How to run a backlink audit (step by step)

A backlink audit is one of the few SEO tasks that pays off whether your link profile is healthy or a mess. Done well, it tells you which links are actually working, which pages are quietly hoarding authority, where you have lost links worth reclaiming, and whether anything in your profile is genuinely worth worrying about. Done badly, it sends you chasing phantom problems and disavowing perfectly good links. This guide walks through how to run a backlink audit properly: which tools to use (including a free option most articles skip over), how to consolidate your data, how to judge link quality on signals that matter, and how to tell a genuinely risky link from a harmless one. We will also be straight about something the SEO industry tends to overstate. Most sites do not need to disavow anything, and Google says as much in its own documentation. The goal here is a clear head and an action plan, not a panic.

What a backlink audit is and when to run one

A backlink audit is a structured review of every link pointing to your website. The aim is to understand your link profile as a whole and link by link: how many referring domains you have, where they come from, how authoritative and relevant they are, what anchor text they use, and whether any of them pose a real risk. If you are new to the wider topic, our building links sets the context, and our explainer on what backlinks are and how you earn them covers the fundamentals.

An audit is not only a hunt for bad links. In practice the more valuable output is usually positive: which pages attract links, which content formats earn them, which relationships are worth nurturing, and which lost links you can win back. Backlinks remain a meaningful ranking factor, and the scarcity is real. Ahrefs found that 66.31% of pages have no backlinks at all, so every quality link you already hold is an asset worth protecting.

You do not need to audit constantly. For most established sites, once or twice a year is sufficient. Run an audit sooner if you have just taken over a site or client and want to know what you have inherited, if you are about to migrate or restructure URLs, if rankings have dropped without an obvious on-page cause, or if you have received a manual action notice in Google Search Console. Outside those triggers, light ongoing monitoring is enough.

The tools you need (free and paid)

You can run a credible audit with free tools alone. The single most important source is Google Search Console, which is free and shows the links Google itself has recorded for your site. Inside the Links report you can see your top linked pages, top linking sites, and most common anchor text, then export the full list. Because this data comes straight from Google, it is the most authoritative view of what Google actually associates with your domain.

Search Console has limits. It caps the number of links it exports and gives you no quality metrics, no referral traffic figures, and limited filtering. That is where paid link building tools earn their place. Ahrefs and Semrush both maintain large independent link indexes and layer on useful signals: domain and URL level authority scores, estimated organic traffic to the linking page, anchor text breakdowns, link attributes, and the ability to filter and sort thousands of links quickly. Other options such as Moz and Majestic offer their own metrics.

One caveat worth stating plainly. Authority scores such as Domain Rating and Authority Score are proprietary metrics invented by each tool. They are useful for comparison and triage, but they are not Google scores and Google does not use them. Treat them as a helpful shorthand, not gospel. If budget is tight, start with Search Console and add one paid tool only when you need the extra depth.

How to export and consolidate your link data

Each tool sees a slightly different slice of the web, so the cleanest starting point is to pull data from every source you have access to and combine it. Export the full backlink or referring domains report from each tool as a CSV, plus your Search Console export, then bring them into a single spreadsheet.

The practical goal is one row per referring domain rather than one row per individual link. Most sites have far more links than domains, because a single site can link to you dozens of times through sitewide footers, navigation, or repeated mentions. Auditing at the domain level keeps the job manageable and reflects how search engines think: ten links from one site count for far less than links from ten different relevant sites.

De-duplicate your combined list so each referring domain appears once, keeping the strongest available data point for authority and the anchor text you most want to review. Add a few working columns you will fill in as you go: a relevance note, an authority figure, an estimate of real traffic, and a simple verdict such as keep, monitor, or investigate. This sheet becomes the backbone of the audit. It also gives you a clean before and after record, which is invaluable when you want to measure progress later.

How to assess link quality

Once your data is consolidated, judge each referring domain on the signals that genuinely indicate value. No single metric tells the whole story, so weigh several together. For a deeper treatment, see our guide to spotting a good link.

• Relevance. Is the linking site topically related to yours, or is the specific page relevant even if the site is broad? A link from a respected publication in your field is worth more than a higher authority link from a site with nothing to do with your subject. Relevance is the signal most audits underweight and the one that matters most.
• Genuine authority. Use authority scores as a first-pass filter, then sanity check by eye. A low score can simply mean a newer or smaller site rather than a bad one, and a high score does not guarantee the link is meaningful. Look at the site itself: real content, a real audience, recognizable as something a person would visit.
• Real traffic. The best links send actual visitors. Tools that estimate organic traffic to the linking page help you spot links that are likely to drive referral clicks as well as ranking signals. A link on a page nobody reads is worth far less than one on a page that ranks and gets visited.
• Placement and attribute. A contextual link inside the main body of an article carries more natural weight than one buried in a sidebar, footer, or author bio. Check the link attribute too. The difference between follow and nofollow links matters, but in short a healthy profile contains a natural mix rather than a wall of identical follow links.

No single signal is decisive on its own, so weigh them together. If a domain looks strong on two or three counts and weak elsewhere, settle it by asking whether a real reader would ever click that link, because that is the question search engines are ultimately trying to answer too.

How to analyze anchor text distribution

Anchor text is the clickable text of a link, and its distribution is one of the clearest signals of whether a profile looks natural or manufactured. When people link to you of their own accord, they tend to use your brand name, your URL, or a casual phrase like the name of the article. A profile dominated by branded and natural anchors is a healthy sign.

The warning sign is an unnatural concentration of exact-match, keyword-rich anchors. If a large share of your links use commercial phrases you are clearly trying to rank for, that pattern can look like manipulation, because it rarely happens organically at scale. To review this, sort your anchor text export by frequency and look at the overall shape. A spread weighted towards branded and URL anchors, with keyword anchors as a minority, is what you want to see. Our your anchor text mix covers healthy ratios in detail, including why brand-led anchors should usually dominate a natural profile.

If you do spot over-optimization, the fix is rarely to disavow. It is more often to dilute the pattern by earning more natural, branded links over time through steady link building, so the keyword-heavy anchors shrink as a proportion. Anchor text is a distribution to manage, not a single link to delete.

How to spot genuinely risky links versus harmless ones

This is where audits most often go wrong. SEO tools assign third-party toxicity or spam scores to links, and it is tempting to treat a high score as proof of harm. It is not. These scores are estimates produced by each tool’s own model. They are not metrics from Google and they do not tell you how Google treats a link. A high toxicity score is a prompt to look more closely, not a verdict.

Manipulative links

It helps to separate the truly risky from the merely ugly. Genuinely risky links are almost always ones you or a previous agency built to game rankings rather than to reach readers, and they follow a few recognizable patterns.

• Private blog networks. Clusters of sites that exist only to pass links, with no audience, no editorial standards and often shared owners, templates and hosting.
• Mass-produced spam. Directory and comment links generated at scale and on purpose, with no relevance to your topic and no human involved.
• Automated link schemes. Bulk exchanges and tools that fire off links by the thousand, leaving footprints that are trivial to detect.
• Over-optimized anchors. Large numbers of exact-match commercial anchors pointing at money pages, which looks engineered rather than earned.

These are the patterns Google’s spam policies target, and they share one trait: no real publication, no editorial judgment and no reader who would ever follow the link. A placement on a genuine site that vets what it publishes is a different thing entirely, and it is not what an audit should be flagging.

False positives

Harmless links are far more common and look worse than they are. Scraper sites that copy content, low-quality directories you never submitted to, spam that appears in a sudden unsolicited spike, and odd foreign-language sites are nearly always something Google ignores on your behalf. Google has been clear that it largely discounts spam links automatically rather than penalizing the target site for them. Our guide to spammy links goes deeper, but the rule of thumb is simple: if you did not build it and it appeared without your involvement, it is very unlikely to be a problem you need to act on.

Why you do not need to disavow

Here is the part of the audit that most guides get wrong. You do not need to disavow. Google ignores spam links, and it has been ignoring them automatically for years, which is why the tool sits unused by the overwhelming majority of sites that rank perfectly well.

Google’s own guidance says so. Its Search Console Help documentation states that in most cases Google can assess which links to trust without additional guidance, so most sites will not need the tool at all. John Mueller has made the same point repeatedly: links a spammer drops are links Google should simply ignore, and its systems are very good at ignoring them. Since Penguin became part of the core algorithm in 2016, spam links are devalued rather than held against the site, and SpamBrain neutralises link spam so it never counts.

Disavowing is not a free precaution either. Disavow files are blunt, and people routinely cut off links that were quietly helping them. Gary Illyes has said that “if it were up to me, I would do away with the disavow link tool because, often enough, it hurts more than it helps many sites,” and Bing retired its equivalent tool in 2023. The one genuine exception is an unnatural-links manual action in Search Console, which is uncommon and which Google notifies you about. If you have one, our guide on disavowing backlinks covers the mechanics. If you do not, leave the links alone.

How to turn the audit into an action plan and ongoing monitoring

An audit is only useful if it produces decisions. Most of those decisions should be constructive rather than defensive. Work through your findings and turn each into a task.

• Fix wasted links. Identify links pointing to pages that now return a 404 error and redirect them to the most relevant live page, so the authority is not lost. Look too for pages that attract plenty of links but little traffic, which often signals a content opportunity rather than a problem.
• Reclaim what you lost. Links disappear when pages are removed, redesigned, or edited. High-value lost links are worth recovering with a short, polite outreach email. Our guide to recovering lost links covers the process for finding and fixing lost backlinks.
• Double down on what works. Note which pages and content formats earn the most links, and let a competitor backlink analysis show you which sources are already linking to rivals in your space. Backlinko’s analysis found that the number one result has 3.8 times more backlinks than positions two to ten, so concentrating link-earning effort on the pages that already attract them is sound strategy. If statistics roundups or original research pull links, make more of them.
• Set a monitoring cadence. You do not need to repeat a full audit often, but you should keep an eye on new and lost links between audits. Most tools can alert you to sudden changes, and Search Console will flag any manual action directly. Reviewing new links monthly and running a full audit once or twice a year keeps you informed, and a clean record lets you measure progress against your the metrics that matter over time.

Work through those and the picture gets clear quickly. The point of an audit is not to hunt for problems, it is to understand what you already have so you know what to build next.