How to evaluate backlink quality before outreach (the list is the work)
Most link building advice wastes your time on the wrong end of the process. Everyone writes about the pitch: subject lines, personalization tokens, follow-up cadence. Nobody writes about the two hours you spend disqualifying sites before a single email goes out. That's where the outcome gets decided, and it's where I've made almost all of my expensive mistakes.
I'll tell you about one. Two years ago I sent 340 outreach emails in a month. I got 19 placements and thought I'd cracked it. Fourteen of those links disappeared within six months. Not broken — removed, or the whole site got deindexed. One of them had a domain rating of 71. Real talk: I had no idea what I was doing, I just sorted a spreadsheet by DR and started pitching.
So this is the pre-outreach qualification workflow. Not the email. The list.
Key Takeaways
- Domain rating is a starting filter, never a decision. It's trivially manipulable and says nothing about who actually reads the site.
- Check the traffic-to-authority ratio: a DR 60 site with 400 organic visits a month is a red flag, not a bargain.
- Look at how the site treats outgoing links. A homepage stuffed with commercial anchors tells you everything about how yours will be treated.
- Quality traffic means relevant country plus genuine search demand. Volume alone can be bought.
- Expect to reject 70 to 85 percent of a raw prospect list. If your rejection rate is lower, your filters are too loose.
- Run the disqualification pass before you write anything. It's cheaper to delete a row than to lose a relationship.
Why domain rating alone fails you
Domain rating measures one thing: the strength of the link graph pointing at a domain, on a logarithmic scale. It does not measure whether anyone visits, whether the audience matches yours, or whether the site will exist in eighteen months.
The ratio I check first
I keep a running benchmark in my head, roughly: a healthy site pulls somewhere between 300 and 800 organic visits per month for every ten points of domain rating. A DR 50 site should sit somewhere in the low thousands of monthly visits. When I see a DR 55 with 300 visits, I don't argue with it. I delete the row.
Same logic in reverse. A DR 25 site with 12,000 monthly visits from my target country is interesting, and almost every outreach guide I've read would have told me to skip it. And the worst part? Those low-DR, high-traffic sites are usually run by a single person who actually answers their email.
What authority metrics cannot show you
Two sites can share the exact same domain rating and be completely different businesses. One is a regional news outlet with a staff, an editorial calendar and a legal department. The other is a domain someone bought at auction, filled with programmatic pages, and wired to a link-selling service. The metric is identical. The link you'd get is not.
| Signal | Green light | Reject |
|---|---|---|
| Organic traffic vs. DR | Proportional or better | DR 50+, under 500 visits |
| Traffic geography | Matches your target market | Majority from markets you don't serve |
| Outgoing links on homepage | Few, mostly editorial | Dozens of commercial anchors |
| Content cadence | Irregular but real, dated | 400 pages published the same week |
| Branded search | People search the site's name | Nobody does |
| Link placement | Inside body copy | Footer, sidebar, "partners" block |
That table is the whole article compressed. Everything below is how you actually populate it.
The operational disqualification signals
This is the part missing from most checklists. Everyone tells you to check relevance. Nobody tells you what a link network looks like from the outside.
Indexation patterns
Run a site search on the domain. If the results look like a phone book — thousands of near-identical pages, tag archives with no content, author pages with no authors — you're looking at a content farm. Legitimate publishers have messy, irregular footprints. Farms have uniformity.
I also check whether the site's recent posts are indexed at all. Publishing 500 articles that never appear in search results is a strong hint that the site is built for sellers, not readers.
Outgoing link anchors
Open the homepage and count commercial anchors. I'm not talking about links to sources in articles. I mean the sidebar block labeled something like "our partners" or "useful sites". If half of them are exact-match commercial anchors, the site sells links, and yours will sit next to a payday loan site within a month.
One more tell: hover over the outbound links and look at the rel attribute. Sites that mark everything nofollow except the paid ones aren't subtle about it.
Network fingerprints
PBNs leak. They use the same hosting ranges, the same WordPress theme, the same footer text, the same Google Analytics property, and they all link to each other. If the site links out to four other domains with the same layout, you've found a network, not a publication.
- Identical theme across several domains in the same niche
- Footers with near-identical wording
- Contact page pointing to the same generic email address
- All domains registered within a few weeks of each other
None of these alone is fatal. Three together and you're done.
Checking traffic quality, not just volume
Volume is the easiest thing to fake. A site can buy a few thousand bot visits a month and look healthy in any dashboard. What you want to know is whether real people with buying intent are arriving.
Branded vs. non-branded search
If a site's traffic comes almost entirely from people searching its own name, that's fine — it means they have a community. But those visitors aren't discovering new content, so a link there mostly helps your brand, not your rankings. Ideally you want a healthy mix: some branded demand, a lot of long-tail discovery traffic.
Geography and language
I sell to a European audience. A site with 200,000 monthly visitors, 90 percent from a market I don't serve, is worth less to me than a 4,000-visit site in my own country. The math here is brutal and most outreach tools hide it behind a single "traffic" column. Filter by country before you filter by anything else.
Which brings up an obvious problem: most prospecting tools give you a country breakdown that's estimated, not measured. Treat it as a hint, not a fact. When a site matters to me, I open it and read three articles. Five minutes tells me more than any estimate.
A qualification workflow that survives a real list
Here's the order I use. It matters, because each step is cheaper than the next.
- Export the raw list. No filters yet. Mine usually start around 600 domains.
- Kill the obvious. Wrong country, no organic traffic, or a DR under 15 with no compensating traffic. This step alone removes about 60 percent.
- Check the traffic-to-authority ratio. Anything that looks like a bought domain gets cut. Another 10 to 15 percent.
- Manual pass on the survivors. Open the site. Read the last five posts. Count commercial anchors on the homepage. This is where the real judgment happens and where I spend the most time — roughly 40 minutes per 50 domains.
- Only now write anything. Pitch email, topic angle, personalization note. All of it last.
My rejection rate lands between 78 and 85 percent. When it drops below 70, I know I've gotten lazy, and the placements I get in those batches tend to disappear within a year. That correlation is the most useful thing I've learned about link building, and it took me two years and probably €3,000 worth of wasted effort to see it.
What about high-DR sites with bad traffic?
Occasionally you'll find a DR 80 site with genuinely poor organic traffic that's still worth pursuing — usually an old media brand or a university department whose authority is real but whose SEO is neglected. These are rare, and I'd rather spend my time on twenty mid-authority sites with real readers than gamble on one of these. In my experience, the correlation between "high DR, no traffic" and "link farm" is strong enough that I no longer chase exceptions.
How long should qualification take per prospect?
For the first pass, about 20 seconds per domain. For the second, two to three minutes. For the ones you actually pitch, five to ten minutes. If you're spending fifteen minutes on a site you're going to reject, you're doing the work in the wrong order.
The metrics will keep changing, new tools will keep promising to automate this, and every year someone will launch a "vetted link marketplace" that's just a PBN with better branding. The judgment doesn't automate. You still have to open the site and decide whether a human would ever read it.
What I'd genuinely like to know is whether anyone reading this has found a faster way to spot a network before the manual pass. I haven't. But I'd rather ask than keep deleting rows for another two years.
