Three features that decide whether a link buy survives an audit
An inventory you can filter on thirteen fields, ownership proven before a listing exists, and a record that follows every order to its end. What each one does, and what none of them promises.
Every marketplace claims a feature list. Ours is short on purpose. Three things do the work that agencies, publishers and finance teams actually ask about, and everything else in the product exists to serve them: an inventory you can filter, proof that a publisher owns the site they are selling, and a record of what you bought that still makes sense six months later.
One: an inventory you can filter on thirteen fields
The usual buying experience is a spreadsheet in an email attachment. It has a domain, a Domain Rating, a price, and a promise. You cannot sort it against another broker’s sheet, you cannot tell when the numbers were taken, and you cannot ask it a question.
The marketplace here carries thirteen fields per listing, and every one of them is filterable. Price, placement kind, Domain Rating from Ahrefs, Domain Authority from Moz, organic traffic, referring domains, organic keywords, country, language, niche, turnaround, links allowed and the sensitive-niche policy. The metric fields refresh daily rather than being pasted in once at listing time, so the number you filter on is the number a buyer sees today.
What that changes in practice is the order of decisions. You set a traffic floor first, because a site with no readers is not worth a link at any Domain Rating. Then you set the authority floor, then country and language, then the price ceiling. Four moves and 1,284 sites become a shortlist you can read in a minute. Save that combination as a preset and it becomes the buying brief: the same floors, the same excluded niches, re-runnable next month without re-arguing it.
Two smaller details matter more than they sound. The column picker is per workspace, so a media buyer can work in price, Domain Rating and turnaround while an SEO lead keeps referring domains and keyword counts on screen, from the same inventory. And listing domains stay masked until an order is approved, which stops the catalogue being scraped and the publishers being approached directly around the platform.
Two: ownership proven before a listing exists
A price and a metric tell you what a site is worth. They tell you nothing about whether the person selling it can publish on it. That gap is where most bad link buys begin, and it is the reason a submitted website here does not reach buyers on submission.
A publisher proves control of the domain in one of three ways: a DNS TXT record on the zone, a meta tag in the document head, or a file uploaded to the server root. Each is checked against the live domain, and the check is rate limited so it cannot be brute forced. Passing it means one narrow, useful thing: this seller controls this domain.
Then a human looks at the site. An operator reviews the submission and either approves it or rejects it with a reason, and only verified and admin-approved websites appear in the buyer marketplace. Two gates, in that order, doing different jobs. Verification answers who controls the domain. Approval answers whether the site belongs in front of buyers at all.
Be precise with clients about what this covers. It is a control check plus an editorial review, not a guarantee of a ranking, and no marketplace can honestly sell the second thing. What it does buy you is the removal of the most common failure: paying for a placement the seller could never have delivered.
Filters decide what you consider. Verification decides what is allowed to be considered at all.
Three: a record that outlives the campaign
An order here is not a receipt. It moves through four states — new request, approved, pending, completed — with three ways out: the publisher rejects it, the buyer cancels it, or it is refunded. Every one of those moves writes an audit line with the actor and the timestamp, so the question “who changed this, and when” always has an answer.
The line itself carries the fields a client asks about, as first-class data rather than notes in an email: anchor text, target URL, placement URL, project, placement kind, turnaround, price and state. That is the difference between reconstructing a month of work from your inbox and exporting it.
Money follows the same discipline. Orders are paid from workspace credit, and when a publisher declines the work the credit returns to the wallet with the reason attached to the order. Nothing sits in limbo waiting for a support thread, and a single rejection in a twelve-line bulk order does not hold up the other eleven — each line is its own order with its own state.
Six months later this is the only feature that still matters. A link either exists at the URL you were given or it does not, and the record is what lets you check without asking anyone.
What the three add up to
Read them as one sentence: you can only buy from sites that proved they exist and passed review, you choose among them on thirteen fields that refresh daily, and every purchase leaves a record you can hand to a client or an auditor. Each feature covers the weakness of the one before it. Filters without verification are a well-organised way to buy nothing. Verification without a record is a promise you cannot evidence later.
And what we are not claiming
Three features means three, so it is worth naming the things people assume and we do not offer. There is no escrow: orders are paid from workspace credit and refunded to it, which is simpler and worth describing accurately. There is no Search Console attribution yet, so the order record and the audit trail are your evidence, not a traffic graph. And no vetting step, ours included, can promise a ranking movement. Those three sentences have cost us fewer awkward calls than any feature we have shipped.