Quick answer

Don't choose a link building platform based on the number of sites, the average DR, or an "unsponsored" list that contains affiliate links. Only buy when the placement serves a real audience or awareness goal, the site has a coherent editorial line, the commercial relationship is transparent, and the link is properly qualified, in particular with rel="sponsored" for Google.

Before any payment, check the publisher, the relevance, the real traffic, the future page, the link neighborhood, the network ownership, the attributes, the retention, the disclosure, the invoice, the right of correction, and the measurable outcome. If the offer mainly sells an exact match followed anchor "to push the ranking," refuse it: that is exactly the kind of scheme Google classifies as link spam.

Key takeaways

  • A marketplace is a media buying channel, not proof of editorial authority.
  • The catalog visible before payment must be verified at the page level, not the domain level.
  • A paid placement should carry audience value even if it never improves a position.
  • Footprints (same owners, inconsistent themes, templates, anchors, and dates) reveal network risk.
  • A responsible ranking publishes method, date, commercial relationship, tested sample, and reasons for exclusion.

The twelve checks

# Check Evidence to request Immediate refusal
1 Publisher identity Legal notice, team, editorial history No findable owner or disposable domains
2 Audience relevance Recent articles and a coherent readership Multi topic site with no identifiable public
3 Real traffic Sources, trend, countries, active pages Undated screenshot or modeled traffic sold as Analytics
4 Page and placement Section, depth, context, other links Orphan page created only to sell
5 Editorial independence Right of rewrite and publisher validation Imposed text with a guaranteed exact anchor
6 Link attribute sponsored, nofollow, ugc, or followed based on relationship Promise of a paid followed link for ranking
7 Disclosure Visible sponsored mention and contract Refusal to state the commercial relationship
8 Neighborhood Outbound links from the page and the site Casino, credit, health, and SaaS sold side by side with no editorial line
9 Ownership and footprint Network, IP/CDN, templates, authors, ad regie Hundreds of cloned domains and the same authors
10 Survival Duration, procedure if the page is removed, refund "Lifetime" with no commitment or recourse
11 Content quality Author, source, review, originality Generic article produced at volume
12 Measurement UTM, referrals, brand searches, review No KPI beyond DR and position

This table doesn't assign a Google score. It turns a purchase decision into a verifiable file.

Why platform rankings are often fragile

A comparison can monetize sign ups, affiliate clicks, or sponsored placement. That model isn't illegitimate, but it must be disclosed. The conflict appears when a page calls itself "unsponsored" while collecting a commission, with no explanation of how that shapes the order, the criteria, or the tests.

A credible methodology answers these questions:

  • which platforms were included and why;
  • which offer and budget were tested;
  • how many sites were sampled;
  • were the placements actually purchased;
  • what verification date applies to prices, inventory, and attributes;
  • how do affiliate links fund the page;
  • which criteria are weighted, and by whom;
  • what evidence was missing;
  • how to request a correction.

Without a mystery shop and a test log, don't publish "best platform." Publish a selection grid instead, and note that features need revalidating.

Due diligence procedure

1. Define the non SEO objective

Write down the value if the link carries no signal at all: sector audience, credibility, campaign, referral, or partnership. If the answer is "none," the purchase depends too much on hoped for manipulation.

2. Sample before you sign

Randomly pull at least twenty sites matching your filter, then five recently sold pages if the platform provides them. Don't let the sales rep pick only their best examples.

3. Verify the reality of the metrics

Compare vendor data, third party tools, brand search, page visibility, and editorial activity. A third party metric is an estimate; it isn't Analytics. Note the discrepancies instead of selecting the favorable number.

4. Map the footprint

Look for owners, ad regies, authors, legal notices, themes, templates, dates, and neighborhood. A single group can legitimately own several media outlets; the risk comes from a lack of editorial line and manipulative scale, not from a shared IP alone.

5. Read the contract and the attributes

Check duration, correction, removal, refund, disclosure, content rights, invoicing, and data handling. Google recommends sponsored for a paid link (Google, Qualify outbound links).

6. Run a limited cohort

Buy only a few comparable placements. Keep control URLs or a baseline period, then measure survival, referrals, impressions, and conversions. Don't change titles, content, internal links, and backlinks all at once if you want to learn something.

7. Decide with a stopping condition

Stop if more than a pre-defined threshold fails: deindexed pages, changed attribute, disappearance, no real audience, or non compliant content. A bad first batch doesn't justify "buying more to dilute it."

Worked example

An agency has 3,000 euros. Platform A offers ten domains with DR 50+; platform B offers three sector media outlets with sponsored, a stated audience, and a newsletter. The goal is demo requests from accounting firms.

Sampling A reveals articles across 25 topics, generic authors, no engagement, and outbound links to risky sectors. The pages are indexed but orphaned. B shows a sector audience, pages reachable from the section, a sponsor policy, and UTMs. Even with three sponsored links, B is the better media buy for the goal. A might show a higher DR with no useful reader.

The agency sets before launch: at least 100 cumulative qualified visits, a reading rate comparable to the site, two assisted demo requests, and 100% of pages still live after six months. These thresholds are internal, not market benchmarks.

Scorecard and decision rule

Use the backlink scorecard.csv. Positive dimensions total 14 points: relevance 0 to 3, audience 0 to 3, independence 0 to 3, context 0 to 2, indexability 0 to 1, verified traffic 0 to 2. Subtract footprint risk of 0 to 3.

The formula orders the review; it doesn't auto-approve anything. Three conditions override the score: a hidden relationship, a placement explicitly sold to manipulate Google, or illegitimate ownership or activity. A site at 11/14 can still be refused.

What the rules prove and don't prove

Google classifies buying or selling links for ranking, excessive exchanges, automation, and guest posts with optimized anchors as link schemes (Spam policies). This policy doesn't say every sponsored article is banned; among other things, it asks that the link be properly qualified. It doesn't provide a "toxicity" threshold or a domain score.

An Ahrefs intervention on three of its own articles disavowed 3,476 links for four weeks, then removed the file; the company observed a drop and then a recovery in its visibility estimates (Ahrefs). Three pages, a short window, no randomized control, and a proprietary metric don't allow you to estimate the effect of a purchased link. Above all, reproducing this test on a client site would be reckless.

Common mistakes

  • Publishing 40 platforms without verifying a single placement.
  • Confusing a DR filter with an editorial review.
  • Buying the lowest price then paying for the rewrite and the removal.
  • Using the same exact anchor across an entire cohort.
  • Accepting a page not linked to the main site.
  • Not checking the link at 30, 90, and 180 days.
  • Comparing a sponsored audience link to an editorial link as if they served the same goal.
  • Forgetting the affiliate disclosure on your own comparison.

How SEOryon fits in

SEOryon can help identify gaps, create an asset that deserves coverage, write a sourced page, and track organic results. It states it refuses low quality link pools. This position doesn't replace the due diligence of an external purchase: owner, contract, attribute, context, and audience must be validated by a responsible person.

Measurable exercise

Sample 20 domains from two platforms without looking at their recommended order. Score them with the scorecard, document three absolute refusals, and calculate the cost per qualified placement instead of the listed cost. The test succeeds if the evidence is archived and your decision stays the same once the platform names are masked.

FAQ

There is no universal winner. The choice depends on audience, market, editorial control, transparency, and risk. A ranking with no dated test doesn't answer your case.

It can bring audience and awareness; Google asks that it be qualified and guarantees no ranking effect. Evaluate it as a media purchase.

Should minimum DR be 30, 40, or 50?

No threshold is a Google criterion. Use these metrics to filter, then check the page, the audience, the relevance, and the independence.

Aim for a duration consistent with the editorial value and the contract, then check its survival. A "lifetime" promise only holds value if the publisher and the page survive.

How do you spot a site network?

Cross-check ownership, templates, authors, sections, neighborhood, dates, policies, and ad regie. No single indicator is enough; the accumulation of footprints and the lack of audience justify refusal.

Main sources

Method note

Method: no named ranking is published without a mystery shop. Sources verified 16 July 2026, translated and edited 22 July 2026; revalidate rules, inventory, and attributes before purchase.