What is Domain Authority and How to Check It
Domain authority is one of those terms that gets thrown around constantly in SEO conversations, but the definition shifts depending on who you ask and which tool they use. If you have ever searched for your domain authority score, received three different numbers from three different tools, and wondered which one actually matters, you are not alone. This article explains what domain authority really means, why the number varies so much across platforms, and introduces a metric that most SEO tools do not yet track but almost certainly will within the next few years.
What domain authority actually measures
Domain authority is not a single official metric. It is a category of measurements, each designed to estimate how well a domain might perform in search results based on the quality and quantity of links pointing to it. The term itself was popularised by Moz, which created a proprietary score called Domain Authority (DA) that runs from 0 to 100. Ahrefs has its own version called Domain Rating (DR). SEMrush calls theirs Authority Score. Majestic uses Trust Flow and Citation Flow.
Every one of these scores is calculated differently, uses a different dataset of known links, and is updated on a different schedule. None of them are official Google metrics. Google has never published a publicly accessible domain authority score, and the old Google PageRank toolbar was retired in 2016. What you see when you check domain authority in any tool is that company's best estimate based on the backlink data they have crawled.
That matters because link databases are expensive to build and maintain. Smaller tools have smaller link databases, which means their scores can be significantly lower than the reality of a site's actual link profile. A domain that Moz scores at 30 might score 45 in Ahrefs simply because Ahrefs has crawled more of the web and found more of the links pointing to that domain.
Why the number keeps changing
Domain authority scores fluctuate for a few reasons. The most obvious is that the web itself changes: new links are created, old links disappear, and domains gain or lose relevance over time. But scores also change when the tool updates its algorithm or expands its link database, even if nothing has changed on your site or in your backlink profile. It is not unusual to see a domain authority score drop by ten points overnight without any negative SEO event, simply because the tool recalibrated how it weights certain types of links.
This volatility is one of the reasons experienced SEOs treat domain authority as a directional indicator rather than a precise measurement. The trend matters more than the number. A domain steadily moving from 20 to 30 to 40 over eighteen months is doing something right, regardless of which tool you use to measure it. A domain bouncing between 25 and 35 with no clear direction has a less convincing story to tell.
Domain authority scores from different tools are not comparable to each other. A Moz DA of 40 does not equal an Ahrefs DR of 40. They measure related but different things. When benchmarking against competitors, always use the same tool for both domains.
The link graph behind the score
Most domain authority tools work by crawling the web, building a map of which domains link to which other domains, and then running algorithms over that map to assign authority scores. The two most famous algorithms for this are PageRank, originally developed by Google founders Larry Page and Sergey Brin, and Harmonic Centrality, a graph theory metric that has been used by Common Crawl since 2017.
PageRank measures endorsement. If a highly authoritative domain links to you, your PageRank increases. Links from less authoritative domains still count, but they count less. The core idea is that authority flows through links the way votes flow in an election: a vote from an influential person carries more weight than a vote from someone with no following.
Harmonic Centrality measures something different: your position in the overall network. Rather than asking who links to you, it asks how close you are to every other domain on the web. Domains that sit at the centre of the web's link graph, connected to many other well-connected domains, have high Harmonic Centrality. Domains that exist on the edges of the web, reachable only through long chains of links, have low Harmonic Centrality.
The practical difference matters more than it sounds. Wikipedia, for example, ranks number fourteen in Harmonic Centrality in Common Crawl's most recent web graph data but only number thirty-seven in PageRank. Despite that gap, Wikipedia is consistently one of the most cited sources in AI-generated responses. Its position as a central hub in the web's link topology appears to contribute to its visibility in ways that raw link authority alone does not capture.
Why Common Crawl's web graph data is different
Common Crawl is a nonprofit organisation that has been crawling and archiving the entire web since 2008. Every month it publishes billions of web pages alongside something called Web Graph data: a complete map of how domains link to each other across the entire crawl, with Harmonic Centrality and PageRank scores calculated for over 120 million domains.
This data is publicly available and free. But it arrives as multiple gigabytes of compressed files that require significant technical work to process and query. Until recently, it was essentially inaccessible to anyone who was not a data scientist or a developer comfortable working with large datasets.
The reason this data matters increasingly to SEOs is that Common Crawl's web graph is one of the primary sources used to train large language models including ChatGPT, Claude, and Gemini. Common Crawl uses Harmonic Centrality to determine which domains get crawled more frequently. Domains with stronger Harmonic Centrality appear more often in Common Crawl's archives. More appearances in the archive means greater representation in the training data that AI models are built on. Greater representation means AI models are more likely to know your site exists and reference your content in their responses.
This is not the same as traditional Google SEO. You can rank well in Google and still be largely invisible to AI systems if your domain sits on the periphery of the web's link graph. Understanding your Harmonic Centrality gives you a signal that traditional domain authority tools simply do not provide.
Harmonic Centrality and traditional domain authority metrics measure related but distinct things. A domain can have strong Moz DA or Ahrefs DR (indicating good backlink authority) while still having weak Harmonic Centrality (indicating a peripheral position in the web's link structure). Both signals are worth tracking.
How to check your domain authority
For traditional domain authority scores, the most widely used free options are Moz's Link Explorer, Ahrefs' free tier, and SEMrush's domain overview tool. Each will give you a different number for the same domain, and each is useful for different purposes. Moz's DA score has the longest history and is most commonly referenced in the industry. Ahrefs' DR is generally considered the most accurate reflection of raw link authority because Ahrefs maintains one of the largest active link databases. SEMrush's Authority Score blends link data with traffic signals for a more holistic view.
For Harmonic Centrality and Common Crawl-based metrics, the options have historically been limited to manually downloading and processing Common Crawl's raw data files. That is a multi-day project even for experienced developers. Two tools now make this data accessible without technical setup: the online version of Tom's AI Rank Checker at tomdahne.com/ai-rank-checker, which queries a full 120 million domain database directly from the server, and the free desktop application which lets you check unlimited domains offline after downloading a local database.
Both tools show you Harmonic Centrality rank, PageRank from the Common Crawl web graph, an AI Visibility Score from 0 to 100, your tier classification across six levels from Elite to Long Tail, and a percentile showing how your domain compares to all 120 million domains in the dataset.
What a good domain authority score looks like
The honest answer is that there is no universally good score. Domain authority is always relative to your competition. If you are a small local business competing for keywords in your region, a Moz DA of 20 might be more than enough to outrank your competitors. If you are a SaaS company trying to rank for high-competition keywords against established players with DA scores of 70 and above, a score of 40 puts you at a structural disadvantage regardless of how good your content is.
The same relative thinking applies to Harmonic Centrality. In Common Crawl's most recent web graph covering 120 million domains, breaking into the top one million means you are in the top one percent of all domains on the web. That is the threshold where AI training data representation starts to become meaningful. Smaller or newer sites will typically sit in the Long Tail tier, which is rank 1,000,001 and beyond. That is not a failure state. It is simply the starting point for most legitimate domains, and it improves as you earn more links from well-connected sites.
The most useful question is not what a good score looks like in isolation but how your score compares to your direct competitors, and whether that score is trending in the right direction month over month.
How to improve domain authority
Traditional domain authority improves through link building: earning links from other websites, particularly from sites with strong authority scores of their own. The tactics for this have not changed fundamentally in years. Publishing genuinely useful content that other sites want to reference, building tools or resources that attract natural links, pursuing relevant directory listings and industry citations, and outreach to sites in your niche are all proven approaches.
Harmonic Centrality improves through the same fundamental activity, but the emphasis shifts slightly. A single link from a site that sits close to the centre of the web's link graph, like a major publication, a university, or a government site, can move your Harmonic Centrality more than dozens of links from smaller, peripheral sites. It is not just about accumulating links. It is about earning links from sites that are themselves well-connected to the wider web.
Practically, this means being selective about where you spend your link-building effort. A link from a small niche blog that has no links from anywhere significant will contribute less to your Harmonic Centrality than a link from a well-established resource site that itself has strong connections. The topology of your backlink profile matters as much as its size.
Common Crawl publishes new web graph data monthly. Because AI models are trained on accumulated data over time, improvements to your Harmonic Centrality today may take months to influence how AI systems perceive your domain. This is a long-term signal, not a quick win.
Domain authority and AI search
The connection between domain authority signals and AI search visibility is still being researched and debated. What is clear from multiple large-scale citation studies is that the domains most frequently cited by AI systems like ChatGPT, Perplexity, and Google's AI Overviews tend to have both strong traditional domain authority scores and strong Harmonic Centrality in Common Crawl's web graph. Whether this is correlation or causation, and how much weight each signal carries in AI responses, is not fully understood.
What is actionable now is that Harmonic Centrality gives you a measurable proxy for your presence in AI training data. You cannot directly control what AI models know about you or how they reference your site. But you can monitor where you stand in the web's link topology, track whether that position is improving over time, and make link-building decisions informed by which potential linking sites would most improve your network position.
The shift from "index and rank" to "train and retrieve" that is underway in search means that domain authority signals matter in a new way. Being well-connected in the web's link graph is no longer just about Google rankings. It is increasingly about whether AI systems know your domain exists at all.
Checking your own numbers
If you want to see where your domain sits in Common Crawl's web graph right now, Tom's AI Rank Checker lets you check up to five domains at once with no account required. Enter your domain alongside a few competitors and you will immediately see how your Harmonic Centrality and PageRank compare across the full 120 million domain dataset, along with your AI Visibility Score and tier classification.
The free desktop version extends this with batch checking of up to 1,000 domains at once, a compare view for side-by-side analysis of up to ten domains, CSV export, and fully offline operation after the initial database download. Both versions are free and require no subscription.