Use case — MFA & made-for-AdSense filtering

Filter made-for-advertising junk out of your recruitment lists

Made-for-AdSense and made-for-advertising sites exist to farm ad impressions, not to send you buyers. We read every candidate against 120M+ classified domains and strip the MFA junk out before it reaches your outreach team — using auditable content and layout signals, never invented traffic.

hundreds of thousands

of candidate domains a single MFA sweep can screen (estimate)

10+

MFA-specific tell-tale signals per domain

invented traffic numbers — content signals only

Request a Free Pilot See Free Samples

The job to be done

MFA sites look like publishers in a list — and convert like nothing

A made-for-advertising site is engineered to maximize ad clicks: ad-dense layouts, thin templated articles, arbitraged traffic, and content built for keywords rather than readers. In a spreadsheet it reads as a real niche publisher. In your program it burns outreach time and, if approved, quietly drains payouts.

Ad-first by design

Content is a wrapper for ad units. The article exists to hold the page together long enough to load the ads, not to inform or convert a reader.

Arbitraged traffic

Many MFA sites buy cheap traffic and monetize it against display ads. That audience has no intent to buy your advertiser's product, so affiliate conversion is near zero.

Camouflaged as a niche blog

MFA operators mimic legitimate niche sites — plausible topic, stock imagery, generic bylines. A keyword or category match cannot tell them apart from the real thing.

How the pipeline filters MFA

From a raw list to a clean one — five steps

Run your discovery or applicant list through the MFA filter and the junk drops out with a labeled reason, so your recruiters only ever see sites worth contacting.

Read the layout & ad density

The site is read for ad-to-content ratio, ad placement patterns, and whether display units dominate the fold — the structural tells of a made-for-advertising build.

Assess content authenticity

An LLM judges whether articles carry real editorial substance or are thin, templated, keyword-stuffed filler generated to fill space around ads.

Check monetization intent

We look at whether the site is built to inform and recommend — with genuine affiliate context — or purely to farm impressions, and whether affiliate links are incidental to an ad-first page.

Score the MFA risk

Signals roll up into an MFA-risk grade with a confidence score. High-risk domains are flagged with the specific tells that triggered them, not a black-box verdict.

Filter & return clean

You get the cleaned recruitment list plus a separate quarantine of flagged MFA domains with reasons — so nothing is deleted silently and you can audit any call.

Re-sweep on cadence

Legitimate sites can degrade into MFA after a sale or redesign. Optional re-sweeps catch approved publishers that turned ad-first later.

The MFA tells

The specific signals that give MFA sites away

Each signal is extracted from what the site actually publishes and shipped with a confidence score. Red signals raise MFA risk; green signals mark the genuine publishers you want to keep.

High ad-to-content ratio Ad units dominate the fold Templated / spun articles Keyword-stuffed thin pages Signs of arbitraged traffic Generic stock-only imagery No real author / about identity Incoherent topic sprawl Aggressive interstitials / redirects Genuine editorial depth Real, identifiable author Coherent niche focus

Content signals, not traffic guesses. The filter is built on what the page contains and how it is laid out — all reproducible. Where traffic context helps, it comes only from CrUX presence and public rank group, never a modeled visitor number.

Side by side

A genuine niche publisher vs. an MFA lookalike

The two can share a topic and a template. The filter separates them on the signals that actually predict conversion. Descriptions are generalized; real domains ship in your deliverable.

Dimension	Genuine niche publisher (keep)	MFA lookalike (filter out)
Ad density	Ads present but subordinate to content	Ad units dominate; content fills gaps
Content	Original, specific, first-hand detail	Templated, thin, keyword-stuffed
Authorship	Named author with a real track record	Generic byline or none
Topic focus	Coherent niche the audience trusts	Sprawling, whatever ranks this week
Monetization intent	Recommends products it stands behind	Farms impressions; affiliate incidental
MFA-risk verdict	Low risk — passes to outreach	High risk — quarantined with reasons

A note on confidentiality

The examples on this page are illustrative and drawn from real screening output, so we follow standard confidentiality practice on public pages: domains are withheld and descriptions generalized. The published profiles are intentionally not traceable to any site — including through a web search — which protects the publishers without changing the underlying data. Your filtered deliverable and quarantine list contain the actual domains with every signal and confidence score, so each call can be verified directly. Start with a free pilot →

The deliverable

A clean list plus an auditable quarantine

You receive the filtered recruitment list and a separate quarantine of flagged domains, each with an MFA-risk grade, the triggering signals, and a confidence score. Nothing is removed silently. Values below are illustrative.

canonical_domain	mfa_risk	ad_density	content	author	topic_focus	reason	conf
candidate-a.example	low	moderate	original	named	coherent	keep	0.92
candidate-b.example	high	very_high	spun	none	sprawl	ad_first_layout	0.95
candidate-c.example	high	high	thin	generic	sprawl	thin_keyword_stuffed	0.89
candidate-d.example	medium	high	mixed	named	coherent	review_manually	0.66

You set the strictness: the MFA-risk threshold that filters versus routes to manual review is tuned to your program, and the quarantine keeps every borderline call reversible.

The measured basis

The same engine, proven at scale

The reading engine behind the MFA filter screened an entire vertical for a large-scale production run on our own classification platform — the reference for how content-quality judgments hold up across millions of domains.

classified domains as the base

vertical domains identified

domains screened in one run

precision on a reviewed sample

120M+

classified base

1.73M

vertical matched

1.2M

screened

~96%

precision confirmed

Reference: a large-scale production run on our own classification platform. Your MFA sweep is scoped to the list you send.

Questions

MFA & made-for-AdSense filtering — FAQ

What exactly is a made-for-advertising site?

It is a site built to farm ad impressions rather than serve readers — ad-dense layouts, thin templated content, and often bought traffic. Made-for-AdSense is the classic sub-type built around display ad networks. For an affiliate program they are dead weight: they look like niche publishers but send no buying intent.

Won't a plain ad-density check produce false positives?

Ad density alone would — plenty of legitimate publishers run display ads. That is why the filter combines layout with content authenticity, authorship, topic coherence, and monetization intent. A high-ad site with genuine original content and a real author is kept; the verdict names the signals that drove it, so you can audit every call.

Do you delete flagged sites or just mark them?

Neither is silent. You get a cleaned list plus a separate quarantine of flagged domains with the reasons and confidence scores. Nothing disappears — you can review the quarantine and pull anything back, which keeps borderline calls reversible.

Can you run this over a list I already have?

Yes. Send your existing recruitment or applicant list and we run the MFA sweep over it, returning the cleaned list and the quarantine. The same filter also runs inline on any full-web discovery run, so new lists arrive already scrubbed.

What does the free pilot show for MFA filtering?

We run the MFA filter on a sample of your list and return the cleaned set alongside the quarantine, each flag backed by its triggering signals and a confidence score — so you can see how the bar behaves against real sites before scoping a full sweep. No cost, no obligation.