RESEARCH · 7 SEPTEMBER 2026
One list, bought for a real campaign, checked page by page and then verified by hand fifty rows at a time. No sampling, no cleaning beforehand, and every row that failed stayed in the denominator. What follows is not an argument against comment links; 37 of these pages would have carried one. It is an argument about what a bought list contains and about the metric everyone sorts it by.
were waste: no comment form, no followed link, or no readable page. 818 had no comment form at all. The list was priced the same either way.
were among the 37 that worked. Not one. Sorting this list by Domain Rating puts every unusable row at the top of it.
The rows that did work were near-invisible domains: 34 of the 37 sat at DR 10 or below, and the strongest was DR 29. No filter in the export would have surfaced them.
This is the finding worth the whole run. The list was bought on domain metrics, and on domain metrics it looked respectable: 285 rows at DR 50 or above, a median of DR 18.
Not one of those 285 rows could carry a followed comment link. Of the 37 that could, 34 sat at DR 10 or below. The correlation between Domain Rating and usefulness on this list was not weak. It pointed the wrong way.
The reason is visible once you group the rows by host. The highest-rated rows are not strong sites. They are subdomains of strong platforms, and Ahrefs rates a hosted subdomain at its platform’s rating:
A DR floor applied to this list would have kept all 38 Blogspot comment pages and thrown away most of the 37 rows that actually worked. That is not a tuning problem. It is what happens when a domain-level number is used to answer a page-level question.
An open comment form is not a followed link. Four rows in five with an open form would give you nothing that passes equity.
Having found the rows that would carry a followed link, the obvious next question is what those pages are. The answer is uncomfortable: 33 of the 37 were unedited theme-demo blog posts: the sample articles that ship with an e-commerce template and never get deleted. Titles like “best-leather-bags”, “wool-jackets” and “greece-travel” repeat across dozens of unrelated shops.
30 of those 33 were on domains at DR 6 or below. On spam risk, 19 of the 37 came back Medium and 8 came back High, leaving 10 rows out of 1,001 that were followed, readable and low-risk at the same time.
Which is the honest summary of the list: about one row in a hundred was worth an outreach email, and you could not have found it by sorting.
The content-type breakdown explains the 81.7% better than any commentary. More than half the list was ordinary articles, but the next two categories were church audio players and forum threads, page types that have never had a comment form and never will.
Two more facts about the shape of the list. 163 rows, 16.3%, were unedited theme-demo posts. And only 118 rows, 11.8%, had a page topic we could identify at all. The rest were too thin, too generic or unreadable. A list where seven pages in eight have no discernible subject cannot be filtered for topical relevance, whatever the seller says.
The list was also far less diverse than its row count suggests: 467 hosts for 1,001 rows, with 69.9% of rows sitting on a host that appears more than once. One Polish forum contributed 30 rows on its own.
This is one list. 18.3% is this list’s open-form rate and 3.7% is its followed rate, on one day. Neither is an industry benchmark and neither is a prediction for yours. A list built from your own research usually does better; a cheap bulk list usually does worse.
Vertical mix is not controlled for, and it matters: 19.8% of this list was retail and e-commerce, and that is where nearly every followed row turned out to live. A list of publishers would behave differently.
Link type on a page with no existing comment links is reported Unknown, not assumed. 33 open rows sit in that state, so the true followed count is somewhere between 37 and 70. We publish the number we measured.
The first three are spreadsheet work you can finish before spending anything on fetching pages, and they run in order in checking a list you were sold. What a comment link is worth once you have found a real one is a separate argument, and the blog commenting guide makes it against these same 1,001 rows.
