ranking outcomes.
Moz Domain Authority (DA) trains a machine-learning model to predict ranking outcomes against Moz's index. The model reads multiple inbound-profile signals and outputs a 0 to 100 score calibrated against actual SERP positions. DA's calibration approach differs from Ahrefs DR's link-graph computation: DR computes PageRank approximation; DA predicts ranking. The two converge on similar 0 to 100 distributions but encode different signal types. Sites comparing prospect lists across vendors see correlated but non-identical numbers.
Majestic Trust Flow and Citation Flow: split-axis link evaluation.
Majestic splits authority into Trust Flow (a quality metric reading proximity to a curated seed set of high-reputation reference sites) and Citation Flow (a volume metric reading raw inbound link count). The split-axis frame approximates the TrustRank seed-set proximity model from the 2004 Gyongyi/Garcia-Molina/Pedersen paper. Trust Flow / Citation Flow ratios above 0.8 indicate a high-quality inbound profile; ratios below 0.5 indicate a low-quality profile loaded with citation-flow volume without trust-flow proximity. The split is closer to what TrustRank actually evaluates than the single-number DR or DA frame.
The single-number score does not equal what Google scores.
Vendor authority metrics collapse multi-axis link evaluation into one number for tool convenience. The algorithm evaluates source-domain TrustRank seed-set proximity, source-page topical relevance (Content Warehouse mention-context scoring), in-page link position, co-citation neighborhood, the destination URL's per-URL anchor distribution (Penguin 4.0), and the link-graph network density (SpamBrain) separately. A DR 30 link from a topically-coherent vertical publication often outperforms a DR 60 link from a generic news aggregator because the multi-axis profile is different even when the single-number score favors the second.
Methodology questions we get during the audit conversation.
What is a good Domain Rating to target for backlinks?
The wrong frame. Ranking depends on the SERP competitor profile at the URL level, the source-domain topical relevance, the per-URL anchor distribution, and the link-graph network position. A DR threshold abstracted from those signals catches generic news aggregators (high DR, low topical relevance) and misses topically-coherent vertical publishers (lower DR, high topical relevance). The right prospect-list filter scopes against topical-cluster maps with DR as a secondary signal, not against DR as the primary filter.
Why do Ahrefs DR, Moz DA, and Majestic Trust Flow give different numbers?
The three metrics encode different signal types. Ahrefs DR computes a logarithmic PageRank approximation from the Ahrefs link-graph crawl. Moz DA trains a machine-learning model to predict ranking outcomes against Moz's index. Majestic Trust Flow scores seed-set proximity against a curated trust seed and Citation Flow scores raw inbound volume. The three indices crawl different link-graph subsets and apply different calculation approaches; the numbers correlate but do not converge on identical scores. Cross-vendor prospect lists typically use one metric as the primary filter and verify with a second.
Does Google use Domain Rating internally?
No. Domain Rating, Domain Authority, and Trust Flow are vendor metrics, not Google's internal signal. Google's ranking surface reads multi-axis signals: TrustRank seed-set proximity (the algorithmic ancestor of Majestic Trust Flow), source-page topical relevance (the Content Warehouse mention-context scoring surfaced in the May 2024 leak), in-page link position, co-citation neighborhood, anchor distribution per URL (Penguin 4.0), and link-graph network density (SpamBrain). Vendor metrics approximate components of the signal stack; the single-number score does not equal what the algorithm evaluates.
Is a low Domain Rating site worth pursuing for outreach?
Yes, when the topical relevance and TrustRank seed-set proximity are strong. A DR 30 vertical publication that consistently cites peer brands in the same topic cluster, sits one hop from a seed-set domain in the trust graph, and carries editorial-discretion inbound placements often passes more signal than a DR 60 generic aggregator. The prospect list scopes against topical-cluster maps and verifies the multi-axis profile; the DR threshold by itself filters out the strongest signal sources in many verticals.
DR and DA are filters. The audit reads the multi-axis profile the algorithm actually scores.
The audit pulls the inbound profile, decomposes the source-domain TrustRank proximity, source-page topical relevance, anchor distribution per URL, and co-citation neighborhood, then names the placements passing signal versus the placements the single-number score overstates.