Key takeaways
A bad placement is charged twice: once by the seller, once by the time spent undoing it.
Six of the seven checks are hard passes. Only editorial substance is a judgement call.
Two failures account for most rejections: no real readership, and a page already saturated with paid links.
Every check runs before a price is discussed, never after one has been agreed.
Why a checklist at all
A link is cheap to buy and expensive to own. The invoice is the small part of the cost: a placement on the wrong site buys a signal that has to be explained later, in an audit, in a disavow file, or to whoever inherits the domain.
The arithmetic is worth doing plainly. A placement bought badly is paid for once, delivers nothing, and then costs a second time when somebody has to find it, judge it and decide whether removing it is riskier than leaving it. In our experience that second cost is the larger one, because it lands months later and lands on people who were not in the room when the decision was made.
So the question is never “is this link worth €400”. It is “would we be comfortable with a reviewer reading this page next to the client’s name”. Everything below is an attempt to answer that before money moves.
Six of the checks are binary. One is not, and we say so where it comes up.
The seven checks
Each of these runs on the specific page the link would sit on, not on the domain in the abstract. A strong domain publishes weak pages, and a weak-looking domain sometimes publishes the one page in the market that the right readers actually open.
01 — Real traffic
We look at organic visits to the host page itself. Not a domain-level score, not an aggregate for the whole site, and not the seller’s screenshot.
Domain scores are a modelled number, and anything modelled can be pushed. A page that nobody reaches does not carry readers, and a link from it is an entry in a database rather than a recommendation. Hard pass.
02 — Topical fit
Sport, finance, entertainment or regional news — somewhere a piece on this subject is plausible to the people who already read the site.
Fit is not a keyword match. It is whether the article would look strange to a regular reader. If it would, the placement is doing the opposite of what a backlink is supposed to signal. Hard pass.
03 — Outbound pattern
How many paid links the site already carries, and who they point to.
This is the check sellers like least, because it is the one that exposes inventory. A page carrying eight commercial links to eight unrelated industries is not a publication; it is a shelf. Saturated sites are dropped whatever their other numbers look like. Hard pass.
04 — Indexation
The page and the section it sits in have to be indexed and reachable from the site’s own navigation.
A placement in a directory that nothing links to, or that carries a noindex header, is a link nobody will ever follow — including the crawler it was bought for. We check the live header, not the seller’s claim. Hard pass.
05 — Editorial substance
A masthead, named writers, and a publishing history that predates the link market.
This is the judgement call. Plenty of good regional titles have thin mastheads and uneven archives; plenty of link shops have invented a convincing one. What we are reading for is whether the site existed to publish something before it existed to sell links. Where it is genuinely unclear, we say so on the shortlist and let the client decide rather than quietly including it. Judgement.
06 — Licence context
Whether the jurisdiction and the site permit links from this category at all.
In regulated markets a placement can breach advertising rules — and the risk often sits with the publisher, which means an editor who later realises it will remove the link and the money with it. In grey markets the exposure is different but not absent. We check this before pitching, not after a piece is written. Hard pass.
07 — Footprint
Hosting, ownership and template overlaps with other sellers in the same pool.
Individually, five placements can each look reasonable. Together, on the same IP range, the same theme and the same contact address, they form a footprint — a pattern that identifies the whole group at once. Anything that clusters is refused, including sites that would otherwise pass all six checks above. Hard pass.
What gets rejected most
Two failures account for the bulk of it.
No real readership. The single most common rejection. A domain score in the sixties, a clean design, and almost no organic traffic to the pages that would carry the link. This is the normal state of the marketplace inventory, and it is why marketplace pricing looks attractive — the price reflects what the link is worth, which is very little.
Saturation. The second most common. The site is real, the readers are real, and the page has already sold the same slot to six other advertisers. Whatever authority passes through is divided by everything else leaving the page, and the commercial pattern is obvious to anyone looking.
A distant third is licence context, and it is the one that surprises clients most often. A site can be perfect on every other measure and still be a bad idea in a specific market in a specific quarter.
What survives all seven is a smaller list than most clients expect, and considerably more expensive per placement. That is the trade: fewer links, each of which can be defended on its own. The full protocol, and what it costs to run, is on the link building page — and the reasoning behind refusing whole categories of inventory is in the piece on rented links.
