Most niche edit problems are visible before you pay. The trouble is that few buyers look. A shortlist arrives, the numbers seem fine and the order goes through. This guide covers the checks that take about ten minutes per article and catch the common failures.
Start with indexing
Copy the exact article address and search for it in Google. If the page comes back, it is at least in the index. If it does not, ask why. Some pages are new, some are blocked and some have been dropped. A link on a page Google does not show is a link with very little chance of being counted.
Also look at the page source or an SEO tool for a noindex tag or a canonical pointing somewhere else. Either one changes the picture.
Judge the topic from the reader's side
Read the full article, not just the paragraph where the link would go. Ask a simple question: would a reader of this page plausibly want the page you are linking to? If the honest answer is a stretch, the link will look like one.
Then check the site. A publication with a clear subject and a consistent voice is a better host than one that covers insurance on Monday, gardening on Tuesday and cryptocurrency on Wednesday.
Look at what else is on the page
Count the outbound links. A few links to well-known sources are normal. A dozen links to unrelated small businesses, each with a different commercial anchor, is a pattern that suggests the page is for sale.
- Do the other links share a topic with the article?
- Do any anchors look like search queries rather than natural phrases?
- Are there links hidden in the footer or sidebar that have nothing to do with the site?
Check the history of the domain
An SEO tool can show whether a site has changed subject, lost a large share of its visibility or carried a very different set of pages in the past. Repurposed domains often show a clean authority number alongside content that no longer matches. Treat a sudden change as a reason to look harder, not as proof of a problem.
Treat traffic numbers as estimates
Traffic figures in third-party tools are modelled, and two tools often disagree. Use them to compare articles against each other, not as a promise about what any one page receives. If you have access to the site owner's own analytics, ask to see them. Many will not share, and that is fair, so rely on the estimate and your other checks.
Read the exact insertion before it goes live
Ask to see the sentence or the short passage that will carry your link, along with the anchor. Ask whether it reads naturally if the link is removed. A good insertion looks like part of the original writer's thinking. A poor one is a sentence that appears to have been added by someone else, because it was.
Think too about where in the article it sits. A link in the opening paragraph, in a paragraph about the right subtopic, usually fits better than one added at the bottom in a block of unrelated names.
Confirm the result afterwards
Once the link is in, open the live URL in a private window. Check that the link points to the right page, that the anchor matches what was agreed and that it is not marked in a way you did not expect, such as a sponsored or nofollow attribute. Save the address and the date. You will want that record later when you review the work.
Know what checks cannot tell you
None of this predicts how Google will treat a link. The checks lower the odds of a poor placement. They do not turn a purchased link into a natural one, and buying links to influence ranking can go against Google's spam policies. If your site is new or has had a manual action, think carefully before buying any links at all.
If you want the standards we aim for in our own shortlists, they are on the inventory criteria page, and the risks and vetting page covers the warning signs in more depth. You can also send us a page and niche through the contact page if you would like to see what a shortlist looks like.
A note on limits: search guidelines and tactics change, and links are one signal among many. Results depend on your site, your competition and your category, and no result can be promised.