How to Vet a Guest Post Site
Before You Pay.

A Traffic Figure Is an Estimate

The monthly visits figure in a pitch is not a count of anyone. A tool outside the site cannot see its analytics, so what Semrush and its competitors publish is assembled from rankings: the terms the tool has seen the domain rank for, the searches it expects each term to draw, and the share of those searches it expects each position to win. Across a broad set of rankings, that can land near the truth. Across a thin one, a single misjudged term can move the whole figure.

So the first question for a seller is where the number came from: which tool, and on what date. A figure with neither is a claim you have no way to check. A figure with both is one you can pull up yourself in a minute, and it is still only as good as the rankings underneath it, which is what the next two sections test.

Zero is where the modeling shows most plainly: 12 of the 84 listings here carry a Semrush estimate of exactly 0 monthly visits, and 11 of those 12 still rank for at least one keyword. That is usually no contradiction. A traffic model credits nothing to rankings it expects nobody to click, and the same arithmetic that finds no visits behind a handful of low rankings can find thousands behind one well-placed term.

The working rule is to treat every traffic figure — a seller's, a tool's, or the one printed on a listing here — as a statement about rankings, and to test it against them. The listings print Semrush's estimate beside the keyword count so the two are read together, and the guide sets out how to read that pairing.

What Keyword Count Exposes

Domain Authority and Domain Rating score a link profile. Neither counts a reader. Moz describes DA as a comparative prediction of ranking ability, and Ahrefs describes DR as a measure of backlink profile strength. Both are models of links, and links can be acquired in weeks.

Organic keyword count cannot be acquired that way. It is the number of distinct search terms a domain currently ranks for, and it is the one field that separates a working publication from a shell with a good link profile. A genuine publisher covering a subject ranks for thousands of terms, because it has hundreds of articles competing on hundreds of topics. A domain built to sell links ranks for a handful.

The distribution is worth seeing rather than being told about. Across the 84 publishers in this directory, the median listing ranks for 19 organic keywords, and 30 of them rank for ten or fewer. Those are not automatically bad placements, but the number states plainly what is being bought: a link on a domain that search engines currently surface for almost nothing.

The clearest case listed here carries Moz DA 60 against Ahrefs DR 0 — 60 points of disagreement between two link models looking at the same profile — while Semrush reports 4 monthly visits across 26 ranking keywords. The DA is the number a seller leads with. The visits figure is the number that says what a link there is worth to you.

Across the whole set, Moz DA runs about +8 points above Ahrefs DR on average. That spread is why a single authority score is never enough: two models are reading the same links and reaching materially different conclusions, so quoting one of them is quoting the kinder one.

The Ten-Second Arithmetic Check

Divide monthly organic traffic by the number of ranking keywords. The result is average visits per keyword, and it catches an inflated traffic claim faster than anything else available to a buyer.

Search demand is long-tailed. A publisher ranking for a few thousand terms takes most of its traffic from a few dozen of them and a trickle from the rest, so the average lands in single or low double digits. When the average comes out in the hundreds or thousands, the arithmetic is telling you the traffic is not coming from a spread of those keywords — it rests on a search volume the model has misjudged, or on one volatile term that will be gone next quarter.

A worked example from this directory: one listing reports 22,603 Semrush visits against 14 ranking keywords. That is roughly 1,615 visits per keyword, where the median across the listings here is 14.0. No ordinary search distribution produces that, on any subject, at any scale. The listing stays published with both figures on it, because they are what the tools report — and it prints the division beside them, so nobody has to read the traffic on its own.

Run the same division on any site you are quoted. It needs two numbers a seller should be able to produce in one screenshot, and the sum takes longer to type than to interpret.

Age, Archive and Cadence

Three facts sit on the domain itself rather than inside a tool, and between them they answer whether a publication existed before it started selling placements.

• Registration dateRead from the registry record, not from a copyright line in the footer. A domain registered eighteen months ago has no track record to judge, whatever its authority score says.
• Indexed pagesA site: query returns the count in one search. It is the size of the archive a search engine will admit to holding, as opposed to the size the navigation implies.
• Update cadenceOpen the archive or a category page and read the dates. A publication updating steadily will still exist in a year; one that stopped in 2023 and resumed only for paid posts will not.

Registrations in this directory run from 1994 (Rox.Com) to 2026 (DallasNewsSarah.Com) — a spread wide enough that age on its own settles very little without the other two readings beside it.

Indexed pages are the more revealing of the three. The median across these listings is 50 indexed pages. Set a figure like that against a claim of tens of thousands of monthly visitors and the two statements are hard to hold at once: a few dozen pages cannot carry that much search traffic unless every one of them ranks nationally for a competitive term.

None of the three disqualifies a site by itself. A young domain with a real archive and honest traffic is a perfectly reasonable placement. A three-year-old domain with forty indexed pages and nothing published between paid articles is a different purchase entirely, and the difference is visible in about a minute.

Reading robots.txt Yourself

Append /robots.txt to the domain and read it. It is a plain text file, it loads instantly, and it is one of the few things about a site that nobody has prepared for a buyer to look at. Google's documentation covers what the directives mean; for vetting, the line that earns its keep is Sitemap:.

A normal publisher declares one sitemap, occasionally two, and both are on its own domain. Then open the sitemap itself. The sitemaps protocol is a simple XML list of URLs with last-modified dates, and those dates tell you when the site was genuinely updated rather than when its theme last rebuilt.

What that catches: a robots.txt that declares dozens of sitemaps, most of them on entirely different domains in unrelated subjects. That is the signature of one shared publishing install writing every site in a network into every site's robots.txt. It is a configuration mistake rather than a rule violation, and it does not make a publication worthless. It does tell you the title is one of many operated together, which is worth knowing before you place a link there and describe it to yourself as independent.

Footprint is the reason this matters beyond tidiness. Google's spam policies describe link schemes as links intended primarily to manipulate rankings, and name buying or selling links that pass ranking credit among them. Thirty domains sharing one install, one template and one link inventory is precisely the pattern a footprint analysis exists to find.

The Part No Tool Measures

Once the numbers survive, what remains is editorial, and it is answered by reading rather than measuring. Open three articles that were not sponsored. Is there an argument in them, or paragraphs assembled around a keyword? Is there a named author with anything behind the name? Does the site publish on its own subject when nobody is paying it to?

Relevance beats authority here more often than buyers expect. A placement on a modest site your readers actually use does more than one on a stronger domain in an unrelated field — which is why the listings here are grouped by subject. Technology titles index and date quickly, business titles hold their relevance far longer, and that trade-off is worth choosing on purpose rather than inheriting from whatever a seller had in stock.

The final check is on the seller rather than the site: ask what they verify, and when they last checked it. We publish what gets measured and what gets left off a listing for exactly that reason, and no rate is printed anywhere on this site. Prices move, a printed price is wrong the day after a publisher changes it, and every listing therefore says “Email for details” instead.

The Checklist

Applicable to any site, listed here or not. The checks are ordered so the cheapest come first — stop at the first hard failure.

1. Ask where the traffic figure came from. Which tool, and on what date. A number nobody will source is a claim, not a reading; a sourced one is still a model's estimate, and the next two checks are how you test it.
2. Read the organic keyword count. A few hundred terms means the domain ranks for very little, whatever its DA suggests. Under ten means it ranks for almost nothing.
3. Divide traffic by keywords. Averages in the hundreds or thousands per keyword do not survive contact with how search demand actually distributes.
4. Compare DA against DR. A wide disagreement means two link models cannot agree about one profile. Not fatal on its own, but ask what is driving it.
5. Check the registration date in the registry record, not the footer, and set it against the size of the archive it is supposed to have produced.
6. Run a site: query. Indexed pages should be plausible for the traffic claimed. A few dozen pages and five-figure traffic is a contradiction, not a bargain.
7. Open /robots.txt and the sitemap. Confirm the sitemaps belong to the domain, then read the dates for real update activity.
8. Read three unsponsored articles. If you would not want your brand beside them, the metrics stop mattering.
9. Ask the price last. A link on a domain nobody reads is expensive at any figure.

On 74 of the 84 listings here, the first five checks are already done and dated 29 Sep 2026. The last four are yours: three take a browser tab, and the price takes an email. The full table filters on subject, authority and domain age, and sorts on traffic or keyword count. Every figure stays on the page, including the unflattering ones.