What High-Authority Sites Will Actually Link To: Placement Data From 1,786 Domains

Every link building plan starts with a target list: the pages on your site that need authority. In practice, the page you want linked and the page a high-authority site is willing to link to are often different pages. We grouped the linked pages into four kinds: blog posts and other editorial content, landing pages, homepages, and product or service pages, and recorded which of them each domain linked to.

Blog posts are close to universal: 97.7% of the 1,384 domains link to them. From there it falls away fast. About half link to a landing page (50.6%) or a homepage (49%), and 44% link to a product or service page. Put another way, more than half of high-authority sites will not link to the page that makes you money.

The gap is not random. An editor adding a link inside an article is protecting the reader and the site's own standing. A link to a useful blog post reads as a citation. A link to a pricing page reads as an advertisement, and editors at authoritative sites treat it as one, however relevant the product is.

How many sites refuse commercial pages outright?

We then looked at the domains that link to blog posts and to nothing else.

43% of the 1,384 domains link to blog posts only. Every homepage, landing page and product page is off the table there, however good the fit. The remaining 57% link to at least one other kind of page, and 44% of all domains in the set link to a product or service page.

For a campaign that needs authority on commercial pages, that cuts the usable site list by more than half before any other filter is applied. It also means the blog-posts-only sites are not a rounding error you can work around. They are close to half the set, and in this data they include many of the most authoritative domains.

Being allowed to link to a page type is the first filter. The second is whether the editor is willing to link to your site at all. Across the 1,786 domains we recorded three checks editors apply to the linked site: a minimum Domain Rating, a cap on the site's spam score in Ahrefs or Moz, and a minimum level of organic traffic.

32.5% of domains hold the linked site to a minimum DR. 26.4% cap its spam score, and 15.6% require a minimum level of organic traffic. The checks overlap, but any one of them is enough, so a site below DR 50 with thin traffic is filtered out at a third or more of these domains before the fit or quality of the content is ever considered.

These are not arbitrary hurdles. An editor who links out to a weak or spammy site puts their own site's reputation at risk, so the default at a high-authority publication is to cite the most authoritative source available. When your page is not that source, the editor swaps it for one that is: a government page, an industry body, a well known brand, or a competitor with a stronger profile. The link survives. It just no longer points at you.

The stronger the site, the higher the bar

We split the 1,786 domains by their own Domain Rating to see whether the most authoritative sites are also the most selective.

They are. 42% of domains at DR 70 and above set a DR floor for the sites they link to, against 24.6% of domains below DR 70. The links that carry the most authority sit behind the toughest checks, which is what you would expect from editors whose own standing depends on who they cite.

This is the part of the data that explains why so many campaigns aimed at top-tier sites stall. The sites are reachable and the content is relevant. The linked site is the problem.

Where the floor sits

For the 581 domains with a DR floor, we recorded where it sits.

The floor clusters at DR 50 to 59, where 44.4% of these domains set it. 18.9% go further and require DR 60 or higher, so 63.3% of domains with a floor set it at DR 50 or above. Only 18.4% set the floor below DR 40.

Page type and authority decide whether a link is possible. Editorial rules decide what it looks like once it is in: where in the article it sits, what the anchor text says, and how much the surrounding text can change. These rules are rarely published. They show up in how editors handle the link, so we recorded them from the placements themselves and grouped them by theme.

Fit comes first. 33.2% of the recorded rules require the link to match the topic of the article it sits in. Placement is next: 22.6% keep links out of the introduction and conclusion, the two parts of an article readers and editors look at most. 21.2% set rules on anchor text, usually that it reads as part of the sentence rather than as a keyword. 18.4% restate the blog-posts-only rule in their own words, 17.4% refuse links to their direct competitors, 11.3% repeat a DR, DA or traffic requirement for the linked site, and 9.4% limit how much of the article can be edited around the link.

Two of these deserve a closer look. The competitor rule works against you when your site competes with the publication itself for readers or rankings: a media site or large blog in your own category is often closed to you regardless of quality, so the most topically relevant sites can be the hardest to earn a link from. The anchor rule means an exact-match keyword anchor is the first thing an editor rewrites. The anchor that gets through is the one that reads like the rest of the sentence.

Can the text around the link be changed?

Finally, we looked at whether editors allow small changes to the sentence around the link. This decides whether a link into an existing article is possible at all, or whether new content is the only route in.

87.4% of the 1,784 domains with a recorded answer allow small text edits around the link. The remaining 12.6% take a link only where it fits the existing sentence, which rules out most link inserts on those sites and leaves guest content as the only option.

That 87.4% is the most encouraging number in the study. It means a well written insert into a relevant article that already ranks is possible at most high-authority sites, provided the page type, the authority check and the fit all hold.

Reading the first three chapters in order gives you a funnel. Start with the domains that link to the kind of page you need: for a product or service page, that is 44% of the 1,384 with a clear read. Then remove the domains whose authority checks your site cannot pass: at DR 70 and above, 42% hold the linked site to a DR floor, and most floors sit at DR 50 or higher. Then remove the domains whose editorial rules do not fit the link you have in mind: the 17.4% that refuse links to their own competitors, and the 12.6% that allow no edits around the link.

What is left is the real target list. It is smaller than the raw list, but every site on it can carry the link you need to the page you need. Two common starting points look like this.

Data. Link placements across 1,786 unique domains, one record per domain, as the data stood in October 2026. For each domain we recorded what the site linked to and what it refused, the checks its editors applied to the linked site, and the editorial rules that shaped the link. The median domain in the set has a DR of 67, so the set describes established, high-authority sites rather than a random sample of the web.

Structured fields. Page types, minimum DR and traffic, spam score caps and text edits come from fixed fields recorded for each domain. Page-type shares are over the 1,384 domains with a clear read on that field. The blog-posts-only share counts domains that link to blog posts and to no other kind of page. A field for refused industries was left out because most records carry the same default list.

Written rules. For 576 domains, editorial rules were recorded in the editor's own words. We tagged them with keyword rules, so an unusually worded rule can be missed. Rules that fewer than 5% of those domains apply are left off the chart. Shares in that chart are of the 576, not of all 1,786.

DR comparison. Domains were split at DR 70 using their Ahrefs Domain Rating at the time of the record.

What we left out. No domain, editor, contact or client is named anywhere in this study or its data. The dataset is published as aggregate shares only.