How do you vet a backlink vendor before you spend a dollar?

I have bought backlinks across every tier from 2018 to 2026 — wholesale catalog listings at $30, agency invoices at $400 sourced from those same catalogs, and a few PBN networks that lost their juice in a Penguin-style update along the way. If your last vendor invoice felt high or your last "white-hat" link disappeared from the index, you have already lived a lesson this post tries to teach. Below: a six-metric checklist, three reliable PBN tells, the wholesale-versus-retail markup math most vendors will not surface, and a clawback-prevention layer that keeps a single bad vendor from dominating a small profile. Start with the entity-based SEO framing for investor sites; this is the link side of that picture.

What does a vetted backlink site actually look like?

A vetted backlink site has real organic search traffic, an editorial process, topical relevance to the buyer, and a clean link neighborhood. Vetting matters most for low-volume real estate keywords because a single bad link can dominate a small profile.

A vetted site is a real publisher. It has a real editorial process, real organic traffic from real users typing real queries into a search engine, and a topical map that overlaps with the buyer's site at the entity level — not the keyword level. Provenance, traffic, topical relevance, and link neighborhood are the four parts of the test. A donor site that fails any one of them is not a vetted site, no matter what its DR says.

The four parts are independent of one another. Provenance is who runs the site and how publication decisions get made. Traffic is whether actual humans arrive from organic search, not whether a tool reports a number. Topical relevance is whether the site's existing entity coverage overlaps with the buyer's topical map — a real estate investor's link profile gains nothing from a "marketing" site that has never published a real estate piece. Link neighborhood is the company the donor site already keeps: who links to the donor site, who the donor site links out to, and whether those neighbors are themselves publishers or shells.

Vetting matters more on a small profile than on a large one. A real estate investor site usually targets low-volume queries with thin SERP competition; the link profile is small by definition, and a single misjudged link from a thin neighborhood can dominate the signal Google reads from the whole profile. Post 1 covers the entity-based framing that makes this risk concrete; this post turns that framing into a vendor checklist.

Technical note. Sites without organic traffic have no audience to give the receiving page; sites with organic traffic but on irrelevant topics give an audience that has no reason to click through. Both fail the test for different reasons.

Which metrics actually predict link value (and which are vanity)?

Six metrics actually predict backlink value: organic search traffic, referring-domain growth pattern, topical relevance, branded-vs-anchor traffic ratio, indexed-page-vs-sitemap count, and outbound link density per article. DR or DA in isolation is a vanity tiebreaker for buyers, not a primary filter.

Organic search traffic is the closest single proxy for link value. A site that earns visitors through search has an audience, an algorithm score, and a track record. A site that posts traffic counts without a source breakdown is asking the buyer to take the number on trust — direct visits, social referrals, and bot hits all count toward the same headline figure, and only the search slice carries the SEO weight a buyer is paying for.

Referring-domain growth pattern matters next. A site that has accumulated referring domains steadily over years looks different from one that gained two thousand RDs in a month and held flat ever since. Pull the RD chart in Ahrefs or Semrush and look at the slope, not the total.

Topical relevance is the third metric, and it is computed at the entity level. A real estate investor site links best from real estate, finance, business, or home-services neighborhoods. A "marketing" or "lifestyle" site that has never covered the buyer's topics gives Google an audience signal that does not match the receiving page.

Branded-vs-anchor traffic ratio separates real publishers from PBNs cleanly. Real publishers earn brand searches over time — readers come back, type the publication name into Google, and arrive directly. PBNs almost never earn brand searches because no one is reading them.

Indexed-page-vs-sitemap count tells you whether Google trusts the site. A donor with five thousand pages in its sitemap and four hundred pages indexed is signaling that most of its publishing is throwaway. Outbound link density per article closes the list — a real editorial publication carries a few external links per piece, while a network site stuffs ten or twenty into a single page.

DR and DA in isolation are tiebreakers, not filters. Ahrefs' published guidance on Domain Rating is direct on this point — DR is correlation with backlink-profile strength, not causation of receiving-site rank. Spam-score badges are proprietary and opaque. Use them last, after the six metrics that actually mean something.

Technical note. DR is a logarithmic score on Ahrefs' index. Moving from DR 30 to DR 40 represents a much smaller absolute backlink gain than moving from DR 60 to DR 70. Apply this comparison within a tier, not across tiers.

How do you detect a PBN before you buy?

Three reliable PBN tells appear before purchase: bulk-registered WHOIS with shared hosting IP and clustered registration dates; recycled site architecture and authors across "different" sites; convergent outbound links pointing at the same handful of money sites from every network member.

WHOIS and hosting are the first tell. A network of "independent" sites that all registered through the same registrar in a three-week window, sit on the same hosting IP class, and share an SSL certificate provider is a network. Run the homepage IP through any hosting lookup tool and the sites in the same network usually show up clustered on the same handful of addresses. The pattern is structural — bulk acquisition leaves bulk fingerprints, and bulk fingerprints stay in WHOIS records and DNS until the operator goes through the painful work of moving each site to a separate host.

Content fingerprints are the second tell. Open three "different" sites in the same network and compare site architecture: the navigation, the page templates, the category structure, the typography. PBN operators rarely build a unique theme per site — the same WordPress template recurs, the same author headshots get recycled with different names, the same generic stock photos repeat across the network. Posting frequency tells the same story from a different angle. A real publication posts on a rhythm; a PBN posts in bursts when the operator runs a content campaign and pauses for months between bursts.

Outbound links are the third tell, and the most reliable. Every site in a PBN points at the same handful of money sites — sometimes through a footer module, sometimes through "editor's picks" sidebars, sometimes through anchor-rich placements in posts. Pull the donor site's external link profile in Majestic's Trust Flow / Citation Flow view and look for a Citation Flow score significantly higher than its Trust Flow score. CF much higher than TF, paired with the WHOIS and hosting cluster, is the fast triangulation.

The full check takes about ten minutes per donor site once the workflow is set up. Most agencies do not run it because the wholesale catalog they source from has already done a basic version. The careful buyer runs it again, because the version a wholesaler ran last quarter does not protect against a network the wholesaler missed this quarter.

What does wholesale vs retail link pricing actually mean?

Wholesale link pricing is the aggregator's flat per-site cost: typically thirty to eighty dollars in real estate niches. Retail pricing is the vendor's marked-up rate covering service overhead, typically one hundred fifty to five hundred dollars for the same placement.

Most paid placements sold as "guest posts" by SEO agencies are sourced from a small number of wholesale publishing networks. The networks aggregate publisher inventory at a flat per-site price and resell it to vendors who pass it through to end buyers at a markup. The same DR 50 real estate publisher appears in multiple agency catalogs at different retail prices, because the underlying inventory is shared and the retail markup is set independently by each vendor.

The wholesale price band for real estate niches typically runs $30 to $80 per placement. The retail price band typically runs $150 to $500 for the identical placement on the identical site. The 3x to 5x spread is not a markup on link quality. The link is the same link. The spread covers vendor-side relationship management — the briefing call with the buyer, the article draft, the editorial revisions, the publication coordination, the post-publish tracking. Each of those is real labor, and each is priced into the link the buyer sees on the invoice.

REI Spark's marketplace surfaces the wholesale catalog directly. Each site in the catalog shows the wholesale price plus a flat $15 fee or 20% markup, whichever is higher. A site at $30 wholesale ships at $45. A site at $80 wholesale ships at $96. A site at $400 wholesale ships at $480. The fee structure is published, the wholesale price is visible, and the buyer can decide whether the in-house labor of briefing, revisions, and coordination is worth handling directly or worth paying a retail vendor to handle.

This is the model the operator who built REI Spark uses for personal acquisitions across three cash-buying brands. The catalog is the same one those brands buy from — no inventory split between the operator's own brands and the marketplace customers.

Technical note. The wholesale catalog is REI Spark's wholesale publishing network — a shared pool of vetted publisher inventory that agencies and direct buyers both draw from. Multiple agencies and direct buyers source from the same pool simultaneously, which is why the same site appears in many catalogs.

Why do most agencies mark links up 3–5x?

Most agencies mark wholesale links up three to five times for legitimate client-service overhead: briefing, revisions, account management, retainer bundling. The markup is rational. The lack of any source transparency prevents the buyer from doing the math themselves before purchase.

The first answer most agencies give to this question is "service quality." That answer is partially true. It does not explain a 3x to 5x spread on identical inventory.

The structural reason an agency must charge that much is overhead. Every retail-priced link carries an account manager's time on the briefing call, an editor's time on the draft, an account director's time on the quarterly report, and a fraction of the agency's senior-staff and software costs. Each is real labor. The labor is priced into the link because most agencies do not separate strategy from execution on a line-item invoice. A retainer bundles strategy, link acquisition, reporting, and revisions into one number, and the link is the deliverable that gets named.

The second structural reason is opacity. Most agencies do not publish the wholesale source. The buyer cannot price-compare across vendors because the buyer does not know what the same site costs at a different vendor or directly from the wholesaler. Without that comparison, a $400 invoice line is just $400 — no benchmark to test it against. Vendors who publish the wholesale source give up that opacity in exchange for buyers who do the math and trust the result.

The trade-off framing is the right one. Paying retail through an agency is rational when the operator does not have the time or in-house capacity to vet sites, write briefs, manage revisions, or follow up on placements. Paying wholesale directly is rational when the operator has that capacity. Neither choice is wrong. The choice should be made with full information about what is bundled into the price, not with the bundle hidden.

This section does not call out specific agencies. The point is pricing transparency — what the buyer is paying for, what the buyer is paying twice for when a retainer also covers strategy, and what the buyer can opt out of when in-house capacity exists.

How do you avoid toxic-link clawbacks?

Toxic-link clawback is more often silent devaluation than formal penalty. Three prevention practices: keep a link log; cap any single vendor at twenty percent of acquisitions per ninety days; distrust replacement clauses tied to devaluation evidence GSC does not surface.

A clawback is what happens when Google's classifier later decides a link came from a site it now considers low-quality or part of a manipulative network. The post-Penguin reality is that Google does not always penalize the receiving site for the bad link. More often, Google simply ignores the link going forward — the SEO benefit disappears silently, and the receiving site loses the rank lift it paid for without ever getting a manual action notice. The outcome is the same as never having bought the link, except the money is already spent.

Three practices reduce clawback exposure in advance. The first is a link log. Keep a spreadsheet with date acquired, vendor, donor URL, anchor text, target URL, and price paid. When a future audit happens — your own or someone else's — the log makes the analysis straightforward. Without a log, six months from now you will not remember which vendor placed which link on which site.

The second is a vendor concentration cap. Any single vendor should account for no more than 20% of new link acquisitions in a 90-day window. Concentration is itself a signal — a profile where one source provides most of the new links is structurally more vulnerable to a single-vendor devaluation event than a diversified profile. The cap is a portfolio rule, not a quality judgment on any individual vendor.

The third is healthy skepticism toward "if-Google-devalues-we-replace" clauses. Most replacement clauses require evidence of devaluation that Search Console does not surface — Google does not flag individual links as devalued. The clause sounds like protection but is rarely triggerable. Read the fine print before treating it as an insurance policy. Google's disavow links tool exists for the worst cases but is a last resort, not a planning instrument. The second clawback-prevention layer is anchor-text distribution — what ratio of branded, naked-URL, and exact-match anchors keeps a profile inside the unmanipulated boundary.

How does REI Spark vet sites in its marketplace?

REI Spark's marketplace draws from a forty-eight-thousand-site wholesale publishing network and applies organic-traffic, topical-relevance, DR-tiebreaker, branded-anchor, and PBN-exclusion filters before any site enters the catalog. The catalog is the same one the operator's own three cash-buying brand sites buy from.

The marketplace pulls inventory from a 48,000-site wholesale publishing network across real estate, finance, business, and home services. Every candidate site runs through a layered filter before it gets a catalog entry.

The first filter is organic traffic from search. Sites without a verifiable organic share — distinct from direct, social, and referral traffic — do not enter the catalog regardless of DR. The second is topical relevance to the buyer's vertical: a site has to overlap with real estate, finance, business, or home-services entity coverage to be eligible for an investor-site buyer. The third is a DR floor used as a tiebreaker between otherwise-equivalent sites, never as a primary signal — the framing earlier in this post applies inside the catalog as much as it does outside it.

The fourth filter is the donor site's own outbound link history. Sites that have already over-anchored their outbound links toward "money" anchors get capped or excluded; the buyer's link inherits the donor's outbound profile, and a donor with a clean outbound mix is more durable than one already dialed up. The fifth is PBN exclusion: any domain whose registrar, hosting IP, or content fingerprint matches a known PBN cluster gets blocked from the catalog by registrar/IP cluster, regardless of DR, traffic, or topical match.

Sites that pass become catalog entries with the wholesale price visible plus a flat $15 fee or 20% markup. The operator who built REI Spark's vetted link marketplace for real estate investors runs three cash-buying brand sites — Yuba Home Buyer, Fast Home Buyer California, and Sell My House Fast In CA — on the same catalog inventory. No inventory split, no selection bias, no separate tier for the operator's own brands.

What to do today

Pick three sites a current vendor has pitched and run all six metrics from the earlier section against each — organic search traffic, referring-domain growth pattern, topical relevance, branded-vs-anchor traffic ratio, indexed-page-vs-sitemap count, and outbound link density per article — before approving any of them. If even one fails, ask the vendor to swap it. For the entity-based framework underneath these metrics, start with the entity-based SEO root post. To skip the agency markup on the same inventory, browse REI Spark's vetted link marketplace.

By YK Kuliev, California DRE #02006033 — founder of REI Spark, operator of Yuba Home Buyer and Fast Home Buyer California.