Page Authority is Moz’s 0 to 100 score predicting how well a single page might rank. It’s a third-party estimate built from a third-party crawl, Google has never seen it, and every major tool has its own incompatible version.
What it measures
Moz calculates Page Authority mainly from link data in its own index, using a model trained to predict rankings. Ahrefs has URL Rating, Semrush has Authority Score, Majestic has Trust Flow and Citation Flow. Same rough idea, different crawls, different formulas.
| Tool | Page-level metric | Domain-level metric |
|---|---|---|
| Moz | Page Authority (PA) | Domain Authority (DA) |
| Ahrefs | URL Rating (UR) | Domain Rating (DR) |
| Semrush | Page Authority Score | Authority Score |
| Majestic | Trust Flow / Citation Flow (per URL) | Trust Flow / Citation Flow (per domain) |
The numbers are not interchangeable. A DR 60 site and a DA 60 site are making different claims computed from different data, and comparing them is a category error that appears in a depressing number of proposals.
Two things people forget
The scale is logarithmic
Moving from 20 to 30 is achievable with steady work. Moving from 70 to 80 is a different order of effort entirely. A linear reading of these scores makes progress look like stagnation at the top and a miracle at the bottom.
Google does not use them
Google has said this repeatedly, most often about Domain Authority. These are estimates built by third parties to approximate something Google does internally with data nobody else has.
The complication worth knowing
The 2024 Content Warehouse leak referenced a stored site-level authority value inside Google’s own systems, which reopened a long argument about whether Google has something conceptually similar to a domain score.
Read that carefully. The documents show a field exists. They do not show its weight, its current use, or how it’s computed. Treating a leaked field name as vindication of a third-party metric is exactly the overreach the leak invites and the evidence doesn’t support.
What these scores are genuinely good for
- Relative comparison. Is this page more linked-to than that one? Reliable enough.
- Tracking your own trend. Has our profile grown this quarter? Useful, as long as you remember the scale is logarithmic and the index gets recalibrated.
- Triage at volume. Sorting 500 link prospects into rough tiers before a human looks.
What they’re bad for
- Minimum thresholds for outreach. A DR 30 filter excludes relevant niche sites your customers read and admits irrelevant general ones they don’t.
- Client reporting as a performance metric. It measures links, not traffic, rankings or revenue. Reporting it invites a conversation about the wrong number.
- Cross-tool comparison. Covered above and worth repeating because it keeps happening.
- Pricing links. The entire paid-link economy is priced on DR, which tells you what that economy is optimising for.
Related
PageRank is the real mechanism these metrics approximate. Off-page SEO covers what actually moves them.