Blog commenting for SEO, judged from 1,001 real pages

Blog commenting is the tactic everyone has an opinion about and almost nobody has numbers for. The two camps both write the same article: one says it is dead, one says it still works if you do it right, and neither shows you a list.

We opened one. A purchased commenting list, 1,001 rows, every page checked and then verified by hand. What follows is what was on those pages, and what it implies about spending your afternoon this way.

The short version

One purchased list, 999 distinct URLs across 467 hosts, opened and verified by hand on 7 September 2026. One list is one list. A list built from your own reading will do better; a cheaper bulk list will do worse.

So: about one row in a hundred was worth writing a comment on. That is the number both camps are arguing around without stating.

What the dead pages actually were

The interesting part is not the 3.7%. It is what the other 96.3% turned out to be, because it says more about the list than about the tactic.

818 rows had no comment form of any kind. Not closed. Never had one. Breaking the list down by what the page actually was:

Nearly a quarter of what we paid for was page types that have never had a comment form and never will. Church sermon players were a large block of it. Whoever assembled the list was not opening pages, and the price was the same as if they had.

The finding that should change how you sort

Of the 285 rows at Ahrefs DR 50 or above, not one was among the 37 that worked. 34 of the 37 sat at DR 10 or below and the median was DR 1.

The reason is that the high-scoring rows were hosted subdomains. All 38 rows at DR 95 were *.blogspot.com, inheriting a rating that belongs to Google. 76 rows on one Polish forum platform all carried DR 61. 48 church media pages all carried DR 68.

This matters more than the headline percentage, because it means the standard cleaning move makes the list worse. Applying a DR 30 floor to this list would have removed every single usable row. There is a longer piece on why DR inverts on lists like this.

What the 37 good rows were, honestly

Having found them, the obvious question is what they are. The answer is not flattering: 33 of the 37 were unedited theme-demo posts, the sample articles that ship with an e-commerce template and never get deleted. Titles repeat across dozens of unrelated shops. 30 of those 33 sat at DR 6 or below.

On spam risk, 19 of the 37 came back Medium and 8 came back High. Ten rows survived everything.

A followed link on a page no human will ever read is worth approximately what it sounds like. The link exists. The audience does not.

So is blog commenting dead

No, and the data does not say that. What the data says is narrower and more useful: a purchased commenting list is mostly filler, and the standard way of sorting it makes it worse. Those are two facts about lists and metrics, not about the tactic.

Comment links still do three things, and only one of them is about ranking:

• Referral traffic and visibility on a page with an actual audience. A thoughtful comment under a well-read post gets read. This works whether the link is followed or not, which is the part the followed-link argument keeps missing.
• Starting a relationship. A useful comment is often the cheapest first contact with someone whose site you want a real placement on later. It is outreach with a lower rejection cost.
• A followed link, occasionally. Rare, worth little on its own, and never the reason to bother.

The narrow case where it is worth your afternoon

1. You found the page by reading, not by buying. Every problem above comes from the list, not from commenting. A page you found because you actually read it does not have these failure modes.
2. The post is recent and the thread is alive. A comment on a 2011 post with the last reply in 2013 is a message to nobody.
3. The existing comments are people, not pharmacy links. If moderation has given up, your comment is sitting in a dump and so is your name.
4. You have something to say about that specific post. Not “great article, check out my site”. The version that works is indistinguishable from ordinary participation, because it is ordinary participation.
5. You treat the link as secondary. If the link is the point, the comment reads like it, and a human moderator will see that faster than any algorithm.

When to skip it entirely

• You bought the list. Check it before you write a word: see the vetting sequence.
• You are working at volume. The economics do not survive contact with a 1% yield.
• The campaign needs authority links. Comment links on DR 2 theme demos are not that, and no amount of them adds up to it.
• You are reporting followed comment links to a client as a deliverable. On this evidence you will not find enough of them to report.

If links are already built and nothing has moved, the cause is usually earlier in the chain than the tactic. Working through that in order — the target page first, then the links at page level — is more useful than adding another round of comments to the pile.

If you are going to work a list anyway

Check it before you write anything. The cost of qualifying a thousand rows by hand is roughly 33 hours at two minutes a page, and the arithmetic is why nobody does it and why the lists stay this bad.

The Backlink Prospect Checker opens each exact URL and reports whether a comment form is open, what the page already gives its commenters, the Ahrefs DR and how spammy the existing thread looks. It reads only: it never posts, never signs in, never answers a CAPTCHA. Where a page has an open form but no existing comment links, it reports Unknown rather than guessing from the platform.

The free allowance is 500 checks on first use, which at one credit a URL is five full scans of 100 URLs. That is enough to find out whether the list you were sold is better or worse than the one measured here.

The full run, with the method and the limits, is at 1,001 backlink prospects, opened by hand.