Link Building Metrics: What to Look At Before You Buy

Domain Rating measures one thing: the strength of the links pointing at a domain, on a logarithmic scale where moving from 20 to 30 is far easier than moving from 70 to 80. It says nothing about traffic, relevance or editorial standards.

It is a reasonable first filter because it is hard to fake slowly — but it is not hard to fake quickly. Buying a few hundred links at a site will move DR within weeks, which is why a domain with DR 60 and no organic traffic is a common and deliberate shape.

Use it to rule sites out at the bottom rather than to rank them at the top. A DR of 5 usually means nobody links to the site at all; a DR of 65 means somebody, somewhere, links to it a lot, and the next question is who.

Domain Authority is Moz's version of the same idea, computed from a different index. The two correlate but do not match, and a site can be DR 40 / DA 25 or the other way round without either being wrong.

Worth knowing mainly because sellers quote whichever of the two flatters the site. If a listing leads with DA and you cannot find its DR, that is usually the reason.

This is the single most useful number on a listing, because it is the hardest of them to manufacture cheaply. Traffic means the site ranks for things people search for, which means search engines already trust it for something.

It is an estimate: the tool sees the keywords a site ranks for, guesses the click-through rate, and multiplies. Real traffic is usually higher than the estimate for large sites and lower for small ones. Treat the figure as an order of magnitude rather than a measurement — the difference between 200 and 20,000 is what matters, not the difference between 12,000 and 14,000.

The shape of the curve says more than the number at the end of it. A site at 20,000 visits that was at 60,000 eighteen months ago has been hit by something — an algorithm update, a competitor, a redesign — and a link from it is worth less than the current figure suggests, because the figure is still falling.

A sharp vertical spike is the other pattern to be wary of. Real editorial sites grow in steps; a cliff-edge rise usually means expired-domain rebuilds, aggressive link buying, or a single piece of luck that is already over.

Flat and boring over two years is an excellent sign, and almost nobody advertises it.

A site with 50,000 monthly visitors is not useful to a UK business if 45,000 of them are in Indonesia. The link still passes signals, but the audience is not one that can become a customer, and relevance to your market is part of what search engines are measuring.

This is where a lot of cheap inventory falls down. Sites with large but geographically irrelevant audiences are common and are priced as though traffic is traffic.

If you sell in one country, weigh a site with 3,000 visitors in that country above one with 40,000 spread across places you do not operate.

Relevance is the metric with the least precise number attached and the most weight behind it. A link from a site that genuinely writes about your industry, in a paragraph that makes sense without the link, is the thing everything else is a proxy for.

Majestic's Topical Trust Flow is the closest thing to a measurement: it categorises the sites linking in, so a domain whose inbound links are mostly from finance sites is probably a finance site. Press Parrot shows the top categories for each listing from that data.

The cruder test is better: read three recent articles. If you cannot tell what the site is about, neither can a search engine, and that is the actual problem with general-purpose "we cover everything" publishers.

More informative than a raw backlink count, because a thousand links from one site is one relationship and ten links from ten sites is ten. Referring domains is the count that resists inflation.

What to look for is the ratio between referring domains and total backlinks. A site with 400 referring domains and 900 backlinks looks like a site people cite. A site with 400 referring domains and 90,000 backlinks has something sitewide — a footer link, a widget, a template — and the profile is less organic than the domain count suggests.

A page carrying twelve outbound commercial links is diluting every one of them, and the pattern is obvious to anybody looking — including the systems designed to look.

The number to check is not on the domain but on the articles. Open two or three recent posts and count the external links that point at businesses. One or two in a two-thousand-word piece is normal editorial behaviour. Seven is a page that exists to sell links.

This is the check that most reliably separates a publisher from a link farm, and it takes about a minute.

A link inside the body of an article, in a sentence that would still make sense without it, is worth considerably more than the same link in an author bio, a footer, or a paragraph obviously bolted on at the end.

It is also the thing you have most control over. When you supply or commission the article, you decide where the link goes and what the sentence around it says. That is the main practical advantage a guest post has over a link inserted into an existing piece.

Ask what the link policy is before ordering: whether links are dofollow, whether a sponsored tag is applied, and whether the placement is in-content. Press Parrot listings state this upfront.

The judgement no metric replaces. Open the site. Does it publish anything other than sponsored posts? Is there a named author, a contact page, a history? Do the articles read like they were written for a reader or assembled for a crawler?

Sites that fail this are usually easy to spot within thirty seconds, and they are frequently the ones with the most impressive numbers, because the numbers are what they were built to have.

A useful proxy: look at what the site published in the month before it started selling links. If there is nothing there, the site is not a publisher.