How to Vet a Guest Post Publisher Before You Pay (The Auditor's Checklist)

Every week, thousands of marketers pay for guest post placements based on a single number in a seller’s spreadsheet. And every week, a meaningful share of that money buys links from sites that exist for no other reason than to be sold — sites whose metrics were manufactured precisely because buyers check metrics and nothing else.

Vetting a publisher properly takes about fifteen minutes. Here is the full method, the way an auditor would run it — trust nothing that’s claimed, verify everything that’s checkable — ending with the 10-point checklist we run before any placement gets paid for.

Why authority metrics are manufactured exactly where you’re looking

Domain Rating, Domain Authority and their cousins are third-party estimates built mostly from backlink data. That makes them useful weather vanes — and trivially gameable, because backlinks are exactly the thing a link seller knows how to manufacture. Point a few expired domains with legacy equity at a shell site, interlink it with the rest of the farm, wait a few weeks, and the dashboard reads DR 65. Nothing about the site’s actual standing with Google has changed, because Google doesn’t use any of these scores.

Here’s the economic logic worth internalising: metric inflation concentrates wherever buyers rely on metrics. Guest post marketplaces are the highest-scrutiny, highest-volume link market on the internet — so that’s precisely where the most sophisticated fakery lives. The DR number on a marketplace listing isn’t evidence of quality; it’s the product’s packaging. In competitive verticals — ours is iGaming, where restricted supply pushes prices up and fakery follows the margin — the packaging is often all there is.

None of this makes DR useless. It makes DR a tiebreaker among sites that already passed real checks — never a reason to buy.

Check 1: the live-traffic test

Open the site in Ahrefs or Semrush and look at the organic traffic graph over 2–3 years — the curve, not the current number.

What a real publisher looks like: a curve that breathes. Gradual growth or gentle decline, seasonal waves, visible but survivable dips at algorithm updates, traffic distributed across dozens or hundreds of pages, and a keyword profile that deepens over time. Real sites are boring in the best way.

What a farm looks like: geometry. A dead-flat line at near-zero followed by a vertical cliff upward (equity injection); a sawtooth of sharp spikes that decay within months (churned expired-domain tricks, repeated); a single hero page carrying 80% of all traffic; or a recent all-time peak on a site whose content is visibly thin. Any of these shapes on a site with a high authority score is the mismatch that tells the whole story: manufactured authority produces metrics without traffic; earned authority produces both.

As a working floor, a publisher worth paying standard guest-post rates should show at least four figures of monthly organic traffic, earned across many pages. Below that, whatever the score says, you’re buying a link from a site search engines barely visit.

Check 2: keyword-market match — the costume test

Traffic volume alone isn’t enough; look at where it comes from and what it ranks for.

A seller pitches you a “premium US business site”. The traffic tab shows 85% of visits from geographies irrelevant to your market, ranking for keyword salads in languages the “US business site” doesn’t publish in. That’s not a publisher — that’s a costume. The site is wearing an audience it doesn’t have, usually because programmatic pages or inherited expired-domain rankings generate visits somewhere cheap while the storefront targets buyers somewhere expensive.

The match test has two halves:

Geography: the site’s top traffic countries should overlap your target market, or at minimum the market the placement claims to serve. If you’re building for an audience in Canada and the site’s readers are entirely elsewhere, the link’s relevance signal is close to zero.
Topical coherence: scan the top 50 organic keywords. A genuine publication clusters around themes. A farm ranks for an incoherent scatter — kitchen appliances, forex, CBD, moving companies — because it publishes whatever anyone pays for. Incoherence is the diagnosis; no further checks needed.

Check 3: outbound-link forensics

Metrics can be dressed up; editorial behaviour can’t. Open the ten most recent posts and audit them like a forensic accountant:

Outbound ratio. Count external dofollow links per post. Real editorial writing links out a few times per piece, mostly to references, occasionally to a commercial site where it genuinely fits. Recent posts averaging three or more commercial dofollow links each — pointing at unrelated businesses — describe a site selling by the placement.
Niche scatter. Where do those links go? A personal-finance site linking to a plumbing firm, a crypto exchange and a dental clinic within a week isn’t diversifying; it’s invoicing.
Follow status honesty. Spot-check the HTML. Some sellers promise dofollow, publish nofollow (or sponsored), and trust you’ll never view source. Others do the reverse — sitewide dofollow on obviously paid content, which is the riskier profile for you. Either way, thirty seconds of inspection tells you whether the site’s word matches its markup.
Publishing rhythm. Five “contributed” posts a day, every day, each by a different unknown author with a suspiciously polished bio? That cadence is industrial, and Google’s spam systems are pattern-matching the same thing you are.

Check 4: update-survival history

Every core update and spam update is a public stress test that Google runs on the entire web, free of charge. Use it. Overlay the site’s traffic curve against the dates of the major updates from the past two years and read the verdicts:

Passed: flat or rising through updates — the site’s model has Google’s ongoing approval. This is the strongest single quality signal available to you.
On probation: one visible hit with a slow, genuine recovery (recovered rankings on the same pages, not new spike pages) — acceptable with the other checks green.
Failed: repeated drawdowns, each partially recovered by fresh tricks; or a catastrophic single collapse the site never came back from. Links from these domains are claims on an asset Google has already written down. It does not matter what the DR says — the landlord already condemned the building.

This check takes half a minute and kills more bad inventory than everything else combined.

The price-vs-value logic: cheap is the most expensive thing you can buy

Genuine placements have a cost floor, because real publications have editors, standards and refusal rights — and in restricted verticals, scarcity pricing on top. When you see “DR 70, dofollow, $60”, the price is the disclosure: no real site with real traffic sells its editorial voice for less than the content costs to produce.

Run the actual arithmetic. A $60 farm link that passes zero value — or worse, sits in your profile as a liability you’ll later pay a cleanup consultant to disavow — has infinite cost-per-value. A $500 placement on a publisher that passes all four checks, keeps compounding as the publisher grows, and survives every update is among the cheapest marketing you’ll ever buy. Divide price by expected surviving value, not by the number on the invoice. The cheapest DR70 on the marketplace is, by this math, usually the most expensive link on your ledger — you pay for it three times: at purchase, in results that never come, and at disavow time.

This is the entire reason our link building service publishes its quality gates instead of a price-per-DR menu — and why, in a vertical as scrutinised as ours, iGaming SEO programs live or die on supplier discipline. We covered the vertical-specific half of that discipline — anchors, velocity, tier-2 — in the iGaming link building playbook.

The 10-point pre-payment checklist

Print it, pin it, run it before any money moves. A placement needs 8 of 10 to pass; any single ❌ on points 1–3 or 8 is an automatic no.

Live traffic, not screenshots — four figures of monthly organic minimum, pulled in Ahrefs/Semrush yourself; if the seller’s screenshot and the live curve disagree, you already have your answer.
The shape of the curve — the 2–3 year graph breathes (gradual, seasonal); no dead-flat line ending in a cliff, no sawtooth of decaying spikes, no single hero page carrying the whole site.
The costume test — top traffic countries and top-50 keywords match the market and themes the site claims to serve; a “US business site” ranking somewhere else entirely for keyword salads is wearing a costume.
Outbound-link forensics — the ten most recent posts average under ~2 commercial dofollow links each; a publication links to sources, a billboard links to buyers, and forty per post settles which one you’re reading.
Niche scatter — those outbound links land in coherent niches; finance-to-plumbing-to-crypto within the same week isn’t editorial range, it’s invoicing.
Follow-status honesty — the markup matches the promise, verified in thirty seconds of view-source; treat “dofollow forever” as a promise nobody can honestly make.
Publishing rhythm — the cadence looks human; ten posts a day for a month then silence, each by a different one-time “contributor”, is industrial batching that Google’s spam systems pattern-match as readily as you can.
Update survival — flat or rising through the last two years of core and spam updates; one honest recovery at most, and no serial resurrection by fresh tricks after every hit.
Price vs surviving value — divide the quote by the odds the link still exists and still counts next year; by that arithmetic the deep-discount “premium” site is confessing, and the cheapest DR70 on the marketplace is usually the most expensive link on your ledger.
Editorial friction — there are guidelines, a review step, and things they turned down last month; a site that never says no to anyone is saying something about itself.

Score honestly and you’ll reject most of what marketplaces offer you. That’s not the checklist being harsh — that’s the market being what it is. The placements that pass are the ones that still count in three years.

Want your current suppliers scored against this checklist by people who run it every week? Send us your list — the first reply tells you which links you paid for and which you merely bought.