Link Building · July 23, 2026 · 8 min read
Backlink Audit Checklist: 18 Checks to Find Toxic Links Before They Tank You
Run a backlink audit using these 18 checks to identify toxic links, disavow spam, and protect your site's search rankings before damage sets in.
By FluxWriter Team
A backlink audit is one of the most high-leverage SEO tasks you can do — and one of the most frequently botched. Most site owners either ignore their link profile entirely or nuke it indiscriminately with disavow files, both of which cost rankings. This checklist walks through 18 specific checks to help you triage every link: keep it, flag it, or kill it.
Before You Start: Pull Your Link Data
You need two sources minimum. Export your backlinks from Google Search Console (Links report → Top linking sites → Export) and run a parallel crawl through Ahrefs, Semrush, or Moz. GSC is authoritative but incomplete; third-party tools catch domains Google knows about but hasn't credited. Reconcile both exports into one master spreadsheet before running any check below.
Section 1: Domain-Level Red Flags
Check 1 — Domain Spam Score
Pull the Moz Spam Score or Semrush Authority Score for every linking root domain. Any domain scoring 7+/17 on Moz Spam Score or below 10 on Semrush AS warrants a closer look. A score alone isn't a death sentence — a brand-new legitimate blog scores low too — but it's a filter.
Check 2 — TLD Distribution
Flag domains ending in .xyz, .top, .click, .gq, .tk, .ml, and .cf. These TLDs host a disproportionate share of link spam. If more than 5% of your referring domains use these extensions, that cluster deserves manual review.
Check 3 — Country of Origin vs. Target Market
If your site is English-language and US-targeted, a link from a Vietnamese auto-parts site or a Russian coupon blog almost certainly has no editorial basis. Filter by country code via the Ahrefs Referring Domains report and review any geographic cluster that doesn't make sense for your audience.
Check 4 — Domain Age
Use DomainIQ, Whois Lookup, or Ahrefs' domain registration data. Domains registered within the last 6 months that already rank highly in link-count metrics are often link farms in their early crawl cycle. Cross-reference with Wayback Machine snapshots — if the domain showed entirely different content 3 months ago (auto insurance → gambling → your niche), it's been recycled.
Check 5 — Outbound Link Ratio
Sites with 200+ outbound links on a single page are operating as link directories or link networks, not editorial resources. Check the linking page directly. If it's a wall of hyperlinks with no coherent topical thread, the link provides no authority transfer and may signal a pattern Google penalizes.
Section 2: Page-Level Red Flags
Check 6 — Indexation Status
Copy a sample of linking URLs and paste them into Google as site:example.com/page-path. If the page isn't indexed, Google isn't counting the link. If it's actively deindexed, the domain may be penalized. Deindexed linking pages at scale = a sign of a sitewide manual action on the referring domain.
Check 7 — Page Relevance
Open the actual linking page. Does the content have any topical overlap with your site? A personal injury law firm getting 40 links from recipe blogs with no food-law content is an obvious editorial mismatch. Relevance isn't required for every link, but mass irrelevance is a pattern.
Check 8 — Link Placement
Links buried in footers, sidebars, or injected into old WordPress posts (especially below the byline or in plugin license blocks) are programmatic, not editorial. Editorial links appear within the body copy of a page, in context that justifies their existence.
Check 9 — Surrounding Anchor Text and Co-Citation
What other links appear on the same page? If your link sits next to 15 other links pointing to payday lenders, gambling operators, and pharma sites, that neighborhood association is a liability regardless of the linking domain's metrics.
Check 10 — Link Velocity on the Referring Domain
Pull the referring domain's link history in Ahrefs (Referring Domains → chart over time). A site that went from 10 referring domains to 4,000 in 90 days is a link farm. Natural link profiles grow incrementally. Even legitimate viral content rarely produces that slope.
Section 3: Anchor Text Analysis
Check 11 — Exact-Match Anchor Ratio
Pull your site-wide anchor text distribution. Here's a rough benchmark for a healthy commercial site:
| Anchor Type | Healthy Range |
|---|---|
| Branded (your brand name) | 40–60% |
| Naked URL (yoursite.com) | 15–25% |
| Generic ("click here", "read more") | 10–20% |
| Exact-match keyword | Under 5% |
| Partial-match keyword | 5–15% |
If exact-match keyword anchors exceed 10–15% of your total link profile, you have either a manual optimization footprint or a negative SEO attack. Both require action.
Check 12 — Anchor Clustering
Even if your exact-match percentage looks fine in aggregate, check whether those anchors cluster on specific pages. If 80 out of 100 exact-match anchors point to a single money page, that page's link pattern looks unnatural regardless of the sitewide ratio.
Check 13 — Suspicious Anchor Themes
Search your anchor text export for pharmaceutical terms (Viagra, Cialis), gambling terms (casino, slots, poker), and adult content keywords. If any appear and you don't operate in those verticals, you have a negative SEO attack or a hacked-site injection. Disavow those links immediately — don't wait for the audit to complete.
Section 4: Link Pattern Analysis
Check 14 — Reciprocal Link Networks
Search your outbound links against your inbound link list. If site A links to you and you link back to site A, that's normal. If you find a triangle — A links to B, B links to C, C links back to A — and your site is one of those nodes, you're inside a link ring. These are explicitly against Google's link spam policies.
Check 15 — Paid Link Signals
Look for footprints: "sponsored," "advertorial," or "paid placement" labels without a rel="sponsored" attribute, or the inverse — links on sites with advertiser rate cards but missing any disclosure. Also check if the anchor text is unnaturally keyword-rich for what's supposed to be editorial content (example: a random news article about local events with a dofollow anchor "best SEO agency in Chicago" pointing to your competitor, then to you six months later).
Check 16 — Sitewide Links
A sitewide link (appearing in the header, footer, or sidebar of every page on a domain) generates hundreds of linking pages from one root domain. Google generally devalues or discounts sitewide links, especially from lower-quality domains. If you bought a sitewide footer link from a web design client's site, that's a disavow candidate.
Section 5: Your Own Link Profile Health
Check 17 — Lost Links
Ahrefs and Semrush both track "lost" backlinks — links that pointed to you and now 404 or redirect elsewhere. A lost link from a high-authority domain is link equity you can reclaim. Reach out to the linking site; if the page still exists and just changed URLs, request a correction. If the page is gone, check whether your content has moved and submit a redirect request.
Check 18 — Link Growth vs. Traffic Correlation
Open your GSC Performance report and overlay referring domain growth from your link tool on the same timeline. Links should correlate loosely with traffic growth. If your link count tripled in Q3 but organic traffic dropped, you added low-quality links that diluted rather than boosted authority. If link count flatlined while traffic grew, you have strong topical authority that isn't dependent on aggressive link building — protect it carefully.
What to Do With Your Findings
Keep: Links from indexed, topically relevant, naturally written pages with mixed anchor text. Even lower-DA links from real editorial sources are fine.
Monitor: Links from domains with borderline spam scores, recent registrations, or slightly off-topic content. Add them to a watchlist and reassess quarterly.
Disavow: Links featuring pharmaceutical/gambling/adult anchors unrelated to your niche, obvious link farm pages, sitewide paid links without sponsorship attributes, and any link that appeared during a suspicious velocity spike.
Build your disavow file at the domain level, not the page level, when you see multiple toxic pages from one root domain. Format: domain:spammydomain.com. Submit through Google Search Console's Disavow Tool — and keep a copy of the file for your own records before every update, since Google's UI doesn't version it.
Rerun this checklist every 90 days, or immediately after any traffic drop that isn't explained by algorithm updates or technical issues.
FAQ
How long does it take Google to process a disavow file? Google states there's no guaranteed timeline, but most practitioners report seeing ranking changes — positive or negative — within 4–8 weeks of submission. The disavow file affects how Google processes future crawls of those domains; it doesn't retroactively uncount links from the current index immediately.
Should I disavow every low-DA link? No. Domain Authority and similar metrics are third-party proxies, not Google signals. A low-DA link from a genuine editorial source (a local newspaper with poor technical SEO, a niche forum, a professor's course page) is often worth keeping. Disavow on the basis of spam signals, not authority scores alone.
Can I recover from a manual action caused by unnatural links? Yes, but it requires a documented cleanup. You need to demonstrate you contacted linking sites requesting removal (save all outreach emails), built a thorough disavow file for links you couldn't get removed, and submitted a reconsideration request through Search Console with a detailed explanation. Recovery typically takes one to three months after the reconsideration request is approved.
Takeaway
A backlink audit isn't a one-time cleanup — it's a recurring diagnostic. Run through these 18 checks, prioritize the most obviously toxic signals first (pharmaceutical anchors, link farms, deindexed pages), and build a disavow file grounded in evidence rather than anxiety. If you're producing content consistently, tools like FluxWriter can help you generate the kind of link-worthy material that earns editorial placements and gradually dilutes whatever low-quality links you're cleaning up.