Back to blog
Link Building

How to Vet a Website Before You Buy a Link on It

A plain checklist for vetting a site before you buy a link: traffic, outbound links, content, the page itself and the red flags to walk away from.

How to Vet a Website Before You Buy a Link on It - BuildLinks blog

Before you pay for a link, check the site in a set order: its traffic first, then the sites it links to, then its content, then the exact page your link will sit on, then a few red flags. Stop at the first check it fails. Most bad links could have been spotted before anyone paid for them.

Vetting usually gets skipped because someone is rushing through a seller's list of hundreds of sites. This checklist is worth slowing down for. It is step 3 of the six-step process we use to buy backlinks for clients.

1. Check traffic first, not authority scores

DR (Domain Rating) and DA (Domain Authority) are 0 to 100 scores that estimate how strong a site's links are. They are the first numbers people quote, and the easiest to inflate. A site can keep a high score from old links while almost nobody finds it in Google anymore.

Organic traffic (visitors who arrive from search results) is much harder to fake. So check it first:

  • Look at the traffic trend for the last 12 to 24 months, not just today's number. You want a steady line or gentle growth.
  • A sudden, steep drop usually means a Google update or penalty hit the site. Do not put your link there.
  • Check where visitors come from. A "US lifestyle blog" whose visitors are mostly from unrelated countries may have bought its traffic.

Also check how many other sites it links to. A site that links out to thousands of domains spreads its value thin. The reason lies in what moves Domain Rating.

The quickest sign that a site sells links to anyone is the list of sites it links to. Open five recent posts and look at where the links go. Casino, CBD, essay-writing and payday loan links inside a parenting blog are not editorial choices. They are paid ads. When a vendor sells the placement itself as an editorial link, our guide to telling a real editorial link from a paid one sets out the questions to ask first.

Outbound links from five recent posts on two sites: an editorial blog links to useful sources, a lifestyle blog links to casinos, lenders and essay services
The same five-post check on two sites. One odd link is noise; one in almost every post is a price list. Illustrative example, not real sites.

Volume matters too. A site that publishes thirty "guest posts" a month, each with two or three links to sales pages, is a link farm: a site that exists to sell links. Links from it can drag your client down with it. We set out six checks that catch a link farm in a vendor list, starting with how many sites it links to compared with how many link to it.

One good sign: if the site already links to two or more of your client's competitors, its editors clearly cover that topic. Who a site links to is one of the five relevance layers we score before buying.

Read two or three articles like a normal reader. The writing does not need to be brilliant. It just needs a reason to exist beyond the link it carries.

Warning signs include thin posts that say nothing, generic intros, and paragraphs written only to hold a link. Also check the categories. If "Finance", "Health" and "Travel" all started on the same day under the same author, the site was built to sell links in every niche at once. For guest posts, we also check what separates a guest post that counts from one that only holds a link.

A site can pass every check above and still put your link on a page that does nothing. These page checks take about five minutes:

  • Google can index the page. "Indexed" means Google has the page in its search results. Look in the page source for a "noindex" tag. Google's guide to blocking indexing with noindex says a noindexed page is dropped from Search completely.
  • For a niche edit, the page is already indexed. A niche edit is a link added to an article that already exists. Search Google for the exact page address. If it does not show up, Google has not picked up the page. If it does, read the page-level URL Rating and what feeds it rather than the site's DR alone.
  • The page is not a list of links. An article with a dozen links to unrelated businesses is a directory, not an article.
  • No competitor on the same page. Ask the seller, in writing, whether a competitor can buy a link on the same article later.
  • Your link goes in the main text. Not in an author bio, a "resources" box or the footer. A footer or sidebar slot repeats on every page, and sitewide links sold in a template block are a different purchase with their own risks.

5. Red flags that end the conversation

  • The domain was recently bought and its new content has nothing to do with its past. This is an old domain rebuilt to sell links, often one of several in a private blog network, and a vendor's list can hide several of them.
  • No real people anywhere: no author pages and no about page.
  • The same site is offered by three different resellers at three different prices.
  • The number of pages Google has indexed keeps falling, month after month.

What Google's rules say

Google's spam policies explain why these flags matter. In plain terms:

  • Link spam: buying or selling links to boost rankings counts as spam, and so do keyword-stuffed links in guest posts. Paid links are fine when they carry rel="nofollow" or rel="sponsored", tags that tell Google the link should not pass ranking value. This is one reason we plan anchor text for each page before ordering.
  • Expired domain abuse: buying an old domain and reusing it mainly to game rankings.
  • Site reputation abuse: a respected site that hosts unrelated paid content just to borrow its reputation.

Google's September 2026 spam update rolled out from September 24 to October 8, 2026. Google did not say what it targeted. If a site lost a big share of its traffic in that window, treat it as the sudden drop from step 1.

Timeline of the traffic history to check, with the September 2026 spam update window marked at the end
A large traffic loss during the September 2026 spam update is a reason to reject the site.

Vetting while a core update is rolling out

A core update is not a spam update, and the two call for different reactions. Google started the November 2026 core update on November 10, 2026 and said the rollout may take up to two weeks. Traffic often swings both ways during a rollout and settles afterwards, so a drop that starts inside that window is not yet a verdict on the site.

For a site that passes every other check, hold the order and look at the traffic graph again two weeks after Google marks the rollout complete. If traffic has recovered, approve it. If it settled at half or less of its old level, reject it as you would any other sudden drop. A site that also fails step 2 or step 3 needs no waiting: reject it now.

Pass and reject thresholds we use

A checklist only saves time if everyone on the team reads a check the same way. These are the defaults we start from. Tighten them for a client in a sensitive niche such as finance or health, and write any change into the client's brief so the next person applies it too.

CheckPassReject
Organic traffic trend, 12 to 24 monthsFlat or growing; any dip recovered within two to three monthsLost half or more of its traffic and stayed down, or lost a large share during a spam update
Organic traffic levelAt least 1,000 visits a month for a paid guest post; for a niche edit, the page itself ranks for somethingUnder a few hundred visits a month on a site sold at a mid-market price
Traffic countryMost visits come from the client's target marketMost visits come from countries the site's content never mentions
Outbound links in five recent postsNo casino, loan, CBD or essay links, or one odd link at mostTwo or more of those posts carry them
Guest post volumeA handful of contributed posts a month alongside the site's own contentMost new posts are guest posts with links to sales pages
Target pageIndexable; for a niche edit, already indexed; link in the main textNoindex, not indexed, a dozen unrelated links, or a bio or footer link
PriceWithin the usual range for its DR and trafficFar above that range with nothing to justify it, or offered by several resellers at different prices

The traffic level is the threshold people argue about most. A small site with 800 visits a month from exactly the right readers can be worth more than a general site with 20,000, so treat that row as a prompt to look closer, not an automatic rejection. The outbound link and target page rows are the ones not to bend.

Example: three sites from one seller's list

These sites are made up. A seller offers three guest post options for a client that sells restaurant payroll software, sorted by DR, which is how most seller lists arrive.

SiteWhat happenedVerdict
X: food magazine, DR 58Traffic fell from about 40,000 to 6,000 a month, mostly in one monthRejected at step 1
Y: small business blog, DR 41Steady traffic, but casino or loan links in four of five recent postsRejected at step 2
Z: restaurant operations blog, DR 33Growing traffic, useful links, named authors, clean pageApproved

The takeaway: the order of checks saves time. Site X took one look at a traffic graph. Only Site Z needed a full review, and a DR sort had put it at the bottom of the list.

6. Write down the result, so you only vet once

Teams often check the same sites again and again because nobody recorded the answer. The fix is simple: keep one list of every site you have checked, one row per site. Record:

  • The site and the date you checked it
  • Traffic now and a year ago, and the main country
  • What you found on links, content and the page
  • DR on that date, and the price you were quoted
  • Approved or rejected, the step it failed at, and who checked it

Over a quarter, the same list shows which vendors keep sending sites that fail, a number worth reading next to the live rate and survival rate of each vendor's links.

Compare the price with typical guest post prices by DR. Also compare it with the rates you charge clients, so you catch a good site that would still lose you money.

If you can, keep the list in software built for managing link building rather than a spreadsheet. Checking a site once a year is careful. Checking it on every order is wasted time.

We built BuildLinks around this idea. It has a shared site list your whole team orders from, and new sites must be approved before anyone can quote them. The checking one person does in March still pays off in November.

Vetting does not stop once you pay. When the link goes live, and again from time to time, confirm that:

  • The link is on the page you were told, pointing at the right address.
  • It uses the agreed anchor text (the clickable words) and the agreed tag. Read the rel sponsored tag on the live page to confirm it.
  • The page can still be indexed, and the link has not been switched to nofollow. If a new page is still missing from Google after 30 days, follow our steps for getting a paid placement indexed.

Check niche edits more often, because the site owner keeps editing those pages, and time each recheck so you can still claim under the vendor's terms; our guide to catching lost links before a replacement window closes sets out the schedule. If a site you bought from later turns into a link seller, decide whether its toxic backlinks need a tag, removal or a disavow, and what to do with links already placed. For a new client whose links you inherited, run these same site checks inside a link audit of the client's existing profile before the first order. Feed the results into your monthly client report.

If you resell, use a full quality checklist for white label link building. It should cover both guest posts and niche edits.

Conclusion

Check every site in the same order and stop at the first failure: the 12 to 24 month traffic trend, who its recent posts link to, whether the content would exist without the links, then the exact page your link will sit on. Write your pass and reject thresholds down, and never bend the ones on outbound links and the target page. Reject sites hit by the September 2026 spam update, rebuilt old domains and sites with no real people. If a site only dipped during the November 2026 core update, look again two weeks after the rollout ends.

Record every result once, with the date, traffic, DR, price and verdict, so the next order starts from your notes. After the link goes live, confirm the page, anchor, tag and indexing, and recheck niche edits most often.

Frequently Asked Questions

We start from at least 1,000 organic visits a month for a paid guest post, with most of them from the client's target market. For a niche edit, the page itself matters more than the site: it should already be indexed and rank for something. A smaller site with exactly the right readers can still pass, so treat the number as a prompt to look closer.

DR (Domain Rating) is a score of how strong a site's links are, and it is easy to inflate. A site can keep a high DR from old links while almost nobody finds it in Google anymore. Traffic is much harder to fake, so check the 12 to 24 month traffic trend first and DR after it.

Look at where recent posts link. Casino, CBD, essay-writing or payday loan links on an unrelated blog are a bad sign, and so are dozens of guest posts a month each linking to sales pages. Thin posts that only exist to hold a link point the same way.

Google's spam policies treat buying or selling links to boost rankings as link spam. Paid links are allowed when they carry rel="nofollow" or rel="sponsored", tags that tell Google not to pass ranking value. Vetting lowers your risk of a bad site, but it does not change how Google treats a paid link.

Checking a site once a year is careful; checking it on every order is wasted time. Keep one list of the sites you have checked, with their numbers, price and the date, so nobody repeats the work.

Vet once, reuse forever.

Keep every vetted domain - with its DA, DR, traffic, and pricing - in one searchable inventory your whole team orders from.

Start free trial