To evaluate backlink quality, you check whether a real site with real readers placed the link on a relevant page because it helps those readers. In practice that means eight checks: topical relevance, real traffic, editorial placement, outbound link patterns, indexation, anchor text, rel attributes and site history. A good backlink passes most of them. A link that fails the link-selling or indexation checks isn’t worth having, whatever its authority score says.
When I vet a site, whether it’s a prospect for outreach or a link that already points at a client’s page, I run the checks in exactly this order, because the cheap pass or fail questions knock out most bad candidates before I spend any time on the slower ones. Below you’ll find the three stop checks I run first, the eight scored checks, a scorecard you can copy, and what to do with each result.
Why a Single Metric Gives You the Wrong Answer
Most people judge a link by one number, usually DR or DA. I get it. It’s quick, it’s in every tool, and a high number feels safe.
The trouble is that those scores measure the linking site’s backlink profile as one vendor’s crawler sees it. They say nothing about whether this particular page is relevant to you, whether anyone reads it, or whether the site sells links to anyone who pays. I’ve explained what those scores are and why Google doesn’t use them in my post on topical authority vs domain authority, so I won’t repeat it here.
A site can score high on authority and still be a link farm with a nice history. A small blog with a modest score can send you buyers every week. The checks below are built to tell those two apart.
What Makes a Backlink Good in Google’s Eyes?
Google hasn’t published a quality checklist for links, so I work backwards from what it calls spam. Its spam policies define link spam as “the practice of creating links to or from a site primarily for the purpose of manipulating search rankings.” The examples include paid links that pass ranking credit, “Low-quality directory or bookmark site links”, and “Widely distributed links in the footers or templates of various sites”.
Flip that around and you get my working definition. A good backlink exists because someone decided your page was worth pointing to, on a page their own readers use. Everything in this scorecard is a way of testing that one idea from a different angle.
Start With Three Stop Checks
Before scoring anything, I run three pass or fail checks. If a site fails any of them, I stop there. No score can rescue it.
- Does the site openly sell links? Look for a “write for us” page with a price list, “sponsored post” packages, or a contact page promising dofollow links. Google names “Exchanging money for links, or posts that contain links” as link spam.
- Is the site a content farm or a hijacked domain? Scroll the last twenty posts. If a plumbing blog is publishing casino, CBD and payday loan articles, it’s selling space.
- Can Google see the linking page at all? If the page is noindexed, blocked or missing from search, the link has little to offer. The indexation check below covers how to test this.
If a site passes all three, it goes to scoring. If you’re vetting links that already exist, a fail here usually means “ignore it”, not “panic”. I’ll get to that later.
How to Evaluate Backlink Quality in 8 Checks
Score each check 0, 1 or 2 points. Zero means it fails, 1 means mixed, 2 means clearly good. The maximum is 16 points.
1. Topical Relevance
Ask whether the linking page covers a subject your customers care about. A kitchen equipment review linking to your restaurant supply store scores 2. A general “business tips” page scores 1. A page about crypto trading linking to your dental clinic scores 0.
Check the page, not just the domain. Big general sites can have highly relevant sections, and niche sites sometimes publish off-topic posts.
2. Real Traffic and Real Readers
Honestly, this is the check I weigh most. In my experience, a site that real people visit for its own topic almost never turns out to be a link seller, while a site with a high score and no genuine readers usually does. Look at the site’s estimated organic traffic trend in your tool of choice, then look at whether it ranks for terms in its own topic. A steady or growing chart built on relevant keywords scores 2.
Remember that third-party traffic numbers are estimates, not analytics data. The shape of the trend often tells you more than the size. A chart that fell off a cliff after a Google update is a warning, and so is traffic that comes almost entirely from unrelated keywords.
3. Editorial Placement
Where does the link sit? A link inside the body text, surrounded by a sentence that explains why it’s there, scores 2. A link in an author bio or a resource list scores 1.
Footer links, sidebar blogrolls and sitewide template links score 0. Google’s policy lists widely distributed footer and template links as link spam, and they rarely carry much weight even when they’re innocent.
4. Outbound Link Patterns
Open three or four recent articles and look at who else they link to. Natural sites link to sources, tools and references that fit the topic. Link sellers show a pattern: every post has two or three exact-match commercial anchors pointing at unrelated businesses.
This matters more since the December 2022 link spam update. Google said SpamBrain “can now detect both sites buying links, and sites used for the purpose of passing outgoing links”, and that it would “neutralize the impact of unnatural links” (Google Search Central). A link from a site Google has flagged as a link seller may simply count for nothing. Clean outbound patterns score 2, a few questionable ones score 1, and obvious selling scores 0.
5. Indexation
Search Google for site: plus the exact linking URL. If it shows up, good. If it doesn’t, check the page source for a noindex robots meta tag and check whether its canonical tag points somewhere else.
Careful here. Google’s own documentation says the site: operator isn’t guaranteed to show every indexed URL. So a missing result is a reason to dig, not proof. Indexed and self-canonical scores 2, unclear scores 1, noindexed or canonicalized away scores 0.
6. Anchor Text Naturalness
Look at the words used to link to you. Your brand name, the page title, a plain URL or a descriptive phrase all look natural. Google’s link guidance says good anchor text is “descriptive, reasonably concise, and relevant”.
Exact-match money keywords are the problem, especially on placed links. Google’s spam policy calls out “links with optimized anchor text in articles, guest posts, or press releases distributed on other sites.” Natural anchors score 2, partial-match ones score 1, and a keyword-stuffed anchor scores 0.
7. Rel Attributes
Inspect the link. A plain link can pass ranking credit. A link with rel="nofollow", rel="sponsored" or rel="ugc" is treated by Google as a hint, so it may pass less or nothing.
Here’s the part many checklists get wrong. A sponsored link isn’t a red flag. (The broader dofollow vs nofollow question deserves its own post, so I’ll keep this to scoring.) It’s an honest one. Google says paid links are fine “as long as they are qualified with a rel=”nofollow” or rel=”sponsored” attribute value”, and that’s the rule I give anyone who asks about paid placements. For ranking value, a followed editorial link scores 2 and a qualified link scores 1. A link you know was paid for but left unqualified scores 0, because it breaks Google’s policy.
8. Site History
Check the domain in the Wayback Machine and skim what it looked like over the last few years. A site that has covered the same topic for years scores 2. A site that changed hands and topic recently scores 1.
A domain that was a dentist’s site until 12 months ago and is now a “general news” blog full of guest posts scores 0. That’s the classic expired-domain pattern, and I’d skip it every time.
The Backlink Quality Scorecard (Copy and Use)
Here’s the scorecard I use. Copy the table into your own sheet or notes, one row per link.
| Check | 0 Points | 1 Point | 2 Points |
|---|---|---|---|
| Stop checks | Any fail: stop, don’t score | n/a | All three pass |
| Relevance | Unrelated topic | Loosely related | Same topic as your page |
| Real traffic | Collapsed or unrelated traffic | Small or flat | Steady, topic-relevant traffic |
| Placement | Footer, sidebar, sitewide | Bio or resource list | In body text with context |
| Outbound links | Obvious link selling | A few questionable links | Natural sources only |
| Indexation | Noindexed or canonicalized away | Unclear | Indexed and self-canonical |
| Anchor text | Keyword-stuffed | Partial match | Brand, URL or descriptive |
| Rel attribute | Paid and unqualified | Nofollow, sponsored or ugc | Followed and editorial |
| Site history | Recent topic switch or expired domain | Some changes | Consistent for years |
And a plain-text version for anyone who prefers notes:
BACKLINK SCORECARD
Linking URL:
Target URL on my site:
Date checked:
STOP CHECKS (any NO = stop)
[ ] No link-selling signals (no price list, no "dofollow" offers)
[ ] Not a content farm / off-topic casino, loans, CBD posts
[ ] Linking page can be indexed (no noindex, canonical OK)
SCORES (0 / 1 / 2)
Relevance ........... _
Real traffic ........ _
Placement ........... _
Outbound links ...... _
Indexation .......... _
Anchor text ......... _
Rel attribute ....... _
Site history ........ _
TOTAL (max 16) ...... _
13-16 = strong 9-12 = acceptable 0-8 = skip
Notes:
The thresholds are my own rule of thumb, not an industry standard. I’ve found that anything scoring 13 points or more is a link I’d be happy to earn, and anything at 8 points or lower isn’t worth the outreach time.
What Can Third-Party Metrics Tell You, and What Can’t They?
Tools like Ahrefs, Semrush and Moz are genuinely useful here. I use them for the traffic trend, the outbound link view, anchor text reports and referring domain counts. If you’re not sure how those counts differ, my post on backlinks vs referring domains explains it.
What they can’t tell you is intent. No score knows whether the site owner took money for a link or whether a human ever reads the page. Treat authority scores as a rough filter for very weak sites, then let the eight checks decide.
Spam Score is a good example of a metric people misread. Moz describes it as the percentage of sites with similar features that it has found to be penalized or banned by Google, based on 27 features its model identified. Moz itself says a high score “doesn’t mean this site is necessarily spammy”, and that it’s a sign to investigate further. That’s exactly how I use it: a prompt to look closer, never a verdict, and I treat the “toxic” labels from other tools the same way in my post on whether toxic backlinks are real.
What Should You Do With a Low-Scoring Link?
The answer depends on whether the link is a prospect or already points at you.
| If the Link Is | And It Scored | Then |
|---|---|---|
| A prospect for outreach | 13–16 | Pursue it and pitch a genuinely useful page |
| A prospect for outreach | 9–12 | Pursue only if relevance scored 2 |
| A prospect for outreach | 0–8 or failed a stop check | Skip it |
| Already pointing at you, you didn’t build it | Any score | Usually leave it alone |
| Already pointing at you, you or a vendor built it | Failed a stop check | Ask for removal; consider disavow only if you have a manual action |
For links you didn’t build, my default is to do nothing. Google says of its disavow tool that “most sites will not need to use this tool”, because it can usually decide which links to trust on its own (Google Search Console Help). Links you paid for or built in bulk are a different story, and that’s where cleanup becomes worth the effort, following the removal-first steps in my guide to disavowing backlinks.
Where This Checklist Fits in Your Link Work
This checklist is for vetting individual links. It doesn’t tell you how many links to build or which tactic to use. If you’re deciding between earning coverage and doing outreach, I compare the two in digital PR vs link building.
For new businesses, the easiest links to pass every check are basic ones: real brand profiles, business listings and niche directories your customers use. That is all my foundation link building service covers, and it stays away from the link types this scorecard flags.
When to Bring In a Second Opinion
If Search Console shows a manual action for unnatural links, stop scoring and start reading Google’s notice carefully. That situation calls for a proper link audit and a reconsideration request, and a specialist who has handled one before will save you weeks.
The same goes if your rankings dropped right after a spam update and you know past vendors built links in bulk. A free SEO audit is a sensible first step before you touch anything.
Frequently Asked Questions
How Can I Check Backlink Quality for Free?
Most of the scorecard needs no paid tool. You can check relevance, placement, outbound links, anchor text and rel attributes by reading the page and its source. Use a site: search for indexation and the Wayback Machine for history. Only the traffic trend really needs a third-party tool, and most offer a limited free version.
How Long Does It Take to Evaluate One Backlink?
Once you’re used to the scorecard, a quick pass takes a few minutes per site. The stop checks go fastest, since a price list or a feed of casino posts is obvious on sight. Site history and outbound link patterns take the longest, so I leave them until a site has passed everything else.
What DR or DA Should a Good Backlink Have?
There’s no safe cutoff. Those scores are vendor estimates built from each tool’s crawl, and Google doesn’t use them. I use them only to skip obviously weak sites quickly, then judge the link on relevance, real traffic and placement.
How Many Checks Must a Backlink Pass?
It must pass all three stop checks. After that, I aim for 13 points or more out of 16 before chasing a link. A link between 9 and 12 is fine when it’s highly relevant to your page.
Last updated: September 2026 by Mizanur Rahman
